PiDP-8/I Software

doc/*.png
examples/*.pt
labels/*.pdf
media/*/*.bin
media/*/*.dsk
media/*/*.pt
media/*/*.rk05
media/*/*.tu56
pics/*/*.jpg
pics/*/*.png

src/scp.*
src/sim_*.[ch]
src/sim_*.in
src/PDP8/pdp8_*.[ch]
src/PDP8/pidp8i.c.in

doc/simh/*.doc

# Creators and Major Contributors to the PiDP-8/I Project

*   **Oscar Vermeulen <oscar.vermeulen@hotmail.com>**:

    -   Creator of the project (both hardware and software)
    
    -   Author of the modifications to the SimH PDP-8 simulator
        necessary to make it use the PiDP-8/I front panel hardware
        
    -   Curator of the default set of binary demo media
    
    -   Author of the simulator setup scripts
    
    -   Initiator of much else in the project
    
    -   Author of the bulk of the documentation
    
    -   Host and major contributor to the PiDP-8/I support forum on
        Google Groups
        
    -   Hardware kit assembler and distributor

*   **Robert M Supnik** Primary author of the SimH PDP-8 simulator upon
    which this project is based.

*   **Mike Barnes** Ported Oscar Vermeulen's SimH 3.9 based PiDP-8/I
    simulator to the new SimH 4.0 code base.

*   **Dylan McNamee** Creator of the "buildroot" feature used in the
    creation of the official 2015.12.15 release versions.

*   **Mark G. Thomas** Creator of the installation scripts for the
    2015.12.15 release, which were folded into the `make install`
    handler within `Makefile.in`. Also wrote the version of the SysV
    init script that came with that release as `rc.pidp8`, shipped here
    as `pidp8i-init`.

*   **Ian Schofield <isysxp@gmail.com>** Modified the LED lamp driving
    code in the simulator to better simulate the incandescent lamps in
    the original PDP-8/I hardware.

*   **Henk Gooijen <henk.gooijen@boschrexroth.nl>** Pushed the PDP-8
    simulator's internal EAE step counter value down into the PiDP-8/I's
    LED manipulation code, without which the step counter LEDs remain
    dark even when using the EAE.

*   **Paul R. Bernard <prb@downspout.ca>** wrote `src/test.c` and the
    core of what now appears as `README-test.md`. (The program builds
    and installs `pidp8i-test`.) He also provided a one-line fix that
    completes the work of Henk Gooijen's step counter patch.

*   **Rick Murphy <k1mu.nospam@gmail.com>** optimized the `pep001.pal` *
    example so that it fits into a single page of PDP-8 core, and
    provided several useful files in his OS/8 disk images that have
    managed to land in this software distribution's OS/8 disk image.

*   **Tony Hill <hill.anthony@gmail.com>** Merged all the upstream SIMH
    changes between late September 2015 and late December 2016 into the
    PiDP-8/I simulator.

*   **Warren Young <tangentsoft@gmail.com>** Did everything listed in
    `ChangeLog.md` that is not attributed to anyone else.

# Licenses

The PiDP-8/I software distribution is an agglomeration of software from
multiple sources.  Several different licenses apply to its parts.  This
file guides you to those individual licenses.


## SIMH License

Most of the files in this software distribution are released under the
terms of the SIMH license, a copy of which typically appears at the top
of each file it applies to. This includes not only SIMH proper but also
several files written by PiDP-8/I software project contributors who
chose to license their contributions under the same license.

For a few files, textual inclusion of the license inside the file itself
was impractical, so this license is applied by reference to [a file
included with the distribution][sl].

[sl]: https://tangentsoft.com/pidp8i/doc/trunk/SIMH-LICENSE.md


## autosetup License

The `configure` script and the contents of the `autosetup` directory are
released under the FreeBSD license given in [`autosetup/LICENSE`][as].

[as]: https://tangentsoft.com/pidp8i/doc/trunk/autosetup/LICENSE


## palbart License

The `palbart` program and its manual page are released under the terms
of the license given in [`palbart/LICENSE.md`][pl].

[pl]: https://tangentsoft.com/pidp8i/doc/trunk/palbart/LICENSE.md


## OS/8 License

The OS/8 media images included with this software distribution are
released under the Digital License Agreement presented in
[`media/os8/LICENSE.md`][dla].

[dla]: https://tangentsoft.com/pidp8i/doc/trunk/media/os8/LICENSE.md


## Other DEC Software

The other files in the [`media`][md] and [`examples`][ed] directories
that originate from Digital Equipment Corporation are believed to fall
under the [public domain license][pdp8pd] DEC released all their PDP-8
software under after it stopped being ecomonmically viable. Documented
releases for specific software (e.g. TSS/8) may be difficult to come by,
however.

[md]: https://tangentsoft.com/pidp8i/dir?ci=trunk&name=media
[ed]: https://tangentsoft.com/pidp8i/dir?ci=trunk&name=examples


## ETOS License

ETOS was a commercial product produced outside of DEC. No public
documented declaration of license is known to be available for it, but
we have [a third-hand report][el] that its creators are fine with ETOS
being redistributed.

[el]: http://mailman.trailing-edge.com/pipermail/simh/2017-January/016169.html

# PiDP-8/I Changes

## Version 2017.01.23

*   When any program that talks to the PiDP-8/I front panel starts up,
    it now prints out a banner modeled on the [Erlang configuration
    line][ecl].  For example, when I run the software in the development
    tree on my PiDP-8/I, I get the following:

        PiDP-8/I trunk:i49cd065c [pi3b] [ils] [serpcb] [gpio]

    It tells me that:
    
    *   I'm running code built from Fossil checkin ID 49cd065c on the
        trunk branch, as opposed to a release version, which would be
        marked `release:v20170123` or similar.  (The `i` here is a tag
        standing for "ID", as in Fossil checkin ID.  Contrast `v` used
        to tag release version numbers.)

    *   I'm running it on a Raspberry Pi 3 Model B with Ian Schofield's
        incandescent lamp simulator (ILS) feature enabled.

    *   The software is built to expect that the PiDP-8/I PCB and the Pi
        board attached to it have had the serial mods made to them.

    *   The GPIO module found the GPIO hardware and was able to attach
        to it.

*   I get a very different result when running it on my desktop machine:

        PiDP-8/I trunk:id8536d91 [cake] [nls] [nopcb] [rt]

    This tells me:

    *   I'm running a different version of the development branch (i.e.
        the "trunk") of the code than what's running on the Pi.

    *   It's not running on a Pi at all.  (Cake ≠ pi.)

    *   I've disabled the ILS feature, so it's running with the "no lamp
        simulator" (NLS) GPIO module.
    
    *   Which is all to the good, because there's no point burning CPU
        power running the ILS code on a host machine that doesn't have a
        PiDP-8/I PCB attached.

    *   The GPIO thread is running with real-time privileges.

*   The ILS feature can now be disabled at `configure` time via the new
    `--no-lamp-simulator` flag.  This option is automatically set when
    building on a single-core Raspberry Pi.  (The flag is there only to
    allow someone building the software on a multi-core host machine to
    disable the ILS.)

*   Tweaked the ILS decay constants to be asymmetric, better mimicking
    the way real incandescent lamps work: they heat up to full
    brightness faster than they fade to perceptively "off."

*   The LED values used by the GPIO thread were being recalculated way
    too often.

    In the ILS case, it was updating the values approximately at the
    same rate as the ILS's PWM core frequency, roughly 7,500 times per
    second, which is far higher than the human persistence of vision
    limit.  While the PWM rate does need to be that fast to do its job,
    the underlying LED state values do not need to change nearly that
    often to fool the human into seeing instantaneous updates.

    The NLS case was actually worse, recalculating the LED values on
    every instruction executed by the PDP-8 CPU simulator, which even on
    a Pi 1 is likely to be a few MHz.

    In both the ILS and NLS cases, we now update the LED values about
    100 times a second, maintaining that rate dynamically based on the
    current execution speed of the simulator.

*   In prior versions, the ILS was only updating at its intended rate
    when the PDP-8 simulator was running flat-out on a current
    multi-core Raspberry Pi.  If you throttled the SIMH simulator to a
    slower execution rate, the display quality would start to degrade
    noticeably below about 1 MIPS.

*   With the prior fix, we now ship 5.script (i.e. the handler for
    starting the simulator with IF=5, or restarting it with IF=5 +
    `SING_STEP`) set to a throttle value of 30 kIPS, which allows the
    human to see each AC/MQ modification.  The built-in delay loops are
    still there, else we'd have to drop this to well under 1 kIPS.

*   The `SING_INST` switch now immediately puts the processor into
    single instruction mode, not requiring a separate press of the
    `STOP` key, as in prior versions.  This is the correct behavior
    according to the 1967-1968 edition of DEC's Small Computer Handbook
    for the PDP-8/I.

*   Greatly simplified the way single-instruction mode, processor
    stoppage, and the `CONT` key work.  The prior implementation was
    error-prone and difficult to understand.  This fixes a particularly
    bad interaction between the way `HLT` instructions and `CONT` key
    presses were handled, causing the processor to not resume correctly
    from `HLT` state.

*   Consolidated and cleaned up the bulk of the PiDP-8/I switch handling
    code so that it is not so intimately tied into the guts of the PDP-8
    CPU emulator.  This will greatly increase the chance that future
    updates to the upstream SIMH code will apply cleanly to our version.

*   Fixed a bug in `examples/bit-rotate.pal` which caused it to skip the
    actual bit rotation step.  We were trying to microcode two
    instructions into one that the PDP-8 won't accept together, and we
    didn't catch it until now because the HLT bug masked it, and the
    `palbart` assembler we ship isn't smart enough to notice the bug.

*   Fully generalized the mechanism for generating `obj/*.lst`,
    `bin/*.pt`, and `boot/*.script` from `examples/*.pal`.  You can now
    drop any PAL assembly language program into the `examples` directory
    and type `make` to build these various output forms automatically
    using the shipped version of `palbart`.  This effectively turns this
    PiDP-8/I software distribution into a PDP-8 assembly language
    development environment: rapidly build, test, and debug your PAL
    programs on your PC before you deploy them to real hardware.  Or,
    write PAL programs to debug the hardware or simulator, as we did
    with `examples/bit-rotate.pal`.

*   Fixed a sorting bug in the tool that generates `boot/*.script` from
    `obj/*.lst`, resulting in `dep` instructions that weren't sorted by
    core address.  This didn't cause any real problem, but it made
    tracing the execution of a PAL assembly program difficult if you
    were trying to refer to the `boot/*.script` file to check that the
    PiDP-8/I's front panel is showing the correct register values.

*   Updated SIMH to the latest upstream version and shipping a subset of
    the SIMH docs as unversioned files from tangentsoft.com.

*   The `configure` script now aborts the build if it sees that you're
    trying to build the software as root, since that means it generates
    the init script and the pidp8i script expecting to run the installed
    software as root, not as your normal user.  The most common way this
    happens is that you have a configured source tree, then change one
    of the `*.in` files and say `sudo make install`, thinking to build
    and install the change in one step.  This fixes that.

*   Several improvements to the build system.

[ecl]: http://stackoverflow.com/q/1182025/142454


## Version 2017.01.16

*   Prior releases did not include proper licensing for many of the
    included files.  This project was, therefore, not a proper Open
    Source Software project.  This problem has been fixed.

    In this release, many files that were previously technically only
    under standard copyright due to having no grant of license now have
    an explicit license, typically the same as SIMH itself.  (Thank you
    to all of the authors who allowed me to apply this license to their
    contributions!)

    For several other files, I was able to trace down some prior license
    and include its text here for the first time.

    There remain a few "gray" items: the TSS/8 and ETOS disk images.
    See the [`COPYING.md` file][copying] for more on the current status
    of these OS images.  If the legal status of these files clarifies in
    the future, this software distribution will react accordingly, even
    if that means removing these files from the distribution if we learn
    that these files are not freely-redistributable, as we currently
    believe them to be today.

*   The Step Counter LEDs on the front panel weren't being lit properly
    when EAE instructions were being used.  Thanks for this patch go to
    Henk Gooijen and Paul R. Bernard.

*   The prior `boot/1.script` and `boot/5.script` files are no longer
    simply opaque lists of octal addresses and machine code.  They are
    generated from PAL assembly files provided in the `examples`
    directory, so that you can now modify the assembly code and type
    `make` to rebuild these boot scripts.

*   The mechanism behind the prior item is fully general-purpose, not
    something that only works with `1.script` and `5.script`.  Any
    `examples/*.pal` file found at `make` time is transformed into a
    SIMH boot script named after the PAL file and placed in the `boot`
    directory.  This gives you an easier way to run PDP-8 assembly code
    inside the simulator.  After saying `make` to transform `*.pal` into
    `*.script` files, you can run the program with `bin/pidp8i-sim
    boot/my-program.script` to poke your program's octal values into
    core and run it.  This round-trip edit-and-run process is far faster
    than any of the options given in the [examples' `README.md`
    file][ex].

*   Disassembled both versions of the RIM loader to commented, labeled
    PAL assembly language files.  If you ever wanted to know what those
    16 mysterious instructions printed on the front panel of your
    PiDP-8/I did, you can now read my pidgin interpretation of these
    programs in `examples/*-rim.loader.pal` and become just as confused
    as I am now. :)

*   The two RIM loader implementations now start with a nonstandard
    `HLT` instruction so that when you fire up the simulator with IF=1 to
    start the high-speed RIM loader, it automatically halts for you, so
    you don't have to remember to STOP the processor manually.

    There is currently [a bug][hltbug] in the way the simulator handles
    `HLT` instructions which prevents you from simply pressing START or
    CONT to enter the RIM loader after you've attached your paper tape,
    so you still have to manually toggle in the 7756 starting address
    and press START to load the tape into core.  (I hope to fix this
    before the next release, but no promises.)

*   Added the `configure --throttle` feature for making the simulator
    run at a different speed than it normally does.  See
    [`README-throttle.md`][rmth] for details.

*   The build system now reacts differently when building the PiDP-8/I
    software on a single-core Raspberry Pi:

    *   If you're building the trunk or release branch, you'll get a
        configure error because it knows you can't run the current
        implementation of the incandescent lamp simulator on a
        single-core Pi.  (Not enough spare CPU power, even with heavy
        amounts of throttling.)

    *   If you're building the no-lamp-simulator branch, it inserts a
        throttle value into the generated `boot/*.script` files that do
        not already contain a throttle value so that the simulator
        doesn't hog 100% of the lone core, leaving some spare cycles for
        background tasks.  The above `--throttle` feature overrides
        this.

    These features effectively replace the manual instructions in the
    old `README-single-core.md` file, which is no longer included with
    this software distribution, starting with this release.

*   Lowered the real-time priority of the GPIO thread from 98 to 4.
    This should not result in a user-visible change in behavior, but it
    is called out here in case it does.  (In our testing, such high
    values simply aren't necessary to get the necessary performance,
    even on the trunk branch with the incandescent lamp simulator.)

*   Since v20161128, when you `make install` on a system with an
    existing PiDP-8/I software installation, the binary OS media images
    were not being overwritten, on purpose, since you may have modified
    them locally, so the installer chose not to overwrite your versions.

    With this release, the same principle applies to the SIMH boot
    scripts (e.g. `$prefix/share/boot/0.script`) since those are also
    things the user might want to modify.

    This release and prior ones do have important changes to some of
    these files, so if you do not wish to overwrite your local changes
    with a `make mediainstall` command, you might want to diff the two
    versions and decide which changes to copy over or merge into your
    local files.

[hltbug]:  https://tangentsoft.com/pidp8i/info/f961906a5c24f5de
[copying]: https://tangentsoft.com/pidp8i/doc/trunk/COPYING.md
[rmth]:    https://tangentsoft.com/pidp8i/doc/trunk/README-throttle.md


## Version 2017.01.05

*   Automated the process for merging in new SIMH updates.  From within
    the PiDP-8/I software build directory, simply say `make simh-update`
    and it will do its best to merge in the latest upstream changes.

    This process is more for the PiDP-8/I software maintainers than for
    the end users of that software, but if you wish to update your SIMH
    software without waiting for a new release of *this* software, you
    now have a nice automated system for doing that.

*   Updated SIMH using that new process.  The changes relevant to the
    PiDP-8/I since the prior update in release v20161226 are:

    *   Many more improvements to the simulator's internal timer system.
        This should make deliberate underclocking more accurate.

    *   It is now possible to get hex debug logs for the simulator console
        port by cranking up the simulator's debug level.

*   The simulator now reports the upstream Git commit ID it is based on
    in its version string, so that if you report bugs upstream to the
    SIMH project, you can give them a version number that will be
    meaningful to them.  (They don't care about our vYYYYMMDD release
    numbers or our Fossil checkin IDs.)


## Version 2016.12.26 (The Boxing Day release)

*   Tony Hill updated SIMH to the latest upstream version.

    This change represents about 15 months worth of work in the
    [upstream project][simh] — plus a fair bit of work by Tony to merge
    it all — so I will only summarize the improvements affecting the
    PDP-8 simulator here:

    *   Many improvements to the internal handling of timers.
    
        The most user-visible improvement is that you can now clock your
        emulated PDP-8 down to well below the performance of a real
        PDP-8 via `SET THROTTLE`, which can be useful for making
        blinkenlights demos run at human speeds without adding huge
        delay loops to the PDP-8 code implementing that demo.

    *   Increased the number of supported terminals from four to either
        twelve or sixteen, depending on how you look at it.  Eight of
        the additional supported terminal devices are conflict-free,
        while the final four variously conflict with one or more of the
        other features of the simulated PDP-8.  If you want to use all
        16, you will be unable to use the FPP, CT, MT and TSC features
        of the system.

        This limitation reflects the way the PDP-8 worked.  It is not an
        arbitrary limitation of SIMH.

    *   Added support for the LS8E printer interface option used by the
        WPS8 word processing system.

    *   The simulator's command console now shows the FPP register
        descriptions when using it as a PDP-8 debugger.

    *   Added the `SHOW TTIX/TTOX DEVNO` SIMH command to display the
        device numbers used for TTIX and TTOX.

    *   The `SHOW TTIX SUMMARY` SIMH command is now case-insensitive.

    *   Upstream improvements to host OS/compiler compatibility.  This
        increases the chances that this software will build out of the
        box on random non-Raspbian systems such as your development
        laptop running some uncommon operating system.

*   When you `make install`, we now disable Deeper Thought 2 and the
    legacy `pidp8` service if we find them, since they conflict with our
    `pidp8i` service.

*   Added the install user to the `gpio` group if you `make install` if
    that group is present at install time.  This is useful when building
    and installing the software on an existing Raspbian SD card while
    logged in as a user other than `pi` or `pidp8i`.

[simh]: https://github.com/simh/simh/


## Version 2016.12.18

*   The entire software stack now runs without explicit root privileges.
    It now runs under the user and group of the one who built the
    software.

    For the few instances where it does need elevated privileges, a
    limited-scope set of sudo rules are installed that permit the
    simulator to run the necessary helper programs.

*   The power down and reboot front panel switch combinations are no
    longer sensitive to the order you flip the switches.

*   Changed the powerdown front panel switch combination to the more
    mnemonically sensible `Sing_Step` + `Sing_Inst` + `Stop`.

    Its prior switch combo — `Sing_Step` + `Sing_Inst` + `Start` — is
    now the reboot sequence, with the mnemomic "restart."

*   Removed the USB stick mount/unmount front panel switch combos.  The
    automount feature precludes a need for a manual mount command, and
    unmount isn't necessary for paper tape images on FAT sticks.

*   The simulator now runs correctly on systems where the GPIO setup
    process fails.  (Basically, anything that isn't a Raspberry Pi.)
    Prior to this, this failure was just blithely ignored, causing
    subsequent code to behave as though all switches were being pressed
    at the same time, causing utter havoc.

    The practical benefit of this is that you can now work with the
    software on your non-Pi desktop machine, losing out only on the
    front panel LEDs and switches.  Everything else works just as on the
    Pi.  You no longer need a separate vanilla SimH setup.

*   Added a locking mechanism that prevents `pidpi8-test` and
    `pidp8i-sim` from fighting over the front panel LEDs.  While
    one of the two is running, the other refuses to run.

*   Added `examples/ac-mq-blinker.pal`, the PAL8 assembly code for the
    `boot/5.script` demo.

*   Fixed two unrelated problems with OS/8's FORTRAN IV implementation
    which prevented it from a) building new software; and b) running
    already-built binaries.  Thanks go to Rick Murphy for providing the
    working OS/8 images from which the files needed to fix these two
    problems were extracted.

*   Added the VT100-patched `VTEDIT` TECO macro from Rick Murphy's OS/8
    images, and made it automatically run when you run TECO from the
    OS/8 disk pack.  Also added documentation for it in `VTEDIT.DC` on
    the disk pack as well as [in the wiki][vteditdoc].

*   The default user name on the binary OS images is now `pidp8i`
    instead of `pi`, its password has changed to `edsonDeCastro1968`,
    and it demands a password change on first login.  I realize it's a
    hassle, but I decided I didn't want to contribute to the plague of
    open-to-the-world IoT boxes.

*   Many build system and documentation improvements.

[vteditdoc][https://tangentsoft.com/pidp8i/wiki?name=Using+VTEDIT]


## Version 2016.12.06

*   The `pidp8i-test` program's LED test routines did not work correctly
    when built against the incandescent lamp simulator version of the
    GPIO module.  Reworked the build so that this test program builds
    against the no-lamp-simulator version instead so that you don't have
    to choose between having the lamp simulator or having a working
    `pidp8i-test` program.

*   More improvements to `examples/pep001.pal`.

*   Extracted improved `PRINTS` routine from that example as
    `examples/routines/prints.pal`.


## Version 2016.12.05

*   This release marks the first binary SD card image released under my
    maintainership of the software.  As such, the major user-visible
    features in this release of the Fossil tree simply support that:

    *   The `pidp8i-init` script now understands that the OS's SSH host
        keys may be missing, and re-generates them.  Without this
        security measure, anyone who downloads that binary OS image
        could impersonate the SSH server on *your* PiDP-8/I.

    *   Added a `RELEASE-PROCESS.md` document.  This is primarily for my
        own benefit, to ensure that I don't miss a step, particularly
        given the complexity of producing the binary OS image.  However,
        you may care to look into it to see what goes on over here on
        the other side of the Internet. :)

*   Added an OS/8 BASIC solution to Project Euler Problem #1, so you can
    see how much simpler it is compared to the PAL8 assembly language
    version added in the prior release.

*   Updated the PAL8 assembly version with several clever optimizations
    by Rick Murphy, the primary effect of which is that it now fits into
    a single page of PDP-8 core memory.


## Version 2016.12.03

*   Debounced the switches.  See [the mailing list post][cdb] announcing
    this fix for details.

*   Merged the [`pidp8i-test` program][testpg] from the mailing list.
    The LED testing portion of this program [currently][gpiols] only works
    correctly without the incandescent lamp simulation patch applied.

*   Added a solution to [Project Euler Problem #1][pep001] in PAL8
    assembly language and wrote the [saga of my battle][p1saga] with
    this problem into the wiki.  This also adds a couple of useful PAL8
    routines in `examples/routines`.

*   Integrated David Gesswein's latest `palbart` program (v2.13) into
    the source distribution so that we don't have to:
    
    1.  ship pre-built RIM format paper tapes for the examples; and

    2.  put up with the old versions that OS package repos tend to have
        (Ubuntu is still shipping v2.4, from 6 years ago!)

*   Fixed a bug in the `make install` script that caused it to skip
    installing `screen` and `usbmount` from the OS's package repo when
    they are found to be missing.

*   Fixed a related bug that prevented it from disabling the serial
    console if you configure the software without `--serial-mod` and
    then install it, causing the serial console and the GPIO code in the
    PiDP-8/I simulator to fight over GPIO pins 14 and 15.

*   Removed the last of the duplicate binary media entries.  This makes
    the zip files for this version well under half the size of those for
    the 2015.12.15 upstream release despite having more features.

[cdb]:    https://groups.google.com/d/msg/pidp-8/Fg9I8OFTXHU/VjamSoFxDAAJ
[testpg]: https://groups.google.com/d/msg/pidp-8/UmIaBv2L9Ts/wB1CVeGDAwAJ
[gpiols]: https://tangentsoft.com/pidp8i/tktview?name=9843cab968
[pep001]: https://projecteuler.net/problem=1
[p1saga]: https://tangentsoft.com/pidp8i/wiki?name=PEP001.PA


## Version 2016.11.28

*   Added an intelligent, powerful build system, replacing the
    bare-bones `Makefile` based build system in the upstream version.
    See [`README.md`][readme] for more info on this.

*   The installation is now completely relocatable via `./configure
    --prefix=/some/other/place`. The upstream version would not work if
    installed somewhere other than `/opt/pidp8` due to many hard-coded
    absolute paths.  (This is enabled by the new build system, but
    fixing it was not simply a matter of swapping out the build system.)

*   Changed all of the various "PDP," "PDP-8", and "PiDP-8" strings to
    variants on "PiDP-8/I", partly for consistency and partly because it
    seems unlikely that this software will ever be used with anything
    other than the PiDP-8/I project.

    Part of this renaming means that the default installation location
    is now `/opt/pidp8i`, which has the nice side benefit that
    installing this version of the software will not overwrite an
    existing installation of the upstream version in `/opt/pidp8`.

    Another user-visible aspect of this change is that the upstream
    version's `pdp.sh` script to [re]enter the simulator is now called
    `pidp8i`.

*   Merged Ian Schofield's [Display update for the PiDP8][dupatch]
    patch.  Currently it is not optional, but there is [a plan][dudis] to
    allow this feature to be disabled via a `configure` script option.

*   The scripts that control the startup sequence of the PiDP-8/I
    simulator now include helpful header comments and emit useful
    status messages to the console.  Stare no more at opaque lists
    of SimH commands, wondering what each script does!

*   Merged `scanswitch` into the top-level `src` directory, since the
    only thing keeping it in a separate directory was the redundant
    `gpio.h` file. There were minor differences between the two `gpio.h`
    files, but their differences do not matter.

*   Installing multiple times no longer overwrites the binary OS/program
    media, since the disk images in particular may contain local
    changes.  If you want your media images overwritten, you can insist
    on it via `make mediainstall`.

*   The installation tree follows the Linux Filesystem Hierarchy
    Standard, so that files are in locations an experienced Linux user
    would expect to find them.  The biggest changes are that the content
    of the upstream `bootscripts` tree is now installed into
    `$prefix/share/boot`, and the OS/program media images which used to
    be in `imagefiles` are now in `$prefix/share/media`.

*   Added a bunch of ancillary material: [wiki articles][wiki],
    [USB stick label artwork][art], a PAL8 assembly [example program][ex]
    for you to toggle in, etc. Also filed a bunch of [tickets][tix]
    detailing feature proposals, known bugs and weaknesses, etc. If you
    were looking for ways to contribute to the development effort, these
    new resources provide a bunch of ideas.

*   Made some efforts toward portability.

    While this project will always center around Raspbian and the
    PiDP-8/I add-on board, the intent is that you should be able to
    unpack the project on any other Unix type system and at least get
    the simulator up and running with the normal SimH manual control
    over execution instead of the nice front panel controls provided by
    the PiDP-8/I board.

    In particular, the software now builds under Mac OS X, though it
    does not yet run properly.  (The modified SimH PDP-8 simulator
    currently misbehaves if the PiDP-8/I panel is not present.  Fixing
    this is on the radar.)

*   Fixed a bunch of bugs!

[readme]:  https://tangentsoft.com/pidp8i/doc/trunk/README.md
[dupatch]: https://groups.google.com/forum/#!topic/pidp-8/fmjt7AD1gIA
[dudis]:   https://tangentsoft.com/pidp8i/tktview?name=e06f8ae936
[wiki]:    https://tangentsoft.com/pidp8i/wcontent
[ex]:      https://tangentsoft.com/pidp8i/doc/trunk/examples/README.md
[art]:     https://tangentsoft.com/pidp8i/dir?c=trunk&name=labels
[tix]:     https://tangentsoft.com/pidp8i/tickets


## Version 2015.12.15

*   The official upstream release of the software, still current as of
    late 2016, at least.

Hacking on the PiDP-8/I Software
====

If you are going to make any changes to the PiDP-8/I software, here are
some rules and hints to keep in mind while you work.


Getting Started with Fossil
----

The PiDP-8/I software project is hosted using the [Fossil][fossil]
[distributed version control system][dvcs].  Fossil provides most of the
features of GitHub under a simpler operating model than Subversion
without tying you to a proprietary web service.

This guide will introduce you to some of the basics, but you should also
at least read the [Fossil Quick Start Guide][fqsd]. For a more thorough
introduction, I recommend [the Schimpf book][fbook]. If you have
questions, it is best to ask them on [its low-volumn mailing list][fml],
though you may also ask me, either on [the PiDP-8/I mailing list][ggml]
or via private email.

Most Raspberry Pi OS distributions include Fossil in their package
repository, and it is also available for all common desktop platforms.
If you started with one of the binary OS images downloaded from
tangentsoft.com, Fossil is already installed.  If you don't like any of
those options, you can also use [the official binaries][fbin].


[fbin]:   http://fossil-scm.org/index.html/uv/download.html
[dvcs]:   http://en.wikipedia.org/wiki/Distributed_revision_control
[fbook]:  http://www.fossil-scm.org/schimpf-book/home
[fml]:    http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/fossil-users
[fossil]: http://fossil-scm.org/
[fqsg]:   http://fossil-scm.org/index.html/doc/trunk/www/quickstart.wiki
[ggml]:   https://groups.google.com/forum/#!forum/pidp-8


Fossil Anonymous Access
----

To clone the code repository anonymously, say:

    $ mkdir ~/museum         # because where else do you store fossils?
    $ fossil clone https://tangentsoft.com/pidp8i ~/museum/pidp8i.fossil
    $ mkdir -p ~/src/pidp8i/trunk
    $ cd ~/src/pidp8i/trunk
    $ fossil open ~/museum/pidp8i.fossil

The second command gets you a file called `pidp8i.fossil` containing the
full history of PiDP-8/I from the upstream 2015.12.15 release onward.
You can call that clone file anything you like and put it in any
directory you like.  Even the `.fossil` extension is just a convention,
not a requirement.  (There is one feature of Fossil that requires that
file extension, but you probably won't use that feature.)


Working With Existing Tags and Branches
----

The directory structure shown in the commands above is more complicated
than strictly necessary, but it has a number of nice properties.

First, it collects other software projects under a common top-level
directory, which I'm calling `~/src`, but you are free to use any scheme
you like.

Second, the top-level project directory stores multiple separate
checkouts, one for each branch or tag I'm actively working with at the
moment.  So, to add a few other checkouts, you could say:

    $ cd ~/src/pidp8i
    $ mkdir -p release          # another branch
    $ mkdir -p v20151215        # a tag this time, not a branch
      ...etc...
    $ cd release
    $ fossil open ~/museum/pidp8i.fossil release
    $ cd ../v20151215
    $ fossil open ~/museum/pidp8i.fossil v20151215
      ...etc...

This gives you multiple independent checkouts.  The branch checkouts
remain pinned to the tip of that branch, so that if someone else checks
changes in on that branch and you say `fossil update`, those changes
appear in your checkout of that branch.  The tag checkouts behave
differently, always showing a specific checkout with the given tag name.

(In Fossil, tags and branches are related, but the details are beyond
our scope here.  See the [Fossil Quick Start Guide][fqsg] and the
documents it links to for more details.)

This directory scheme shows an important difference between Fossil and
Git: with Git, the checkout and the clone are intermingled in the same
directory tree, but in Fossil, they are strictly separate.  Git can
emulate Fossil's normal working style through its [worktree][gitwt]
feature, but it's a kind of lash-up using symlinks and such, whereas
with Fossil, there is no confusion: the repository clone is a single
SQLite database file — here, `pidp8i.fossil` — and the checkouts are
made from the contents of that database.

Another important difference relative to Git is that with Fossil, local
checkins attempt to automatically sync checked-in changes back to the
repository you cloned from.  (This only works if you have a login on the
remote repository, the subject of the next section.)  This solves a
number of problems with Git, all stemming from the fact that Git almost
actively tries to make sure every clone differs from every other in some
important way.

While Fossil does allow offline operation and local independent clones,
its default mode of operation is to try and keep the clones in sync as
much as possible.  Git works the way it does because it was designed to
meet the needs of the Linux kernel development project, which is
inherently federated, so Git tries to operate in a federated model as
well.  Fossil is better for smaller, more coherent teams, where there is
a single, clear goal for the project and a single source for its
official code.  Fossil helps remote developers cooperate, whereas Git
helps remote developers go off on their own tangents for extended
periods of time and optionally sync back up with each other
occasionally.

Fossil is a better match for the way the PiDP-8/I software project
works: we want you to cooperate closely with us, not go off on wild
tangents.

[gitwt]:  https://git-scm.com/docs/git-worktree


Fossil Developer Access
----

If you have a developer account on tangentsoft.com's Fossil instance, just
add your username to the URL like so:

    $ fossil clone http://username@tangentsoft.com/pidp8i pidp8i.fossil

Fossil will ask you for the password for `username` on the remote Fossil
instance, and it will offer to remember it for you.  If you let it
remember the password, operation from then on is scarcely different from
working with an anonymous clone, except that on checkin, your changes
will be sync'd back to the repository on tangentsoft.com if you're
online at the time.

If you're working offline, Fossil will still do the checkin, but you'll
be able to sync with the central repoisitory once you get back online.
It is best to work on a branch when unable to use Fossil's autosync
feature, as you are less likely to have a sync conflict when attempting
to send a new branch to the central server than in attempting to merge
your changes to the tip of trunk into the current upstream trunk, which
may well have changed since you went offline.

You can purposely work offline by disabling autosync mode:

    $ fossil set autosync 0

Until you re-enable it (`autosync 1`) Fossil will stop trying to sync
your local changes back to the central repo.  In this mode, Fossil works
more like Git's default mode, buying you many of the same problems that
go along with that working style.  I recommend disabling autosync mode
only when you are truly going to be offline, and don't want Fossil
attempting to sync when you know it will fail.


Getting Developer Access
----

The administrator of this repository is Warren Young, whose email you
can find on the [official PiDP-8/I project mailing list][ggml].
Developer access is available to anyone who makes a reasonable request.


Creating Branches
----

Creating a branch in Fossil is scary-simple, to the point that those
coming from other version control systems may ask, "Is that really all
there is to it?"  Yes, really, this is it:

    $ fossil ci --branch new-branch-name

That is to say, you make your changes as you normally would; then when
you go to check them in, you give the `--branch` option to the
`ci/checkin` command to put the changes on a new branch, rather than add
them to the same branch the changes were made against.

While developers with login rights to the PiDP-8/I Fossil instance are
allowed to check in on the trunk at any time, we recommend using
branches whenever you're working on something experimental, or where you
can't make the necessary changes in a single coherent checkin.
Basically, `trunk` should always build without error, and it should
always function correctly.  Branches are for isolating work until it is
ready to merge into the trunk.

Here again we have a difference with Git: because Fossil normally syncs
your work back to the central repository, this means we get to see the
branches you are still working on.  This is a *good thing*.  Do not fear
committing broken or otherwise bad code to a branch.  [You are not your
code.][daff]  We are software developers, too: we understand that
software development is an iterative process, and that not all ideas
spring forth perfect and production-ready from the fingers of its
developers.  These public branches let your collaborators see what
you're up to, and maybe lend advice or a hand in the work, but mostly
public branches let your collaborators see what you're up to, so they're
not surprised when the change finally lands in trunk.

This is part of what I mean about Fossil fostering close cooperation
rather than fostering wild tangents.

Jim McCarthy (author of [Dynamics of Software Development][dosd]) has a
presentation on YouTube that touches on this topic at a couple of
points:

* [Don't go dark](https://www.youtube.com/watch?v=9OJ9hplU8XA)
* [Beware of a guy in a room](https://www.youtube.com/watch?v=oY6BCHqEbyc)

Fossil's sync-by-default behavior fights these negative tendencies.

[daff]: http://www.hanselman.com/blog/YouAreNotYourCode.aspx
[dosd]: http://amzn.to/2iEVoBL


Debug Builds
----

By default, the build system creates a release build, but you can force
it to produce a binary without as much optimization and with debug
symbols included:

     $ ./configure --debug-mode


Manipulating the Build System Source Files
----

The [autosetup build system][asbs] is composed of these files and
directories:

     auto.def
     autosetup/
     configure
     Makefile.in

Unlike with GNU Autoconf, which you may be familiar with, the
`configure` script is not output from some other tool.  It is just a
driver for the Tcl and C code under the `autosetup` directory.  If you
have to modify any of these files to get some needed effect, you should
try to get that change into the upstream project, then merge that change
down into the local copy when it lands upstream.

The bulk of the customization to the build system is in `auto.def`,
which is a Tcl script run by `autosetup` via the `configure` script.
Some knowledge of [Tcl syntax][tcldoc] will therefore be helpful in
modifying it.

If you do not have Tcl installed on your system, `configure` builds a
minimal Tcl interpreter called `jimsh0`, based on the [Jim Tcl][jim]
project.  Developers working on the build system are encoruaged to use
this stripped-down version of Tcl rather than "real" Tcl because Jim Tcl
is more or less a strict subset of Tcl, so any changes you make that
work with the `jimsh0` interpreter should also work with "real" Tcl, but
not vice versa.  If you have Tcl installed and don't really need it,
consider uninstalling it to force `autosetup` to build and use `jimsh0`.

The `Makefile.in` file is largely a standard [GNU `make`][gmake] file
excepting only that it has variables substituted into it by
[`autosetup`][asbs] using its `@VARIABLE@` syntax.  At this time, we do
not attempt to achieve compatibility with other `make` programs, though
in the future we may need it to work with [BSD `make`][bmake] as well,
so if you are adding features, you might want to stick to the common
subset of features implemented by both the GNU and BSD flavors of
`make`.  We do not anticpate any need to support any other `make`
flavors.

(This, by the way, is why we're not using some heavy-weight build system
such as the GNU Autotools, CMake, etc.  The primary advantage of GNU
Autotools is that you can generate source packages that will configure
and build on weird and ancient flavors of Unix; we don't need that.
Cross-platform build systems such as CMake ease building the same
software on multiple disparate platforms straightforward, but the
PiDP-8/I software is built primarily on and for a single operating
system, Rasbpian Linux.  It also happens to build and run on other
modern Unix and Linux systems, for which we also do not need the full
power of something like CMake.  `autosetup` and GNU `make` suffice for
our purposes here.)

[asbs]:   http://msteveb.github.io/autosetup/
[bmake]:  https://www.freebsd.org/doc/en/books/developers-handbook/tools-make.html
[gmake]:  https://www.gnu.org/software/make/
[jim]:    http://jim.tcl.tk/
[tcldoc]: http://wiki.tcl.tk/11485


Submitting Patches
----

If you do not have a developer login on the PiDP-8/I software
repository, you can still send changes to the project.

The simplest way is to say this after developing your change against the
trunk of PiDP-8/I:

    $ fossil diff > my-changes.patch

Then attach that file to a new [PiDP-8/I mailing list][ggml] message
along with a declaration of the license you wish to contribute your
changes under.  We suggest using the [SIMH license][simhl], but any
[non-viral][viral] [OSI-approved license][osil] should suffice.

If your change is more than a small patch, `fossil diff` might not
incorporate all of the changes you have made.  The old unified `diff`
format can't encode branch names, file renamings, file deletions, tags,
checkin comments, and other Fossil-specific information.  For such
changes, it is better to send a Fossil bundle:

    $ fossil set autosync 0                # disable autosync
    $ fossil checkin --branch my-changes
      ...followed by more checkins on that branch...
    $ fossil bundle export --branch my-changes my-changes.bundle

After that first `fossil checkin --branch ...` command, any subsequent
changes will also be made on that branch without needing a `--branch`
option until you explicitly switch to some other branch.  This lets you
build up a larger change on a private branch until you're ready to
submit the whole thing as a bundle.

Because you are working on a branch on your private copy of the
PiDP-8/I Fossil repository, you are free to make as many checkins as
you like on the new branch before giving the `bundle export` command.

Once you are done with the bundle, send it to the mailing list just as
with the patch.

If you provide a quality patch, we are likely to offer you a developer
login on [the repository][repo] so you don't have to continue with the
patch or bundle methods.

Please make your patches or experimental branch bundles against the tip
of the current trunk.  PiDP-8/I often drifts enough during development
that a patch against a stable release may not apply to the trunk cleanly
otherwise.

[osil]:  https://opensource.org/licenses
[repo]:  http://tangentsoft.com/pidp8i/
[simhl]: https://tangentsoft.com/pidp8i/doc/trunk/SIMH-LICENSE.md
[viral]: https://en.wikipedia.org/wiki/Viral_license


The PiDP-8/I Software Project Code Style Rules
----

Every code base should have a common code style.  Love it or
hate it, here are PiDP-8/I's current code style rules:

**C Source Code**

File types: `c`, `h`, `c.in`

We follow the SIMH project's pre-existing code style when modifying one
of its source files:

*   Spaces for indents, size 4

*   DOS line endings.  (Yes, even though this is a Linux-based project!
    All decent Linux text editors can cope with this.)

*   Function, structure, type, and variable names are all lowercase,
    with underscores separating words

*   Macro names are in `ALL_UPPERCASE_WITH_UNDERSCORES`

*   Whitespace in the SIMH C files is of a style I have never seen
    anywhere else in my decades of software development.  This example
    shows the important features:

        int some_function (char some_parameter)
        {
        int some_variable = 0;

        if (some_parameter != '\0') {
            int nbytes = sizeof (some_parameter);
            char *buffer = malloc (4 * nbytes);

            switch (some_parameter) {
				case 'a':
					do_something_with_buffer ((char *)buffer); 
				default:
					do_something_else ();
                }
            }
        else {
            some_other_function (with_a_large, "number of parameters",
                wraps_with_a_single, "indent level");
            printf (stderr, "Failed to allocate buffer.\n");
            }
        }
    
    It is vaguely like K&R C style except that:

    -   The top level of statements in a function are not indented

    -   The closing curly brace is indented to the same level as the
        statement(s) it contains

    -   There is a space before all opening parentheses, not just those
        used in `if`, `while`, and similar flow control statements.

        Nested open parentheses do not have extra spaces, however.  Only
        the outer opening parenthesis has a space separating it from
        what went before.

    -   Multiple variables declared together don't have their types and
        variable names aligned in columns.

I find that this style is mostly sensible, but with two serious problems:
I find the indented closing curly braces confusing, and I find that the
loss of the first indent level for the statements inside a function makes
functions all visually run together in a screenful of code.  Therefore,
when we have the luxury to be working on a file separate from SIMH,
we use a variant of its style with these two changes, which you can
produce with this command:

    $ indent -kr -nce -cli4 -nlp -pcs -di1 -i4 -l100 \
        -ncs -ss -nbbo FILES...

That is, start with K&R, then:

-   nce:  don't cuddle else
-   cli4: indent case statement labels 4 spaces
-   nlp:  don't align continued statements at the opening parenthesis
-   pcs:  put a space before the opening parenthesis of a function call
-   di1:  don't line up variable types and names in separate columns
-   i4:   use 4-space indents
-   l100: allow lines up to 100 columns before forcibly breaking them
-   ncs:  don't put a space between a cast and its operand
-   ss:   add a space before semicolon with empty loop body
-   nbbo: don't break long lines before || and && operators

That gives the following, when applied to the above example:

        int some_function (char some_parameter)
        {
			int some_variable = 0;

			if (some_parameter != '\0') {
				int nbytes = sizeof (some_parameter);
				char *buffer = malloc (4 * nbytes);

				switch (some_parameter) {
					case 'a':
						do_something_with_buffer ((char *)buffer); 
					default:
						do_something_else ();
				}
			}
			else {
				some_other_function (with_a_large, "number of parameters",
					wraps_with_a_single, "indent level");
				printf (stderr, "Failed to allocate buffer.\n");
			}
        }
    
If that looks greatly different, realize that it is just two indenting
level differences: add one indent at function level, except for the
closing braces, which we leave at their previous position.

SIMH occasionally exceeds 100-column lines.  I recommend breaking
long lines at 72 columns.  Call me an 80-column traditionalist.

BSD `indent` does't understand the `-kr` option, so you can use this
alternative on BSD and macOS systems:

    $ indent -nce -cli4 -nlp -pcs -di1 -i4 -l100 \
            -bap -ncdb -nfc1 -npsl -nsc FILES...

When in doubt, mimic what you see in the current code.  When still in
doubt, ask on the mailing list.

[indent]: http://linux.die.net/man/1/indent


**Plain Text Files**

File types: `md`, `txt`

*   Spaces for indents, size 4.

*   Unix line endings.  The only common text editor I'm aware of that
    has a problem with this is Notepad, and people looking at these
    files anywhere other than unpacked on a Raspberry Pi box are
    probably looking at them through the Fossil web interface on
    tangentsoft.com.

*   Markdown files must follow the syntax flavor understood by
    [Fossil's Markdown interpreter][fmd].

[fmd]: https://tangentsoft.com/pidp8i/md_rules

########################################################################
# Makefile.in - Processed by autosetup's configure script to generate
#    the GNU make(1) file for building the PiDP-8/I software.
#
# If you are seeing this at the top of a file called Makefile and you
# intend to make edits, do that in Makefile.in.  Saying "make" will then
# re-build Makefile from that modified Makefile.in before proceeding to
# do the "make" operation.
#
# Copyright © 2015-2017 Oscar Vermeulen, Warren Young
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
# IN NO EVENT SHALL THE AUTHORS LISTED ABOVE BE LIABLE FOR ANY CLAIM,
# DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT
# OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
# OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
#
# Except as contained in this notice, the names of the authors above
# shall not be used in advertising or otherwise to promote the sale,
# use or other dealings in this Software without prior written
# authorization from those authors.
########################################################################

# Git commit ID of the latest version of the SIMH 4 project on GitHub 
# that has been merged into this source base.
SGCID=0046905f72e2372a8bf5dfc90425fd5214b0ba8a

CFLAGS = @CFLAGS@ -Wno-unused-result -Wno-parentheses @BUILDMODE@ \
	-DUSE_READER_THREAD -DHAVE_DLOPEN=so -DPIDP8I -DSIM_ASYNCH_IO \
	-DHAVE_REGEX_H -DHAVE_GLOB -DSIM_GIT_COMMIT_ID=$(SGCID) \
	-U__STRICT_ANSI__ \
	-I @srcdir@/src -I @srcdir@/src/PDP8 -I @builddir@/src

SIM = bin/pidp8i-sim
BINS = bin/palbart $(SIM) bin/pidp8i-test libexec/scanswitch

BUILDDIRS = bin libexec obj/PDP8

INSTDIRS = bin etc libexec share/boot share/media share/man/man1

OBJS = \
	obj/gpio-common.o \
	obj/sim_console.o \
	obj/PDP8/pdp8_df.o \
	obj/PDP8/pdp8_cpu.o \
	obj/PDP8/pdp8_clk.o \
	obj/PDP8/pdp8_ct.o \
	obj/PDP8/pdp8_dt.o \
	obj/PDP8/pdp8_fpp.o \
	obj/PDP8/pdp8_lp.o \
	obj/PDP8/pdp8_mt.o \
	obj/PDP8/pdp8_pt.o \
	obj/PDP8/pdp8_rf.o \
	obj/PDP8/pdp8_rk.o \
	obj/PDP8/pdp8_rl.o \
	obj/PDP8/pdp8_rx.o \
	obj/PDP8/pdp8_sys.o \
	obj/PDP8/pdp8_td.o \
	obj/PDP8/pdp8_tsc.o \
	obj/PDP8/pdp8_tt.o \
	obj/PDP8/pdp8_ttx.o \
	obj/PDP8/pidp8i.o \
	obj/scp.o \
	obj/sim_disk.o \
	obj/sim_ether.o \
	obj/sim_fio.o \
	obj/sim_serial.o \
	obj/sim_sock.o \
	obj/sim_tape.o \
	obj/sim_timer.o \
	obj/sim_tmxr.o

LIBS = -lm -ldl -lpthread

EXAMPLES := $(wildcard @srcdir@/examples/*.pal)
EXAMPLES := $(subst @srcdir@/examples,bin,$(EXAMPLES))
EXAMPLES := $(EXAMPLES:.pal=.pt)
LISTINGS := $(EXAMPLES:.pt=.lst)
LISTINGS := $(subst bin/,obj/,$(LISTINGS))
BOOTSCRIPTS := $(LISTINGS:.lst=.script)
BOOTSCRIPTS := $(subst obj/,boot/,$(BOOTSCRIPTS)) \
	boot/1.script \
	boot/5.script

# List of *.in files from auto.def file, except for this present file
# (Makefile.in) which is handled separately.  This list should only
# change when the list of "make-template" calls in auto.def changes.
#
# If the first file listed below changes, change the AUTOREBUILD rule
# near the end of this file to match!
INFILES = \
	@srcdir@/bin/pidp8i.in \
	@srcdir@/boot/0.script.in \
	@srcdir@/boot/2.script.in \
	@srcdir@/boot/3.script.in \
	@srcdir@/boot/4.script.in \
	@srcdir@/boot/6.script.in \
	@srcdir@/boot/7.script.in \
	@srcdir@/etc/pidp8i-init.in \
	@srcdir@/etc/sudoers.in \
	@srcdir@/examples/Makefile.in \
	@srcdir@/src/Makefile.in \
	@srcdir@/src/gpio-common.c.in \
	@srcdir@/src/PDP8/Makefile.in \
	@srcdir@/src/PDP8/pidp8i.c.in \
	@srcdir@/src/scp.c.in \
	@srcdir@/tools/simh-update.in
OUTFILES := $(subst @srcdir@/,,$(INFILES))
OUTFILES := $(subst .in,,$(OUTFILES))

CLTXT = /boot/cmdline.txt

.PHONY: tags

all: $(OUTFILES) $(BUILDDIRS) $(BINS) $(BOOTSCRIPTS) $(LISTINGS) $(EXAMPLES)
	@chmod 755 @builddir@/bin/pidp8i

clean:
	@rm -f $(BINS) $(BOOTSCRIPTS) $(EXAMPLES) $(LISTINGS) $(OBJS) \
		@builddir@/tags \
			@builddir@/obj/*.d \
		@builddir@/obj/*.o \
		@builddir@/obj/PDP8/*.d \
		@srcdir@/examples/*.err
	@for f in $(OUTFILES) ; do test "$$f" = "$${f/Makefile//}" && rm $$f ; done
	@-rmdir -p $(BUILDDIRS) 2> /dev/null || true

distclean: clean
	@rm -f \
		@builddir@/config.log \
		@builddir@/Makefile \
		@builddir@/autosetup/jimsh0 \
		@builddir@/examples/Makefile \
		@builddir@/src/Makefile \
		@builddir@/src/config.h \
		@builddir@/src/PDP8/Makefile

ctags tags:
	ctags -R @srcdir@

install: all
	@echo Installing to @prefix@...

	@# Create any missing install tree directories
	for d in $(INSTDIRS) ; do @INSTALL@ -m 755 -d @prefix@/$$d ; done

	@# Install files into those dirs and set their perms
	for f in $(BINS) ; do @INSTALL@ -m 755 -D -s $$f @prefix@/$$f ; done
	@INSTALL@ -m 755 @srcdir@/bin/pidp8i @prefix@/bin
	-for f in @prefix@/bin/pidp8i-* ; do setcap 'cap_sys_nice=eip' $$f ; done || true
	test -e @MEDIADIR@/os8/os8.rk05 || $(MAKE) mediainstall

	@# If this is a Debian-type system, install needed helper programs
	@test -x /usr/bin/apt-get -a ! -h /media/usb && apt-get -y install usbmount || true
	@test -x /usr/bin/apt-get -a ! -x /usr/bin/screen && apt-get -y install screen || true

	@# Disable competing services if this is a Raspberry Pi
	@(test -x /bin/systemctl && /bin/systemctl disable deeper || true)
	@(test -x /bin/systemctl && /bin/systemctl disable pidp8  || true)

	@# Install the init script if this system is systemd based.
	@INSTALL@ -m 750 @srcdir@/etc/pidp8i-init @prefix@/etc
	@(  test -w /etc/init.d -a -x /bin/systemctl && \
		ln -sf @ABSPREFIX@/etc/pidp8i-init /etc/init.d/pidp8i && \
		/bin/systemctl enable pidp8i \
	) || true

	@# Give the install user permission to use GPIO if done on a Pi
	@grep -q '^gpio:' /etc/group && usermod -a -G gpio @INSTUSR@ || true

	@# Give the install user permission to shut down and reboot the Pi
	@# if this is a systemd/sudo based system.
	@(	test -d /etc/sudoers.d -a -w /etc/sudoers.d -a -x /bin/systemctl && \
		@INSTALL@ -m 440 -o root -g root @srcdir@/etc/sudoers \
		    /etc/sudoers.d/099_pidp8i \
	) || true

	@# Add installation bin dir to the non-root user's PATH unless it's
	@# already in there or we aren't running under sudo.
	@(for p in .profile .bash_profile ; do \
		test -n "$$SUDO_USER" -a -w "/home/$$SUDO_USER/$$p" && \
			! grep -qF "@ABSPREFIX@/bin" "/home/$$SUDO_USER/$$p" && \
			echo "export PATH=\$$PATH:@ABSPREFIX@/bin" >> "/home/$$SUDO_USER/$$p" ; \
	done) || true

	@# Ditto for MANPATH
	@(for p in .profile .bash_profile ; do \
		test -n "$$SUDO_USER" -a -w "/home/$$SUDO_USER/$$p" && \
			! grep -qF "@ABSPREFIX@/share/man" "/home/$$SUDO_USER/$$p" && \
			echo "export MANPATH=\$$MANPATH:@ABSPREFIX@/share/man" >> "/home/$$SUDO_USER/$$p" ; \
	done) || true

	@# If serial mod is disabled, turn off serial console and kgdb stuff
	@# in case they were enabled previously, else they will fight with
	@# our use of GPIO.
	@(  test -z "@PCB_SERIAL_MOD@" -a -r $(CLTXT) && ! -w $(CLTXT) && \
		cp -p $(CLTXT) "$(CLTXT)"_orig && \
		sed -e 's/console\=[a-zA-Z0-9]+,[0-9]+ //' \
		    -e  's/kgdboc\=[a-zA-Z0-9]+,[0-9]+ //' -i $(CLTXT) \
	) || true

	@# Install palbart stuff
	@INSTALL@ -m 755 @builddir@/bin/palbart @prefix@/bin
	@INSTALL@ -m 644 @srcdir@/palbart/palbart.1 @prefix@/share/man/man1

mediainstall:
	@echo "[Re]installing OS and program media..."
	@cd @srcdir@ ; \
	find media \( \
		-name \*.bin  -o \
		-name \*.dsk  -o \
		-name \*.rk05 -o \
		-name \*.tu56 \
	\) -exec @INSTALL@ -D -m 664 -g @INSTGRP@ {} @ABSPREFIX@/share/{} \;
	@INSTALL@ -m 664 -g @INSTGRP@ @builddir@/boot/*.script @BOOTDIR@

os8test:
	@builddir@/$(SIM) @srcdir@/boot/0.script

reconfig:
	@AUTOREMAKE@

release: all
	@srcdir@/tools/mkrel

simh-update simh-update-f:
	@@builddir@/tools/simh-update $(subst simh-update,,$@)


# Rule for compiling *.c to *.o and autogenerating dependency info.
# Explained at http://scottmcpeak.com/autodepend/autodepend.html
#
# Reflect any changes here into near-duplicate below!
obj/%.o: @srcdir@/src/%.c
	$(CC)  -c $(CFLAGS) @srcdir@/src/$*.c -o obj/$*.o
	$(CC) -MM $(CFLAGS) @srcdir@/src/$*.c > obj/$*.d
	@mv -f obj/$*.d obj/$*.d.tmp
	@sed -e 's|.*:|obj/$*.o:|' < obj/$*.d.tmp > obj/$*.d
	@sed -e 's/.*://' -e 's/\\$$//' < obj/$*.d.tmp | fmt -1 | \
		sed -e 's/^ *//' -e 's/$$/:/' >> obj/$*.d
	@rm -f obj/$*.d.tmp

# Near-duplicate of above rule for those *.c files generated from *.c.in
# files in the srcdir.  Needed when building out-of-tree.
obj/%.o: @builddir@/src/%.c
	$(CC)  -c $(CFLAGS) @builddir@/src/$*.c -o obj/$*.o
	$(CC) -MM $(CFLAGS) @builddir@/src/$*.c > obj/$*.d
	@mv -f obj/$*.d obj/$*.d.tmp
	@sed -e 's|.*:|obj/$*.o:|' < obj/$*.d.tmp > obj/$*.d
	@sed -e 's/.*://' -e 's/\\$$//' < obj/$*.d.tmp | fmt -1 | \
		sed -e 's/^ *//' -e 's/$$/:/' >> obj/$*.d
	@rm -f obj/$*.d.tmp

# Rule for building example PAL assembly language programs.
obj/%.lst bin/%.pt: @srcdir@/examples/%.pal bin/palbart
	@builddir@/bin/palbart -lr $< || cat @srcdir@/examples/$*.err
	mv @srcdir@/examples/$*.lst @builddir@/obj/$*.lst
	mv @srcdir@/examples/$*.rim @builddir@/bin/$*.pt

# Rule for translating PAL assembly language program listings to SIMH
# boot scripts.
boot/%.script: obj/%.lst
	@srcdir@/tools/mkbootscript $<

# Rules for making aliases of named example programs translated to boot
# scripts as special numbered boot scripts
boot/1.script: boot/hs-rim-loader.script
	ln -f $< $@
boot/5.script: boot/ac-mq-blinker.script
	ln -f $< $@

bin/palbart: @srcdir@/palbart/palbart.c
	$(CC) -Wall -Wno-strict-prototypes @BUILDMODE@ $< -o $@
	$(CC) -MM $< -o obj/$*.d

$(BUILDDIRS):
	mkdir -p $@

$(SIM): $(OBJS) obj/gpio-@LED_DRIVER_MODULE@ls.o
	$(CC) -o $@ $^ $(LIBS)

bin/pidp8i-test: obj/test.o obj/gpio-nls.o obj/gpio-common.o
	$(CC) -o $@ $^ $(LIBS)

libexec/scanswitch: obj/scanswitch.o obj/gpio-nls.o obj/gpio-common.o
	$(CC) -o $@ $^

ifeq ($(findstring clean,$(MAKECMDGOALS)),)
Makefile: @srcdir@/Makefile.in @srcdir@/auto.def $(INFILES) @AUTODEPS@
	@AUTOREMAKE@ && $(MAKE)

# If you simply make $(OUTFILES) depend on $(INFILES), make(1) thinks it
# can build all of $(OUTFILES) in parallel if given -jN, which causes
# ./configure --args to run N times in parallel, which blows shit up
# right good and proper.  We purposely list only the first file in
# $(INFILES) here to prevent that.  The other files in $(INFILES) will
# also be built, which confuses make(1) somewhat, but it figures things
# out on the subsequent $(MAKE) call.
bin/pidp8i: @srcdir@/bin/pidp8i.in
	@AUTOREMAKE@ && $(MAKE)
endif

-include $(OBJS:.o=.d) obj/scanswitch.d

# PCB Test Program


## Compiling and Installing

`pidp8i-test` is a simple program to test [Oscar Vermeulen's PiDP-8/I
Kit][project] during construction. It is built and installed alongside
the other software with the normal `make` process.


## Running It

If you are building the software on the Pi for the first time, log out
of your user account after installing it, then back in so that the
install script's changes to your user's `PATH` take effect.

Thereafter, simply give these commands:

    $ sudo systemctl stop pidp8i
	$ pidp8i-test

The first command ensures that the modified PDP-8 simulator is stopped
during the test, since only one program can be talking to the switch and
LED array at a given time. (This also applies to other programs like
[Deeper Thought 2][dt2].)


## Test Procedure

You can at any time hit Ctrl-C to stop the test.

The test proceeds as follows:

*   All On test:

    It turns on all LEDs for 5 seconds.

*   All Off test:

    It turns off all LEDs for 5 seconds.

*   Row test:

    It turns on one full row of LEDs and pauses for 5 seconds, then
    switches to the next row.  There are eight rows of LEDs of up to 12
    LEDs each.

*   Column test:

    It then turns on one full column of LEDs and pauses for 5 seconds,
    then switches to the next column.  There are 12 columns of LEDs with
    up to 8 LEDs each.  (Some of the LEDs positions in a column are
    sometimes rather chaotic, it will require intimate knowledge of the
    schematic to verify.  It's somewhat of a useless test but it might
    turn up an assembly error for someone.)

*   Switch test:

    It then goes into a single LED chase pattern and starts looking at
    switches.  This loop is infinite.  Every time it detects a change in
    the switch positions it prints out the full Octal bit pattern for
    the three switch banks.  No attempt is made to name the actual
    switch that has been flipped.  The goal is to verify switch
    functionality, not to debug the design of the circuit or the driver.

    When running this test, if you get a new line printed with a single
    bit change when you flip a single switch, the switch in question is
    working.  If you get no output printed or multiple bits changed in
    the output printed something is wrong.

    If for some reason you need to decode the output bits to physical
    switches they appear as follows:

    | A    | B    | C   
    |-------------------
    | 4000 | 0000 | 0000

    The first twelve bits (labelled A) is the Switch Register.  The bits
    left to right correspond to the SR switches also left to right.  So
    above the SR1 switch is toggled down, ie 1.  Every other SR switch
    is up, ie 0.

    The leftmost 6 bits (labelled B) are the 3 DF switches followed by
    the 3 IF switches.  Again left to right.  The rest of the bits are
    unused in the B section.

    The leftmost 8 bits (labelled C) are the remaining 8 switches
    starting at "START" and ending at "SING INST".  Again Left to right.


## License

This document is licensed under the same terms as the associated
[`src/test.c` program][program].


[project]: http://obsolescence.wix.com/obsolescence#!pidp-8
[dt2]:     https://github.com/VentureKing/Deeper-Thought-2
[program]: https://tangentsoft.com/pidp8i/doc/trunk/src/test.c

# Throttling the Simulator

When you do not give the `--throttle` option to the `configure` script,
the simulator's speed is set based on the number of CPU cores detected
by the `tools/corecount` script.


## Multi-Core Default

If `corecount` detects a multi-core system, the default behavior is to
not throttle the simulator at all, since there are only 2 threads in the
software that take substantial amounts of CPU power.

The most hungry thread is the PDP-8 simulator proper, which runs flat-out,
taking an entire core's available power by default.

The other hungry thread is the one that drives the front panel LEDs,
which takes about 15% of a single core's power on a Raspberry Pi 3 when
you build the software with the incandescent lamp simulator enabled.

This leaves over 2 cores worth of CPU power untapped on multi-core
Raspberry Pis, so the system performance remains snappy even with the
simulator running.

You can force this behavior with `--throttle=none`.


## Single-Core Default

If the `configure` script decides that you're building this on a
single-core system, it purposely throttles the PDP-8 simulator so that
it takes about 75% of a single core's worth of power on the slowest
Raspberry Pi supported by this software. This leaves enough CPU power
for some background tasks on a single-core Pi.

This default assumes you are building without the incandescent lamp
simulator feature enabled, as that currently takes so much CPU power to
run that the simulator runs slower than even a PDP-8/S!  (We're working
on ways to improve the speed of that lamp simulator to let it run on
single-core raspberry Pis.)  Indeed, the build system will actively try
to prevent you from building the incandescent lamp simulator feature on
a single-core Pi.

You can force the build system to select this throttle value even on a
multi-core Pi with `--throttle=single-core`.

You will erroneously get this single-core behavior if you run the
`configure` script on a system where `tools/corecount` has no built-in
way to count the CPU cores in your system correctly, so it returns 1,
forcing a single-core build. That script currently only returns the
correct value on Linux, BSD, and macOS systems.  To fix it, you can
either say `--throttle=none` or you can patch `tools/corecount` to
properly report the number of cores on your system.  If you choose the
latter path, please send the patch to the mailing list so it can be
integrated into the next release of the software.


## Underclocking

If you want the software to run even slower, there are additional
`configure --throttle` option values available to achieve that:

*   `--throttle=STRING`: any value not otherwise understood is passed
    directly to SIMH in `SET THROTTLE` commands inserted into the
    generated `boot/*.script` files. You can use any string here that
    SIMH itself supports; RTFM.

*   `--throttle=CPUTYPE`: if you give a value referencing one of the
    many PDP-8 family members, it selects a value based on the execution
    time of `TAD` in direct access mode on that processor:

    | Value            | Alias For | Memory Cycle Time
    ---------------------------------------------------
    | `pdp8e`          | 416k      | 1.2 µs
    | `pdp8i`, `pdp8a` | 333k      | 1.5 µs
    | `pdp8l`, `pdp8`  | 313k      | 1.6 µs
    | `ha6120`         | 182k      | 2.7 µs
    | `im6100a`        | 200k      | 2.5 µs
    | `im6100`         | 100k      | 5 µs
    | `im6100c`        | 83k       | 6 µs
    | `pdp8s`          | 63k       | 8 µs

    I chose `TAD` because it's a typical instruction for the processor,
    and its execution speed is based on the memory cycle time for the
    processor, an easy specification to find.  Other instructions (e.g.
    most OPR instructions) execute faster than this, while others (e.g.
    IOT) execute far slower.  (See the processor's manual for details.)

    SIMH, on the other hand, does not discriminate.  When you say
    `--throttle=pdp8i`, causing the build system to insert `SET THROTTLE
    333k` commands into the SIMH boot scripts, the SIMH PDP-8 simulator
    does its best to execute exactly 333,000 instructions per second,
    regardless of the instruction type.  Consequently, if you were to
    benchmark this simulator configured with one of the options above,
    there would doubtless be some difference in execution speed,
    depending on the mix of instructions executed.
    
    (See the I/O Matters section below for a further complication.)

    The values for the Intersil and Harris CMOS microprocessors are for
    the fastest clock speed supported for that particular chip. Use the
    `STRING` form of this option if you wish to emulate an underclocked
    microprocessor.

*   `--throttle=human`: Causes the computer to throttle the human.

    "I'm sorry, Dave, but you are not worthy to run this software."

    "Aaackkthhhpptt..."

    No, wait, that can't be right.
    
    Let's see here...ah, yes, what it *actually* does is slows the
    processor down to 10 instructions per second, about the fastest that
    allows the human eye to easily discern LED state changes as
    separate.  If you increase it very much above this, the eye starts
    seeing the LED state changes as a blur.

    This mode is useful for running otherwise-useful software as a
    "blinkenlights" demo.
    
*   `--throttle=trace`: Alias for `--throttle=1`, causing the simulator
    to act more or less like it's in single-instruction mode and you're
    pressing the `CONT` button once a second to step through a program.


## I/O Matters

The throttle mechanism discussed above only affects the speed of the
PDP-8 CPU simulator. It does not affect the speed of I/O operations.

The only I/O channel you can throttle in the same way is a serial
terminal by purposely choosing a slower bit rate for it than the maximum
value. If you set it to 110 bps, it runs at the speed of a Teletype
Model 33 ASR, the most common terminal type used for the PDP-8/I, and
most other early PDP-8 flavors. Later PDP-8s were often paired with (or
integrated into!) glass TTYs such as the VT05, which flew along at 2400
bps. Then things got really fancy with the VT52, which screamed along at
9600 bps. Wowee!

I'm not aware of a way to make SIMH slow the other I/O operations, such
as disk access speeds, in order to emulate the speed of the actual
hardware.


## License

Copyright © 2017 by Warren Young. This document is licensed under
the terms of [the SIMH license][sl].

[sl]: https://tangentsoft.com/pidp8i/doc/trunk/SIMH-LICENSE.md

# Getting Started with the PiDP-8/I Software

## Prerequisites

*   A Raspberry Pi with the 40-pin GPIO connector.  That rules out the
    first-generation Raspberry Pi model A and B boards which had a
    26-pin GPIO connector.

*   An SD card containing Raspbian or something sufficiently close.
    This software is currently tested with the Jessie Lite distribution.

    Ideally, you will install a fresh OS image onto an unused SD card
    rather than use this software to modify an existing OS installation,
    but there is currently no known hard incompatibilty that prevents
    you from integrating this software into an existing OS.

*   This software distribution, unpacked somewhere convenient within the
    Raspberry Pi filesystem.

    Unlike with the [upstream 2015.12.15 release][upst], this present
    release of the software should *not* be unpacked into `/opt/pidp8`.
    I recommend that you unpack it into `$HOME/src`, `/usr/local/src` or
    similar, but it really doesn't matter where you put it, as long as
    your user has full write access to that directory.

*   A working C compiler and other standard Linux build tools, such as
    `make(1)`.  On Debian type systems — including Raspbian — you can
    install such tools with `sudo apt install build-essential`


## Configuring, Building and Installing

This software distribution builds and installs in the same way as most
other Linux/Unix software these days.  The short-and-sweet is:

    $ ./configure && make && sudo make install


### Configure Script Options

You can change a few things about the way the software is built and
installed by giving options to the `configure` script:


#### --prefix

Perhaps the most widely useful `configure` script option is `--prefix`,
which lets you override the default installation directory,
`/opt/pidp8i`.  You could make it install the software under your home
directory on the Pi with this command:

    $ ./configure --prefix=$HOME/pidp8i && sudo make install

Although this is installing to a directory your user has write access
to, you still need to install via `sudo` because the installation
process does other things that do require `root` access.


#### --no-lamp-simulator

If you build the software on a multi-core host, the PDP-8/I simulator is
normally built with Ian Schofield's incandescent lamp simulator feature,
which drives the LEDs in a way that mimics the incandescent lamps used
in the original PDP-8/I.  This feature currently takes too much CPU
power to run on anything but a multi-core Raspberry Pi, currently
limited to the Pi 2 and Pi 3 series.

If you configure the software on a single-core Pi — models A+, B+, and
Zero — the simulator uses the original low-CPU-usage LED driving method
instead.

Those on a multi-core host who want this low-CPU-usage LED driving
method can give the `--no-lamp-simulator` option to `configure`.


#### --serial-mod

If you have done the [serial mod][smod] to your PiDP-8/I PCB and the
Raspberry Pi you have connected to it, add `--serial-mod` to the
`configure` command above.

If you do not give this flag at `configure` time with these hardware
modifications in place, the front panel will not work correctly, and
trying to run the software may even crash the Pi.

If you give this flag and your PCBs are *not* modified, most of the
hardware will work correctly, but several lights and switches will not
work correctly.


#### --throttle

See [`README-throttle.md`][thro] for the values this option takes.  If
you don't give this option, the simulator runs as fast as possible, more
or less.


#### --help

Run `./configure --help` for more information on your options here.


### Installing

The `sudo make install` step in the command above does what most people
want.

That step will not overwrite the operating system and program media
(e.g. the OS/8 RK05 disk cartridge image) when installing multiple times
to the same location, but you can demand an overwrite with:

    $ sudo make mediainstall

This can be helpful if you have damaged your OS/program media or simply
want to return to the pristine versions as distributed.

This will also overwrite the boot scripts in `$prefix/share/boot` with
fresh versions from the source distribution.


## Testing

You can test your PiDP-8/I LED and switch functions with these commands:

    $ sudo systemctl stop pidp8i
	$ pidp8i-test

You may have to log out and back in before the second command will work,
since the installation script modifies your normal user's `PATH` the
first time you install onto a given system.

It is important to stop the PiDP-8/I simulator before running the test
program, since both programs need exclusive access to the LEDs and
switches on the front panel.  After you are done testing, you can start
the PiDP-8/I simulator back up with:

    $ sudo systemctl start pidp8i

See [`README-test.md`][test] for more details.


## Using the Software

For the most part, this software distribution works like the upstream
[2015.12.15 distribution][usd].  Its [documentation][prj] therefore
describes this software too, for the most part.

The largest user-visible difference between the two software
distributions is that all of the shell commands affecting the software
were renamed to include `pidp8i` in their name:

1.  To start the simulator:

        $ sudo systemctl start pidp8i

    This will happen automatically on reboot unless you disable the
    service, such as in order to run one of the various [forks of Deeper
    Thought][dt2].

2.  To attach the terminal you're working on to the simulator:

        $ pidp8i

3.  To detach from the simulator's terminal interface while leaving the
    PiDP-8/I simulator running, type <kbd>Ctrl-A d</kbd>.  You can
    re-attach to it later with a `pidp8i` command.

4.  To shut the simulator down while attached to its terminal interface,
    type <kbd>Ctrl-E</kbd> to pause the simulator, then at the `simh>`
    prompt type `quit`.  Type `help` at that prompt to get some idea of
    what else you can do with the simulator command language, or read
    the [SIMH Users' Guide][sdoc].

5.  To shut the simulator down from the Raspbian command line:

        $ sudo systemctl stop pidp8i

There are [other major differences][mdif] between the upstream
distribution and this one.  See that linked wiki article for details.


## License

Copyright © 2016-2017 by Warren Young. This document is licensed under
the terms of [the SIMH license][sl].


[prj]:  https://tangentsoft.com/pidp8i/
[upst]: http://obsolescence.wixsite.com/obsolescence/pidp-8
[smod]: http://obsolescence.wixsite.com/obsolescence/2016-pidp-8-building-instructions
[usd]:  http://obsolescence.wixsite.com/obsolescence/pidp-8-details
[dt2]:  https://github.com/VentureKing/Deeper-Thought-2
[sdoc]: http://simh.trailing-edge.com/pdf/simh_doc.pdf
[prj]:  http://obsolescence.wixsite.com/obsolescence/pidp-8
[test]: https://tangentsoft.com/pidp8i/doc/trunk/README-test.md
[thro]: https://tangentsoft.com/pidp8i/doc/trunk/README-throttle.md
[mdif]: https://tangentsoft.com/pidp8i/wiki?name=Major+Differences
[sl]:   https://tangentsoft.com/pidp8i/doc/trunk/SIMH-LICENSE.md

# PiDP-8/I Software Release Process

If you are just a user of this software, you need read no further. This
document is for those producing release versions of the software, or for
those curious about what goes into doing so.


## Update ChangeLog.md

Trawl the Fossil timeline for user-visible changes since the last
release, and write them up in user-focused form into the `ChangeLog.md`
file. If a regular user of the software cannot see a given change, it
shouldn't go in the `ChangeLog.md`; let it be documented via the
timeline only.


## Update the Release Branch

Run `make release` to check the `ChangeLog.md` file changes in, tagging
that checkin with a release version tag of the form vYYYYMMDD and merge
those changes into the `release` branch.

It runs entirely automatically unless an error occurs, in which case it
stops immediately, so check its output for errors before continuing.


## Update the Home Page Links

The zip and tarball links on the front page produce files named after
the date of the release. Those dates need to be updated immediately
after tagging the release, since they point at the "release" tag applied
by the previous step, so they begin shipping the new release immediately
after tagging it.


## Produce the Normal Binary OS Image

Start with the latest [Raspbian Lite OS image][os].

1.  If the version of the base OS has changed since the last binary OS
    image was created, download the new one and blast it onto an SD card
    used for no other purpose. Boot it up.

2.  After logging in, retreive and initialize the BOSI process:

        $ wget https://tangentsoft.com/bosi
		$ chmod +x bosi
		$ ./bosi init
 
    It will either reboot the system after completing its tasks
    successfully or exit early, giving the reason it failed.

3.  Test that the software starts up as it should.

4.  Reset the OS configuration:

        $ exec sudo ./bosi reset

    The `exec` bit is required so that the `bosi` invocation is run as
    root without any processes running as `pi` in case the `init` step
    sees that user `pi` hasn't been changed to `pidp8i` here: the
    `usermod` command we give to make that change will refuse to do what
    we ask if there are any running processes owned by user `pi`.

5.  Move the SD card to a USB reader plugged into the Pi, boot the Pi
    from its prior SD card, and shrink the OS image:

        $ wget https://tangentsoft.com/bosi
		$ chmod +x bosi
        $ ./bosi shrink

6.  Move the USB reader to the Mac,¹ then say:

        $ bosi image[-nls] BLOCKS

    `BLOCKS` is the value output at the end of the `shrink` step.

7.  The prior step rewrote the SD card with the image file it created.
    Boot it up and make sure it still works.  If you're happy with it:

        $ bosi finish[-nls]

8.  While the OS image uploads, compose the announcement message, and
    modify the front page to point to the new images.  Post the
    announcement message and new front page once the uploads complete.

[os]: https://www.raspberrypi.org/downloads/raspbian/


## Produce the "No Lamp Simulator" Binary OS Image

Log into the SD card from which you made the regular image above, then
say `./bosi init --no-lamp-simulator`, and continue from step 3 above.

When you get down to the `image` and `test` steps, give `image-nls` and
`test-nls` instead.


----------------------

### Footnotes

1.  The image production steps could just as well be done on a Linux box
    or on a Windows box via Cygwin or WSL, but the commands in that
    final stage change due to OS differences.  Since this document
    exists primarily for use by the one who uses it, there is little
    point in having alternatives for other desktop OSes above.  Should
    someone else take over maintainership, they can translate the above
    commands for their own desktop PC.


### License

Copyright © 2016-2017 by Warren Young. This document is licensed under
the terms of [the SIMH license][sl].

[sl]: https://tangentsoft.com/pidp8i/doc/trunk/SIMH-LICENSE.md

Copyright © 2015-2017 by various authors

Permission is hereby granted, free of charge, to any person obtaining a
copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:

The above copyright notice and this permission notice shall be included
in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS LISTED ABOVE BE LIABLE FOR ANY CLAIM,
DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR
THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Except as contained in this notice, the names of the authors above shall
not be used in advertising or otherwise to promote the sale, use or
other dealings in this Software without prior written authorization from
those authors.

########################################################################
# auto.def - Configure file for the PiDP-8/I software build system,
#            based on autosetup.
#
# Copyright © 2016-2017 Warren Young
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
# IN NO EVENT SHALL THE AUTHORS LISTED ABOVE BE LIABLE FOR ANY CLAIM,
# DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT
# OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
# OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
#
# Except as contained in this notice, the names of the authors above
# shall not be used in advertising or otherwise to promote the sale,
# use or other dealings in this Software without prior written
# authorization from those authors.
########################################################################

define defaultprefix /opt/pidp8i

use cc
use cc-lib

options {
	debug-mode =>        "create a debug build (default is release)"
	no-lamp-simulator => "use simple LED driver instead of incandescent lamp simulator"
	serial-mod =>        "build simulator to expect the PCB serial mods"
	throttle:  =>        "override the throttle values in the boot scripts"
}

if {[opt-bool serial-mod]} {
	msg-result "The simulator will expect the serial mods to the Pi and PiDP-8/I PCBs."
	define PCB_SERIAL_MOD
}

if {[opt-bool debug-mode]} {
	msg-result "Creating a debuggable build."
	define BUILDMODE {-O0 -g}
} else {
	msg-result "Creating a release build."
	define BUILDMODE {-O2}
}

# Swap the incandescent lamp simulator feature out for the original LED
# driving method on single-core hosts.  The user can force this on
# multi-core hosts via --no-lamp-simulator.
set cores [exec tools/corecount]
if {($cores < 2) || [opt-bool no-lamp-simulator]} {
	msg-result "Driving PiDP-8/I front panel LEDs using low-CPU-usage method."
	define LED_DRIVER_MODULE n
	define ILS_MODE 0
} else {
	msg-result "Driving PiDP-8/I front panel LEDs using incandescent lamp simulator."
	define LED_DRIVER_MODULE i
	define ILS_MODE 1
}

# Translate --throttle value to a SIMH command
set tv [opt-val throttle]
set tvsl [string length $tv]
if {($tvsl == 0 && $cores > 1) || $tv == "none"} {
	define SET_THROTTLE {set nothrottle}
	set tv "unlimited"
} else {
	# Rewrite symbolic values with values SIMH can understand.  See
	# README-throttle.md for the justification of these values.
	if {$tv == "single-core" || $tvsl == 0} {
		# value for ~75% CPU usage on a Pi Model B+ with the simple
        # LED driving code; conveniently, 4x the speed of a real PDP-8/I
		set tv "1332k"
	} elseif {$tv == "pdp8e"} {
		set tv "416k"
	} elseif {$tv == "pdp8i" || $tv == "pdp8a"} {
		set tv "333k"
	} elseif {$tv == "pdp8l" || $tv == "pdp8"} {
		set tv "313k"
	} elseif {$tv == "ha6120"} {
		set tv "182k"
	} elseif {$tv == "im6100a"} {
		set tv "200k"
	} elseif {$tv == "im6100"} {
		set tv "100k"
	} elseif {$tv == "im6100c"} {
		set tv "83k"
	} elseif {$tv == "pdp8s"} {
		set tv "63k"
	} elseif {$tv == "human"} {
		set tv "10/1000"
	} elseif {$tv == "trace"} {
		set tv "1/1000"
	}
	# else, assume --throttle was given a legal SIMH throttle value

	define SET_THROTTLE "set throttle $tv"
}
msg-result "Simulator CPU throttle set to $tv IPS"

# Compiler and header checks
cc-check-includes time.h
cc-check-functions clock_nanosleep nanosleep usleep
cc-check-functions sched_yield

# We need to find an install(1) type program that supports -D.  The
# Raspberry Pi OSes typically used with the PiDB-8/I board do have this,
# but this package also runs on non-Linux OSes (e.g. for testing on a
# desktop Mac) so make sure we've got a suitable implementation.  The
# ginstall name is typical on non-Linux systems where GNU Coreutils was
# installed alongside the core OS utilities.
if {[cc-check-progs ginstall]} {
	define INSTALL ginstall
} elseif {[cc-check-progs install]} {
	if {[catch {exec install -D -d . >& /dev/null} result] == 0} {
		define INSTALL install
	} else {
		user-error "install(1) does not support -D; install GNU Coreutils."
	}
} else {
	user-error "No install(1) type program found; install GNU Coreutils."
}
msg-result "Found GNU install(1) program as [get-define INSTALL]."

# Also find GNU readlink in the same way
if {[cc-check-progs greadlink]} {
	set rlprg greadlink
} elseif {[cc-check-progs readlink]} {
	if {[catch {exec readlink -f . >& /dev/null} result] == 0} {
		set rlprg readlink
	} else {
		user-error "readlink(1) does not support -D; install GNU Coreutils."
	}
} else {
	user-error "No readlink(1) type program found; install GNU Coreutils."
}
msg-result "Found GNU readlink(1) as $rlprg."

# Canonicalize some paths which may be relative and generate others from them
define ABSPREFIX [exec $rlprg -f [get-define prefix]]
define BOOTDIR  "[get-define ABSPREFIX]/share/boot"
define MEDIADIR "[get-define ABSPREFIX]/share/media"

# Remember the name and primary group of the user who installed this, since
# we want to give that group write privileges to some files when they're
# installed, and we want them to own the screen(1) session.
set instgrp [exec id -grn]
set instusr [exec id -urn]
if {$instusr == "root"} {
	msg-result "Error: This software will not install and run as root."
	user-error "Reconfigure without sudo!"
}
define INSTGRP $instgrp
define INSTUSR $instusr
msg-result "Install group for user-writeable files will be $instgrp."
msg-result "Owner of screen(1) session will be $instusr."

# Can we use any nonstandard flags here?  We don't bother including
# flags that both GCC and Clang support.  The ones inside the "if"
# block are those that Clang will accept in an autosetup test but
# then will yell about if you try to use them.  The test checks for
# an -f sub-option that Clang doesn't currently support even enough
# to fool autosetup.
cc-check-standards c99
if {![opt-bool debug-mode]} {
	cc-check-flags -fipa-cp-clone
	cc-check-flags -fno-strict-overflow
	cc-check-flags -fpredictive-commoning
	if ([get-define HAVE_CFLAG_FIPA_CP_CLONE]) {
		cc-check-flags -fgcse-after-reload
		cc-check-flags -finline-functions
		cc-check-flags -fno-unsafe-loop-optimizations
		define-append CFLAGS "-D_GNU_SOURCE"
	}
}

# Embed this software's Fossil-based version string into gpio-common.c
define VERSION [exec "[get-define srcdir]/tools/version"]

# Write outputs.
#
# NOTE: If you change the list of files here, change INFILES in
# Makefile.in, too.
make-config-header src/config.h \
	-auto {ENABLE_* HAVE_* PACKAGE_* SIZEOF_*} \
	-bare {ILS_MODE PCB_SERIAL_MOD}
make-template Makefile.in
make-template bin/pidp8i.in
make-template boot/0.script.in
make-template boot/2.script.in
make-template boot/3.script.in
make-template boot/4.script.in
make-template boot/6.script.in
make-template boot/7.script.in
make-template etc/pidp8i-init.in
make-template etc/sudoers.in
make-template examples/Makefile.in
make-template src/Makefile.in
make-template src/gpio-common.c.in
make-template src/PDP8/Makefile.in
make-template src/PDP8/pidp8i.c.in
make-template src/scp.c.in
make-template tools/simh-update.in
exec chmod +x "[get-define builddir]/tools/simh-update"

Unless explicitly stated, all files which form part of autosetup
are released under the following license:

---------------------------------------------------------------------
autosetup - A build environment "autoconfigurator"

Copyright (c) 2010-2011, WorkWare Systems <http://workware.net.au/>

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:

1. Redistributions of source code must retain the above copyright
   notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above
   copyright notice, this list of conditions and the following
   disclaimer in the documentation and/or other materials
   provided with the distribution.

THIS SOFTWARE IS PROVIDED BY THE WORKWARE SYSTEMS ``AS IS'' AND ANY
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL WORKWARE
SYSTEMS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

The views and conclusions contained in the software and documentation
are those of the authors and should not be interpreted as representing
official policies, either expressed or implied, of WorkWare Systems.

This is autosetup v0.6.6. See http://msteveb.github.com/autosetup/

#!/bin/sh
# Copyright (c) 2006-2011 WorkWare Systems http://www.workware.net.au/
# All rights reserved
# vim:se syntax=tcl:
# \
dir=`dirname "$0"`; exec "`$dir/find-tclsh`" "$0" "$@"

set autosetup(version) 0.6.6

# Can be set to 1 to debug early-init problems
set autosetup(debug) 0

##################################################################
#
# Main flow of control, option handling
#
proc main {argv} {
	global autosetup define

	# There are 3 potential directories involved:
	# 1. The directory containing autosetup (this script)
	# 2. The directory containing auto.def
	# 3. The current directory

	# From this we need to determine:
	# a. The path to this script (and related support files)
	# b. The path to auto.def
	# c. The build directory, where output files are created

	# This is also complicated by the fact that autosetup may
	# have been run via the configure wrapper ([getenv WRAPPER] is set)

	# Here are the rules.
	# a. This script is $::argv0
	#    => dir, prog, exe, libdir
	# b. auto.def is in the directory containing the configure wrapper,
	#    otherwise it is in the current directory.
	#    => srcdir, autodef
	# c. The build directory is the current directory
	#    => builddir, [pwd]

	# 'misc' is needed before we can do anything, so set a temporary libdir
	# in case this is the development version
	set autosetup(libdir) [file dirname $::argv0]/lib
	use misc

	# (a)
	set autosetup(dir) [realdir [file dirname [realpath $::argv0]]]
	set autosetup(prog) [file join $autosetup(dir) [file tail $::argv0]]
	set autosetup(exe) [getenv WRAPPER $autosetup(prog)]
	if {$autosetup(installed)} {
		set autosetup(libdir) $autosetup(dir)
	} else {
		set autosetup(libdir) [file join $autosetup(dir) lib]
	}
	autosetup_add_dep $autosetup(prog)

	# (b)
	if {[getenv WRAPPER ""] eq ""} {
		# Invoked directly
		set autosetup(srcdir) [pwd]
	} else {
		# Invoked via the configure wrapper
		set autosetup(srcdir) [file-normalize [file dirname $autosetup(exe)]]
	}
	set autosetup(autodef) [relative-path $autosetup(srcdir)/auto.def]

	# (c)
	set autosetup(builddir) [pwd]

	set autosetup(argv) $argv
	set autosetup(cmdline) {}
	# options is a list of known options
	set autosetup(options) {}
	# optset is a dictionary of option values set by the user based on getopt
	set autosetup(optset) {}
	# optdefault is a dictionary of default values for options
	set autosetup(optdefault) {}
	set autosetup(optionhelp) {}
	set autosetup(showhelp) 0

	# Parse options
	use getopt

	# At the is point we don't know what is a valid option
	# We simply parse anything that looks like an option
	set autosetup(getopt) [getopt argv]

	#"=Core Options:"
	options-add {
		help:=local  => "display help and options. Optionally specify a module name, such as --help=system"
		version      => "display the version of autosetup"
		ref:=text manual:=text
		reference:=text => "display the autosetup command reference. 'text', 'wiki', 'asciidoc' or 'markdown'"
		debug        => "display debugging output as autosetup runs"
		install:=.   => "install autosetup to the current or given directory (in the 'autosetup/' subdirectory)"
		force init:=help   => "create initial auto.def, etc.  Use --init=help for known types"
		# Undocumented options
		option-checking=1
		nopager
		quiet
		timing
		conf:
	}

	if {[opt-bool version]} {
		puts $autosetup(version)
		exit 0
	}

	# autosetup --conf=alternate-auto.def
	if {[opt-val conf] ne ""} {
		set autosetup(autodef) [opt-val conf]
	}

	# Debugging output (set this early)
	incr autosetup(debug) [opt-bool debug]
	incr autosetup(force) [opt-bool force]
	incr autosetup(msg-quiet) [opt-bool quiet]
	incr autosetup(msg-timing) [opt-bool timing]

	# If the local module exists, source it now to allow for
	# project-local customisations
	if {[file exists $autosetup(libdir)/local.tcl]} {
		use local
	}

	# Now any auto-load modules
	foreach file [glob -nocomplain $autosetup(libdir)/*.auto $autosetup(libdir)/*/*.auto] {
		automf_load source $file
	}

	if {[opt-val help] ne ""} {
		incr autosetup(showhelp)
		use help
		autosetup_help [opt-val help]
	}

	if {[opt-val {manual ref reference}] ne ""} {
		use help
		autosetup_reference [opt-val {manual ref reference}]
	}

	# Allow combining --install and --init
	set earlyexit 0
	if {[opt-val install] ne ""} {
		use install
		autosetup_install [opt-val install]
		incr earlyexit
	}

	if {[opt-val init] ne ""} {
		use init
		autosetup_init [opt-val init]
		incr earlyexit
	}

	if {$earlyexit} {
		exit 0
	}

	if {![file exists $autosetup(autodef)]} {
		# Check for invalid option first
		options {}
		user-error "No auto.def found in \"$autosetup(srcdir)\" (use [file tail $::autosetup(exe)] --init to create one)"
	}

	# Parse extra arguments into autosetup(cmdline)
	foreach arg $argv {
		if {[regexp {([^=]*)=(.*)} $arg -> n v]} {
			dict set autosetup(cmdline) $n $v
			define $n $v
		} else {
			user-error "Unexpected parameter: $arg"
		}
	}

	autosetup_add_dep $autosetup(autodef)

	define CONFIGURE_OPTS ""
	foreach arg $autosetup(argv) {
		define-append CONFIGURE_OPTS [quote-if-needed $arg]
	}
	define AUTOREMAKE [file-normalize $autosetup(exe)]
	define-append AUTOREMAKE [get-define CONFIGURE_OPTS]


	# Log how we were invoked
	configlog "Invoked as: [getenv WRAPPER $::argv0] [quote-argv $autosetup(argv)]"

	# Note that auto.def is *not* loaded in the global scope
	source $autosetup(autodef)

	# Could warn here if options {} was not specified

	show-notices

	if {$autosetup(debug)} {
		msg-result "Writing all defines to config.log"
		configlog "================ defines ======================"
		foreach n [lsort [array names define]] {
			configlog "define $n $define($n)"
		}
	}

	exit 0
}

# @opt-bool ?-nodefault? option ...
#
# Check each of the named, boolean options and if any have been explicitly enabled
# or disabled by the user, return 1 or 0 accordingly.
#
# If the option was specified more than once, the last value wins.
# e.g. With --enable-foo --disable-foo, [opt-bool foo] will return 0
#
# If no value was specified by the user, returns the default value for the
# first option. If -nodefault is given, this behaviour changes and
# -1 is returned instead.
#
proc opt-bool {args} {
	set nodefault 0
	if {[lindex $args 0] eq "-nodefault"} {
		set nodefault 1
		set args [lrange $args 1 end]
	}
	option-check-names {*}$args

	foreach opt $args {
		if {[dict exists $::autosetup(optset) $opt]} {
			return [dict get $::autosetup(optset) $opt]
		}
	}

	if {$nodefault} {
		return -1
	}
	# Default value is the default for the first option
	return [dict get $::autosetup(optdefault) [lindex $args 0]]
}

# @opt-val option-list ?default=""?
#
# Returns a list containing all the values given for the non-boolean options in 'option-list'.
# There will be one entry in the list for each option given by the user, including if the
# same option was used multiple times.
# If only a single value is required, use something like:
#
## lindex [opt-val $names] end
#
# If no options were set, $default is returned (exactly, not as a list).
#
proc opt-val {names {default ""}} {
	option-check-names {*}$names

	foreach opt $names {
		if {[dict exists $::autosetup(optset) $opt]} {
			lappend result {*}[dict get $::autosetup(optset) $opt]
		}
	}
	if {[info exists result]} {
		return $result
	}
	return $default
}

proc option-check-names {args} {
	foreach o $args {
		if {$o ni $::autosetup(options)} {
			autosetup-error "Request for undeclared option --$o"
		}
	}
}

# Parse the option definition in $opts and update
# ::autosetup(setoptions) and ::autosetup(optionhelp) appropriately
#
proc options-add {opts {header ""}} {
	global autosetup

	# First weed out comment lines
	set realopts {}
	foreach line [split $opts \n] {
		if {![string match "#*" [string trimleft $line]]} {
			append realopts $line \n
		}
	}
	set opts $realopts

	for {set i 0} {$i < [llength $opts]} {incr i} {
		set opt [lindex $opts $i]
		if {[string match =* $opt]} {
			# This is a special heading
			lappend autosetup(optionhelp) $opt ""
			set header {}
			continue
		}

		#puts "i=$i, opt=$opt"
		regexp {^([^:=]*)(:)?(=)?(.*)$} $opt -> name colon equal value
		if {$name in $autosetup(options)} {
			autosetup-error "Option $name already specified"
		}

		#puts "$opt => $name $colon $equal $value"

		# Find the corresponding value in the user options
		# and set the default if necessary
		if {[string match "-*" $opt]} {
			# This is a documentation-only option, like "-C <dir>"
			set opthelp $opt
		} elseif {$colon eq ""} {
			# Boolean option
			lappend autosetup(options) $name

			if {$value eq "1"} {
				set opthelp "--disable-$name"
			} else {
				set opthelp "--$name"
			}

			# Set the default
			if {$value eq ""} {
				set value 0
			}
			dict set autosetup(optdefault) $name $value

			if {[dict exists $autosetup(getopt) $name]} {
				# The option was specified by the user. Look at the last value.
				lassign [lindex [dict get $autosetup(getopt) $name] end] type setvalue
				if {$type eq "str"} {
					# Can we convert the value to a boolean?
					if {$setvalue in {1 enabled yes}} {
						set setvalue 1
					} elseif {$setvalue in {0 disabled no}} {
						set setvalue 0
					} else {
						user-error "Boolean option $name given as --$name=$setvalue"
					}
				}
				dict set autosetup(optset) $name $setvalue
				#puts "Found boolean option --$name=$setvalue"
			}
		} else {
			# String option.
			lappend autosetup(options) $name

			if {$equal eq "="} {
				# String option with optional value
				set opthelp "--$name?=$value?"
			} else {
				# String option with required value
				set opthelp "--$name=$value"
			}
			dict set autosetup(optdefault) $name $value

			# Get the values specified by the user
			if {[dict exists $autosetup(getopt) $name]} {
				set listvalue {}

				foreach pair [dict get $autosetup(getopt) $name] {
					lassign $pair type setvalue
					if {$type eq "bool" && $setvalue} {
						if {$equal ne "="} {
							user-error "Option --$name requires a value"
						}
						# If given as a boolean, use the default value
						set setvalue $value
					}
					lappend listvalue $setvalue
				}

				#puts "Found string option --$name=$listvalue"
				dict set autosetup(optset) $name $listvalue
			}
		}

		# Now create the help for this option if appropriate
		if {[lindex $opts $i+1] eq "=>"} {
			set desc [lindex $opts $i+2]
			#string match \n* $desc
			if {$header ne ""} {
				lappend autosetup(optionhelp) $header ""
				set header ""
			}
			# A multi-line description
			lappend autosetup(optionhelp) $opthelp $desc
			incr i 2
		}
	}
}

# @module-options optionlist
#
# Like 'options', but used within a module.
proc module-options {opts} {
	set header ""
	if {$::autosetup(showhelp) > 1 && [llength $opts]} {
		set header "Module Options:"
	}
	options-add $opts $header

	if {$::autosetup(showhelp)} {
		# Ensure that the module isn't executed on --help
		# We are running under eval or source, so use break
		# to prevent further execution
		#return -code break -level 2
		return -code break
	}
}

proc max {a b} {
	expr {$a > $b ? $a : $b}
}

proc options-wrap-desc {text length firstprefix nextprefix initial} {
	set len $initial
	set space $firstprefix
	foreach word [split $text] {
		set word [string trim $word]
		if {$word == ""} {
			continue
		}
		if {$len && [string length $space$word] + $len >= $length} {
			puts ""
			set len 0
			set space $nextprefix
		}
		incr len [string length $space$word]
		puts -nonewline $space$word
		set space " "
	}
	if {$len} {
		puts ""
	}
}

proc options-show {} {
	# Determine the max option width
	set max 0
	foreach {opt desc} $::autosetup(optionhelp) {
		if {[string match =* $opt] || [string match \n* $desc]} {
			continue
		}
		set max [max $max [string length $opt]]
	}
	set indent [string repeat " " [expr $max+4]]
	set cols [getenv COLUMNS 80]
	catch {
		lassign [exec stty size] rows cols
	}
	incr cols -1
	# Now output
	foreach {opt desc} $::autosetup(optionhelp) {
		if {[string match =* $opt]} {
			puts [string range $opt 1 end]
			continue
		}
		puts -nonewline "  [format %-${max}s $opt]"
		if {[string match \n* $desc]} {
			puts $desc
		} else {
			options-wrap-desc [string trim $desc] $cols "  " $indent [expr $max + 2]
		}
	}
}

# @options options-spec
#
# Specifies configuration-time options which may be selected by the user
# and checked with opt-val and opt-bool. The format of options-spec follows.
#
# A boolean option is of the form:
#
## name[=0|1]  => "Description of this boolean option"
#
# The default is name=0, meaning that the option is disabled by default.
# If name=1 is used to make the option enabled by default, the description should reflect
# that with text like "Disable support for ...".
#
# An argument option (one which takes a parameter) is of the form:
#
## name:[=]value  => "Description of this option"
#
# If the name:value form is used, the value must be provided with the option (as --name=myvalue).
# If the name:=value form is used, the value is optional and the given value is used as the default
# if it is not provided.
#
# Undocumented options are also supported by omitting the "=> description.
# These options are not displayed with --help and can be useful for internal options or as aliases.
#
# For example, --disable-lfs is an alias for --disable=largefile:
#
## lfs=1 largefile=1 => "Disable large file support"
#
proc options {optlist} {
	# Allow options as a list or args
	options-add $optlist "Local Options:"

	if {$::autosetup(showhelp)} {
		options-show
		exit 0
	}

	# Check for invalid options
	if {[opt-bool option-checking]} {
		foreach o [dict keys $::autosetup(getopt)] {
			if {$o ni $::autosetup(options)} {
				user-error "Unknown option --$o"
			}
		}
	}
}

proc config_guess {} {
	if {[file-isexec $::autosetup(dir)/config.guess]} {
		exec-with-stderr sh $::autosetup(dir)/config.guess
		if {[catch {exec-with-stderr sh $::autosetup(dir)/config.guess} alias]} {
			user-error $alias
		}
		return $alias
	} else {
		configlog "No config.guess, so using uname"
		string tolower [exec uname -p]-unknown-[exec uname -s][exec uname -r]
	}
}

proc config_sub {alias} {
	if {[file-isexec $::autosetup(dir)/config.sub]} {
		if {[catch {exec-with-stderr sh $::autosetup(dir)/config.sub $alias} alias]} {
			user-error $alias
		}
	}
	return $alias
}

# @define name ?value=1?
#
# Defines the named variable to the given value.
# These (name, value) pairs represent the results of the configuration check
# and are available to be checked, modified and substituted.
#
proc define {name {value 1}} {
	set ::define($name) $value
	#dputs "$name <= $value"
}

# @undefine name
#
# Undefine the named variable
#
proc undefine {name} {
	unset -nocomplain ::define($name)
	#dputs "$name <= <undef>"
}

# @define-append name value ...
#
# Appends the given value(s) to the given 'defined' variable.
# If the variable is not defined or empty, it is set to $value.
# Otherwise the value is appended, separated by a space.
# Any extra values are similarly appended.
# If any value is already contained in the variable (as a substring) it is omitted.
#
proc define-append {name args} {
	if {[get-define $name ""] ne ""} {
		# Make a token attempt to avoid duplicates
		foreach arg $args {
			if {[string first $arg $::define($name)] == -1} {
				append ::define($name) " " $arg
			}
		}
	} else {
		set ::define($name) [join $args]
	}
	#dputs "$name += [join $args] => $::define($name)"
}

# @get-define name ?default=0?
#
# Returns the current value of the 'defined' variable, or $default
# if not set.
#
proc get-define {name {default 0}} {
	if {[info exists ::define($name)]} {
		#dputs "$name => $::define($name)"
		return $::define($name)
	}
	#dputs "$name => $default"
	return $default
}

# @is-defined name
#
# Returns 1 if the given variable is defined.
#
proc is-defined {name} {
	info exists ::define($name)
}

# @all-defines
#
# Returns a dictionary (name value list) of all defined variables.
#
# This is suitable for use with 'dict', 'array set' or 'foreach'
# and allows for arbitrary processing of the defined variables.
#
proc all-defines {} {
	array get ::define
}


# @get-env name default
#
# If $name was specified on the command line, return it.
# If $name was set in the environment, return it.
# Otherwise return $default.
#
proc get-env {name default} {
	if {[dict exists $::autosetup(cmdline) $name]} {
		return [dict get $::autosetup(cmdline) $name]
	}
	getenv $name $default
}

# @env-is-set name
#
# Returns 1 if the $name was specified on the command line or in the environment.
# Note that an empty environment variable is not considered to be set.
#
proc env-is-set {name} {
	if {[dict exists $::autosetup(cmdline) $name]} {
		return 1
	}
	if {[getenv $name ""] ne ""} {
		return 1
	}
	return 0
}

# @readfile filename ?default=""?
#
# Return the contents of the file, without the trailing newline.
# If the file doesn't exist or can't be read, returns $default.
#
proc readfile {filename {default_value ""}} {
	set result $default_value
	catch {
		set f [open $filename]
		set result [read -nonewline $f]
		close $f
	}
	return $result
}

# @writefile filename value
#
# Creates the given file containing $value.
# Does not add an extra newline.
#
proc writefile {filename value} {
	set f [open $filename w]
	puts -nonewline $f $value
	close $f
}

proc quote-if-needed {str} {
	if {[string match {*[\" ]*} $str]} {
		return \"[string map [list \" \\" \\ \\\\] $str]\"
	}
	return $str
}

proc quote-argv {argv} {
	set args {}
	foreach arg $argv {
		lappend args [quote-if-needed $arg]
	}
	join $args
}

# @suffix suf list
#
# Takes a list and returns a new list with $suf appended
# to each element
#
## suffix .c {a b c} => {a.c b.c c.c}
#
proc suffix {suf list} {
	set result {}
	foreach p $list {
		lappend result $p$suf
	}
	return $result
}

# @prefix pre list
#
# Takes a list and returns a new list with $pre prepended
# to each element
#
## prefix jim- {a.c b.c} => {jim-a.c jim-b.c}
#
proc prefix {pre list} {
	set result {}
	foreach p $list {
		lappend result $pre$p
	}
	return $result
}

# @find-executable name
#
# Searches the path for an executable with the given name.
# Note that the name may include some parameters, e.g. "cc -mbig-endian",
# in which case the parameters are ignored.
# Returns 1 if found, or 0 if not.
#
proc find-executable {name} {
	# Ignore any parameters
	set name [lindex $name 0]
	if {$name eq ""} {
		# The empty string is never a valid executable
		return 0
	}
	foreach p [split-path] {
		dputs "Looking for $name in $p"
		set exec [file join $p $name]
		if {[file-isexec $exec]} {
			dputs "Found $name -> $exec"
			return 1
		}
	}
	return 0
}

# @find-an-executable ?-required? name ...
#
# Given a list of possible executable names,
# searches for one of these on the path.
#
# Returns the name found, or "" if none found.
# If the first parameter is '-required', an error is generated
# if no executable is found.
#
proc find-an-executable {args} {
	set required 0
	if {[lindex $args 0] eq "-required"} {
		set args [lrange $args 1 end]
		incr required
	}
	foreach name $args {
		if {[find-executable $name]} {
			return $name
		}
	}
	if {$required} {
		if {[llength $args] == 1} {
			user-error "failed to find: [join $args]"
		} else {
			user-error "failed to find one of: [join $args]"
		}
	}
	return ""
}

# @configlog msg
#
# Writes the given message to the configuration log, config.log
#
proc configlog {msg} {
	if {![info exists ::autosetup(logfh)]} {
		set ::autosetup(logfh) [open config.log w]
	}
	puts $::autosetup(logfh) $msg
}

# @msg-checking msg
#
# Writes the message with no newline to stdout.
#
proc msg-checking {msg} {
	if {$::autosetup(msg-quiet) == 0} {
		maybe-show-timestamp
		puts -nonewline $msg
		set ::autosetup(msg-checking) 1
	}
}

# @msg-result msg
#
# Writes the message to stdout.
#
proc msg-result {msg} {
	if {$::autosetup(msg-quiet) == 0} {
		maybe-show-timestamp
		puts $msg
		set ::autosetup(msg-checking) 0
		show-notices
	}
}

# @msg-quiet command ...
#
# msg-quiet evaluates it's arguments as a command with output
# from msg-checking and msg-result suppressed.
#
# This is useful if a check needs to run a subcheck which isn't
# of interest to the user.
proc msg-quiet {args} {
	incr ::autosetup(msg-quiet)
	set rc [uplevel 1 $args]
	incr ::autosetup(msg-quiet) -1
	return $rc
}

# Will be overridden by 'use misc'
proc error-stacktrace {msg} {
	return $msg
}

proc error-location {msg} {
	return $msg
}

##################################################################
#
# Debugging output
#
proc dputs {msg} {
	if {$::autosetup(debug)} {
		puts $msg
	}
}

##################################################################
#
# User and system warnings and errors
#
# Usage errors such as wrong command line options

# @user-error msg
#
# Indicate incorrect usage to the user, including if required components
# or features are not found.
# autosetup exits with a non-zero return code.
#
proc user-error {msg} {
	show-notices
	puts stderr "Error: $msg"
	puts stderr "Try: '[file tail $::autosetup(exe)] --help' for options"
	exit 1
}

# @user-notice msg
#
# Output the given message to stderr.
#
proc user-notice {msg} {
	lappend ::autosetup(notices) $msg
}

# Incorrect usage in the auto.def file. Identify the location.
proc autosetup-error {msg} {
	autosetup-full-error [error-location $msg]
}

# Like autosetup-error, except $msg is the full error message.
proc autosetup-full-error {msg} {
	show-notices
	puts stderr $msg
	exit 1
}

proc show-notices {} {
	if {$::autosetup(msg-checking)} {
		puts ""
		set ::autosetup(msg-checking) 0
	}
	flush stdout
	if {[info exists ::autosetup(notices)]} {
		puts stderr [join $::autosetup(notices) \n]
		unset ::autosetup(notices)
	}
}

proc maybe-show-timestamp {} {
	if {$::autosetup(msg-timing) && $::autosetup(msg-checking) == 0} {
		puts -nonewline [format {[%6.2f] } [expr {([clock millis] - $::autosetup(start)) % 10000 / 1000.0}]]
	}
}

proc autosetup_version {} {
	return "autosetup v$::autosetup(version)"
}

##################################################################
#
# Directory/path handling
#

proc realdir {dir} {
	set oldpwd [pwd]
	cd $dir
	set pwd [pwd]
	cd $oldpwd
	return $pwd
}

# Follow symlinks until we get to something which is not a symlink
proc realpath {path} {
	while {1} {
		if {[catch {
			set path [file readlink $path]
		}]} {
			# Not a link
			break
		}
	}
	return $path
}

# Convert absolute path, $path into a path relative
# to the given directory (or the current dir, if not given).
#
proc relative-path {path {pwd {}}} {
	set diff 0
	set same 0
	set newf {}
	set prefix {}
	set path [file-normalize $path]
	if {$pwd eq ""} {
		set pwd [pwd]
	} else {
		set pwd [file-normalize $pwd]
	}

	if {$path eq $pwd} {
		return .
	}

	# Try to make the filename relative to the current dir
	foreach p [split $pwd /] f [split $path /] {
		if {$p ne $f} {
			incr diff
		} elseif {!$diff} {
			incr same
		}
		if {$diff} {
			if {$p ne ""} {
				# Add .. for sibling or parent dir
				lappend prefix ..
			}
			if {$f ne ""} {
				lappend newf $f
			}
		}
	}
	if {$same == 1 || [llength $prefix] > 3} {
		return $path
	}

	file join [join $prefix /] [join $newf /]
}

# Add filename as a dependency to rerun autosetup
# The name will be normalised (converted to a full path)
#
proc autosetup_add_dep {filename} {
	lappend ::autosetup(deps) [file-normalize $filename]
}

##################################################################
#
# Library module support
#

# @use module ...
#
# Load the given library modules.
# e.g. 'use cc cc-shared'
#
# Note that module 'X' is implemented in either 'autosetup/X.tcl'
# or 'autosetup/X/init.tcl'
#
# The latter form is useful for a complex module which requires additional
# support file. In this form, '$::usedir' is set to the module directory
# when it is loaded.
#
proc use {args} {
	foreach m $args {
		if {[info exists ::libmodule($m)]} {
			continue
		}
		set ::libmodule($m) 1
		if {[info exists ::modsource($m)]} {
			automf_load eval $::modsource($m)
		} else {
			set sources [list $::autosetup(libdir)/${m}.tcl $::autosetup(libdir)/${m}/init.tcl]
			set found 0
			foreach source $sources {
				if {[file exists $source]} {
					incr found
					break
				}
			}
			if {$found} {
				# For the convenience of the "use" source, point to the directory
				# it is being loaded from
				set ::usedir [file dirname $source]
				automf_load source $source
				autosetup_add_dep $source
			} else {
				autosetup-error "use: No such module: $m"
			}
		}
	}
}

# Load module source in the global scope by executing the given command
proc automf_load {args} {
	if {[catch [list uplevel #0 $args] msg opts] ni {0 2 3}} {
		autosetup-full-error [error-dump $msg $opts $::autosetup(debug)]
	}
}

# Initial settings
set autosetup(exe) $::argv0
set autosetup(istcl) 1
set autosetup(start) [clock millis]
set autosetup(installed) 0
set autosetup(msg-checking) 0
set autosetup(msg-quiet) 0

# Embedded modules are inserted below here
set autosetup(installed) 1
# ----- module asciidoc-formatting -----

set modsource(asciidoc-formatting) {
# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module which provides text formatting
# asciidoc format

use formatting

proc para {text} {
    regsub -all "\[ \t\n\]+" [string trim $text] " "
}
proc title {text} {
    underline [para $text] =
    nl
}
proc p {text} {
    puts [para $text]
    nl
}
proc code {text} {
    foreach line [parse_code_block $text] {
        puts "    $line"
    }
    nl
}
proc codelines {lines} {
    foreach line $lines {
        puts "    $line"
    }
    nl
}
proc nl {} {
    puts ""
}
proc underline {text char} {
    regexp "^(\[ \t\]*)(.*)" $text -> indent words
    puts $text
    puts $indent[string repeat $char [string length $words]]
}
proc section {text} {
    underline "[para $text]" -
    nl
}
proc subsection {text} {
    underline "$text" ~
    nl
}
proc bullet {text} {
    puts "* [para $text]"
}
proc indent {text} {
    puts " :: "
    puts [para $text]
}
proc defn {first args} {
    set sep ""
    if {$first ne ""} {
        puts "${first}::"
    } else {
        puts " :: "
    }
    set defn [string trim [join $args \n]]
    regsub -all "\n\n" $defn "\n ::\n" defn
    puts $defn
}
}

# ----- module formatting -----

set modsource(formatting) {
# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module which provides common text formatting

# This is designed for documenation which looks like:
# code {...}
# or
# code {
#    ...
#    ...
# }
# In the second case, we need to work out the indenting
# and strip it from all lines but preserve the remaining indenting.
# Note that all lines need to be indented with the same initial
# spaces/tabs.
#
# Returns a list of lines with the indenting removed.
#
proc parse_code_block {text} {
    # If the text begins with newline, take the following text,
    # otherwise just return the original
    if {![regexp "^\n(.*)" $text -> text]} {
        return [list [string trim $text]]
    }

    # And trip spaces off the end
    set text [string trimright $text]

    set min 100
    # Examine each line to determine the minimum indent
    foreach line [split $text \n] {
        if {$line eq ""} {
            # Ignore empty lines for the indent calculation
            continue
        }
        regexp "^(\[ \t\]*)" $line -> indent
        set len [string length $indent]
        if {$len < $min} {
            set min $len
        }
    }

    # Now make a list of lines with this indent removed
    set lines {}
    foreach line [split $text \n] {
        lappend lines [string range $line $min end]
    }

    # Return the result
    return $lines
}
}

# ----- module getopt -----

set modsource(getopt) {
# Copyright (c) 2006 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Simple getopt module

# Parse everything out of the argv list which looks like an option
# Everything which doesn't look like an option, or is after --, is left unchanged
# Understands --enable-xxx and --with-xxx as synonyms for --xxx to enable the boolean option xxx.
# Understands --disable-xxx and --without-xxx to disable the boolean option xxx.
#
# The returned value is a dictionary keyed by option name
# Each value is a list of {type value} ... where type is "bool" or "str".
# The value for a boolean option is 0 or 1. The value of a string option is the value given.
proc getopt {argvname} {
	upvar $argvname argv
	set nargv {}

	set opts {}

	for {set i 0} {$i < [llength $argv]} {incr i} {
		set arg [lindex $argv $i]

		#dputs arg=$arg

		if {$arg eq "--"} {
			# End of options
			incr i
			lappend nargv {*}[lrange $argv $i end]
			break
		}

		if {[regexp {^--([^=][^=]+)=(.*)$} $arg -> name value]} {
			# --name=value
			dict lappend opts $name [list str $value]
		} elseif {[regexp {^--(enable-|disable-|with-|without-)?([^=]*)$} $arg -> prefix name]} {
			if {$prefix in {enable- with- ""}} {
				set value 1
			} else {
				set value 0
			}
			dict lappend opts $name [list bool $value]
		} else {
			lappend nargv $arg
		}
	}

	#puts "getopt: argv=[join $argv] => [join $nargv]"
	#array set getopt $opts
	#parray getopt

	set argv $nargv

	return $opts
}
}

# ----- module help -----

set modsource(help) {
# Copyright (c) 2010 WorkWare Systems http://workware.net.au/
# All rights reserved

# Module which provides usage, help and the command reference

proc autosetup_help {what} {
    use_pager

    puts "Usage: [file tail $::autosetup(exe)] \[options\] \[settings\]\n"
    puts "This is [autosetup_version], a build environment \"autoconfigurator\""
    puts "See the documentation online at http://msteveb.github.com/autosetup/\n"

    if {$what eq "local"} {
        if {[file exists $::autosetup(autodef)]} {
            # This relies on auto.def having a call to 'options'
            # which will display options and quit
            source $::autosetup(autodef)
        } else {
            options-show
        }
    } else {
        incr ::autosetup(showhelp)
        if {[catch {use $what}]} {
            user-error "Unknown module: $what"
        } else {
            options-show
        }
    }
    exit 0
}

# If not already paged and stdout is a tty, pipe the output through the pager
# This is done by reinvoking autosetup with --nopager added
proc use_pager {} {
    if {![opt-bool nopager] && [getenv PAGER ""] ne "" && [isatty? stdin] && [isatty? stdout]} {
        if {[catch {
            exec [info nameofexecutable] $::argv0 --nopager {*}$::argv |& {*}[getenv PAGER] >@stdout <@stdin 2>@stderr
        } msg opts] == 1} {
            if {[dict get $opts -errorcode] eq "NONE"} {
                # an internal/exec error
                puts stderr $msg
                exit 1
            }
        }
        exit 0
    }
}

# Outputs the autosetup references in one of several formats
proc autosetup_reference {{type text}} {

    use_pager

    switch -glob -- $type {
        wiki {use wiki-formatting}
        ascii* {use asciidoc-formatting}
        md - markdown {use markdown-formatting}
        default {use text-formatting}
    }

    title "[autosetup_version] -- Command Reference"

    section {Introduction}

    p {
        See http://msteveb.github.com/autosetup/ for the online documentation for 'autosetup'
    }

    p {
        'autosetup' provides a number of built-in commands which
        are documented below. These may be used from 'auto.def' to test
        for features, define variables, create files from templates and
        other similar actions.
    }

    automf_command_reference

    exit 0
}

proc autosetup_output_block {type lines} {
    if {[llength $lines]} {
        switch $type {
            code {
                codelines $lines
            }
            p {
                p [join $lines]
            }
            list {
                foreach line $lines {
                    bullet $line
                }
                nl
            }
        }
    }
}

# Generate a command reference from inline documentation
proc automf_command_reference {} {
    lappend files $::autosetup(prog)
    lappend files {*}[lsort [glob -nocomplain $::autosetup(libdir)/*.tcl]]

    section "Core Commands"
    set type p
    set lines {}
    set cmd {}

    foreach file $files {
        set f [open $file]
        while {![eof $f]} {
            set line [gets $f]

            # Find lines starting with "# @*" and continuing through the remaining comment lines
            if {![regexp {^# @(.*)} $line -> cmd]} {
                continue
            }

            # Synopsis or command?
            if {$cmd eq "synopsis:"} {
                section "Module: [file rootname [file tail $file]]"
            } else {
                subsection $cmd
            }

            set lines {}
            set type p

            # Now the description
            while {![eof $f]} {
                set line [gets $f]

                if {![regexp {^#(#)? ?(.*)} $line -> hash cmd]} {
                    break
                }
                if {$hash eq "#"} {
                    set t code
                } elseif {[regexp {^- (.*)} $cmd -> cmd]} {
                    set t list
                } else {
                    set t p
                }

                #puts "hash=$hash, oldhash=$oldhash, lines=[llength $lines], cmd=$cmd"

                if {$t ne $type || $cmd eq ""} {
                    # Finish the current block
                    autosetup_output_block $type $lines
                    set lines {}
                    set type $t
                }
                if {$cmd ne ""} {
                    lappend lines $cmd
                }
            }

            autosetup_output_block $type $lines
        }
        close $f
    }
}
}

# ----- module init -----

set modsource(init) {
# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module to help create auto.def and configure

proc autosetup_init {type} {
	set help 0
	if {$type in {? help}} {
		incr help
	} elseif {![dict exists $::autosetup(inittypes) $type]} {
		puts "Unknown type, --init=$type"
		incr help
	}
	if {$help} {
		puts "Use one of the following types (e.g. --init=make)\n"
		foreach type [lsort [dict keys $::autosetup(inittypes)]] {
			lassign [dict get $::autosetup(inittypes) $type] desc
			# XXX: Use the options-show code to wrap the description
			puts [format "%-10s %s" $type $desc]
		}
		return
	}
	lassign [dict get $::autosetup(inittypes) $type] desc script

	puts "Initialising $type: $desc\n"

	# All initialisations happens in the top level srcdir
	cd $::autosetup(srcdir)

	uplevel #0 $script
}

proc autosetup_add_init_type {type desc script} {
	dict set ::autosetup(inittypes) $type [list $desc $script]
}

# This is for in creating build-system init scripts
#
# If the file doesn't exist, create it containing $contents
# If the file does exist, only overwrite if --force is specified.
#
proc autosetup_check_create {filename contents} {
	if {[file exists $filename]} {
		if {!$::autosetup(force)} {
			puts "I see $filename already exists."
			return
		} else {
			puts "I will overwrite the existing $filename because you used --force."
		}
	} else {
		puts "I don't see $filename, so I will create it."
	}
	writefile $filename $contents
}
}

# ----- module install -----

set modsource(install) {
# Copyright (c) 2006-2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module which can install autosetup

proc autosetup_install {dir} {
	if {[catch {
		cd $dir
		file mkdir autosetup

		set f [open autosetup/autosetup w]

		set publicmodules [glob $::autosetup(libdir)/*.auto]

		# First the main script, but only up until "CUT HERE"
		set in [open $::autosetup(dir)/autosetup]
		while {[gets $in buf] >= 0} {
			if {$buf ne "##-- CUT HERE --##"} {
				puts $f $buf
				continue
			}

			# Insert the static modules here
			# i.e. those which don't contain @synopsis:
			puts $f "set autosetup(installed) 1"
			foreach file [lsort [glob $::autosetup(libdir)/*.tcl]] {
				set buf [readfile $file]
				if {[string match "*\n# @synopsis:*" $buf]} {
					lappend publicmodules $file
					continue
				}
				set modname [file rootname [file tail $file]]
				puts $f "# ----- module $modname -----"
				puts $f "\nset modsource($modname) \{"
				puts $f $buf
				puts $f "\}\n"
			}
		}
		close $in
		close $f
		exec chmod 755 autosetup/autosetup

		# Install public modules
		foreach file $publicmodules {
			autosetup_install_file $file autosetup
		}

		# Install support files
		foreach file {config.guess config.sub jimsh0.c find-tclsh test-tclsh LICENSE} {
			autosetup_install_file $::autosetup(dir)/$file autosetup
		}
		exec chmod 755 autosetup/config.sub autosetup/config.guess autosetup/find-tclsh

		writefile autosetup/README.autosetup \
			"This is [autosetup_version]. See http://msteveb.github.com/autosetup/\n"

	} error]} {
		user-error "Failed to install autosetup: $error"
	}
	puts "Installed [autosetup_version] to autosetup/"

	# Now create 'configure' if necessary
	autosetup_create_configure
}

proc autosetup_create_configure {} {
	if {[file exists configure]} {
		if {!$::autosetup(force)} {
			# Could this be an autosetup configure?
			if {![string match "*\nWRAPPER=*" [readfile configure]]} {
				puts "I see configure, but not created by autosetup, so I won't overwrite it."
				puts "Remove it or use --force to overwrite."
				return
			}
		} else {
			puts "I will overwrite the existing configure because you used --force."
		}
	} else {
		puts "I don't see configure, so I will create it."
	}
	writefile configure \
{#!/bin/sh
dir="`dirname "$0"`/autosetup"
WRAPPER="$0"; export WRAPPER; exec "`$dir/find-tclsh`" "$dir/autosetup" "$@"
}
	catch {exec chmod 755 configure}
}

# Append the contents of $file to filehandle $f
proc autosetup_install_append {f file} {
	set in [open $file]
	puts $f [read $in]
	close $in
}

proc autosetup_install_file {file dir} {
	if {![file exists $file]} {
		error "Missing installation file '$file'"
	}
	writefile [file join $dir [file tail $file]] [readfile $file]\n
}

if {$::autosetup(installed)} {
	user-error "autosetup can only be installed from development source, not from installed copy"
}
}

# ----- module markdown-formatting -----

set modsource(markdown-formatting) {
# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module which provides text formatting
# markdown format (kramdown syntax)

use formatting

proc para {text} {
    regsub -all "\[ \t\n\]+" [string trim $text] " " text
    regsub -all {([^a-zA-Z])'([^']*)'} $text {\1**`\2`**} text
    regsub -all {^'([^']*)'} $text {**`\1`**} text
    regsub -all {(http[^ \t\n]*)} $text {[\1](\1)} text
    return $text
}
proc title {text} {
    underline [para $text] =
    nl
}
proc p {text} {
    puts [para $text]
    nl
}
proc codelines {lines} {
    puts "~~~~~~~~~~~~"
    foreach line $lines {
        puts $line
    }
    puts "~~~~~~~~~~~~"
    nl
}
proc code {text} {
    puts "~~~~~~~~~~~~"
    foreach line [parse_code_block $text] {
        puts $line
    }
    puts "~~~~~~~~~~~~"
    nl
}
proc nl {} {
    puts ""
}
proc underline {text char} {
    regexp "^(\[ \t\]*)(.*)" $text -> indent words
    puts $text
    puts $indent[string repeat $char [string length $words]]
}
proc section {text} {
    underline "[para $text]" -
    nl
}
proc subsection {text} {
    puts "### `$text`"
    nl
}
proc bullet {text} {
    puts "* [para $text]"
}
proc defn {first args} {
    puts "^"
    set defn [string trim [join $args \n]]
    if {$first ne ""} {
        puts "**${first}**"
        puts -nonewline ": "
        regsub -all "\n\n" $defn "\n: " defn
    }
    puts "$defn"
}
}

# ----- module misc -----

set modsource(misc) {
# Copyright (c) 2007-2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module containing misc procs useful to modules
# Largely for platform compatibility

set autosetup(istcl) [info exists ::tcl_library]
set autosetup(iswin) [string equal windows $tcl_platform(platform)]

if {$autosetup(iswin)} {
	# mingw/windows separates $PATH with semicolons
	# and doesn't have an executable bit
	proc split-path {} {
		split [getenv PATH .] {;}
	}
	proc file-isexec {exec} {
		# Basic test for windows. We ignore .bat
		if {[file isfile $exec] || [file isfile $exec.exe]} {
			return 1
		}
		return 0
	}
} else {
	# unix separates $PATH with colons and has and executable bit
	proc split-path {} {
		split [getenv PATH .] :
	}
	proc file-isexec {exec} {
		file executable $exec
	}
}

# Assume that exec can return stdout and stderr
proc exec-with-stderr {args} {
	exec {*}$args 2>@1
}

if {$autosetup(istcl)} {
	# Tcl doesn't have the env command
	proc getenv {name args} {
		if {[info exists ::env($name)]} {
			return $::env($name)
		}
		if {[llength $args]} {
			return [lindex $args 0]
		}
		return -code error "environment variable \"$name\" does not exist"
	}
	proc isatty? {channel} {
		dict exists [fconfigure $channel] -xchar
	}
} else {
	if {$autosetup(iswin)} {
		# On Windows, backslash convert all environment variables
		# (Assume that Tcl does this for us)
		proc getenv {name args} {
			string map {\\ /} [env $name {*}$args]
		}
	} else {
		# Jim on unix is simple
		alias getenv env
	}
	proc isatty? {channel} {
		set tty 0
		catch {
			# isatty is a recent addition to Jim Tcl
			set tty [$channel isatty]
		}
		return $tty
	}
}

# In case 'file normalize' doesn't exist
#
proc file-normalize {path} {
	if {[catch {file normalize $path} result]} {
		if {$path eq ""} {
			return ""
		}
		set oldpwd [pwd]
		if {[file isdir $path]} {
			cd $path
			set result [pwd]
		} else {
			cd [file dirname $path]
			set result [file join [pwd] [file tail $path]]
		}
		cd $oldpwd
	}
	return $result
}

# If everything is working properly, the only errors which occur
# should be generated in user code (e.g. auto.def).
# By default, we only want to show the error location in user code.
# We use [info frame] to achieve this, but it works differently on Tcl and Jim.
#
# This is designed to be called for incorrect usage in auto.def, via autosetup-error
#
proc error-location {msg} {
	if {$::autosetup(debug)} {
		return -code error $msg
	}
	# Search back through the stack trace for the first error in a .def file
	for {set i 1} {$i < [info level]} {incr i} {
		if {$::autosetup(istcl)} {
			array set info [info frame -$i]
		} else {
			lassign [info frame -$i] info(caller) info(file) info(line)
		}
		if {[string match *.def $info(file)]} {
			return "[relative-path $info(file)]:$info(line): Error: $msg"
		}
		#puts "Skipping $info(file):$info(line)"
	}
	return $msg
}

# If everything is working properly, the only errors which occur
# should be generated in user code (e.g. auto.def).
# By default, we only want to show the error location in user code.
# We use [info frame] to achieve this, but it works differently on Tcl and Jim.
#
# This is designed to be called for incorrect usage in auto.def, via autosetup-error
#
proc error-stacktrace {msg} {
	if {$::autosetup(debug)} {
		return -code error $msg
	}
	# Search back through the stack trace for the first error in a .def file
	for {set i 1} {$i < [info level]} {incr i} {
		if {$::autosetup(istcl)} {
			array set info [info frame -$i]
		} else {
			lassign [info frame -$i] info(caller) info(file) info(line)
		}
		if {[string match *.def $info(file)]} {
			return "[relative-path $info(file)]:$info(line): Error: $msg"
		}
		#puts "Skipping $info(file):$info(line)"
	}
	return $msg
}

# Given the return from [catch {...} msg opts], returns an appropriate
# error message. A nice one for Jim and a less-nice one for Tcl.
# If 'fulltrace' is set, a full stack trace is provided.
# Otherwise a simple message is provided.
#
# This is designed for developer errors, e.g. in module code or auto.def code
#
#
proc error-dump {msg opts fulltrace} {
	if {$::autosetup(istcl)} {
		if {$fulltrace} {
			return "Error: [dict get $opts -errorinfo]"
		} else {
			return "Error: $msg"
		}
	} else {
		lassign $opts(-errorinfo) p f l
		if {$f ne ""} {
			set result "$f:$l: Error: "
		}
		append result "$msg\n"
		if {$fulltrace} {
			append result [stackdump $opts(-errorinfo)]
		}

		# Remove the trailing newline
		string trim $result
	}
}
}

# ----- module text-formatting -----

set modsource(text-formatting) {
# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module which provides text formatting

use formatting

proc wordwrap {text length {firstprefix ""} {nextprefix ""}} {
    set len 0
    set space $firstprefix
    foreach word [split $text] {
        set word [string trim $word]
        if {$word == ""} {
            continue
        }
        if {$len && [string length $space$word] + $len >= $length} {
            puts ""
            set len 0
            set space $nextprefix
        }
        incr len [string length $space$word]

        # Use man-page conventions for highlighting 'quoted' and *quoted*
        # single words.
        # Use x^Hx for *bold* and _^Hx for 'underline'.
        #
        # less and more will both understand this.
        # Pipe through 'col -b' to remove them.
        if {[regexp {^'(.*)'([^a-zA-Z0-9_]*)$} $word -> bareword dot]} {
            regsub -all . $bareword "_\b&" word
            append word $dot
        } elseif {[regexp {^[*](.*)[*]([^a-zA-Z0-9_]*)$} $word -> bareword dot]} {
            regsub -all . $bareword "&\b&" word
            append word $dot
        }
        puts -nonewline $space$word
        set space " "
    }
    if {$len} {
        puts ""
    }
}
proc title {text} {
    underline [string trim $text] =
    nl
}
proc p {text} {
    wordwrap $text 80
    nl
}
proc codelines {lines} {
    foreach line $lines {
        puts "    $line"
    }
    nl
}
proc nl {} {
    puts ""
}
proc underline {text char} {
    regexp "^(\[ \t\]*)(.*)" $text -> indent words
    puts $text
    puts $indent[string repeat $char [string length $words]]
}
proc section {text} {
    underline "[string trim $text]" -
    nl
}
proc subsection {text} {
    underline "$text" ~
    nl
}
proc bullet {text} {
    wordwrap $text 76 "  * " "    "
}
proc indent {text} {
    wordwrap $text 76 "    " "    "
}
proc defn {first args} {
    if {$first ne ""} {
        underline "    $first" ~
    }
    foreach p $args {
        if {$p ne ""} {
            indent $p
        }
    }
}
}

# ----- module wiki-formatting -----

set modsource(wiki-formatting) {
# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# Module which provides text formatting
# wiki.tcl.tk format output

use formatting

proc joinlines {text} {
    set lines {}
    foreach l [split [string trim $text] \n] {
        lappend lines [string trim $l]
    }
    join $lines
}
proc p {text} {
    puts [joinlines $text]
    puts ""
}
proc title {text} {
    puts "*** [joinlines $text] ***"
    puts ""
}
proc codelines {lines} {
    puts "======"
    foreach line $lines {
        puts "    $line"
    }
    puts "======"
}
proc code {text} {
    puts "======"
    foreach line [parse_code_block $text] {
        puts "    $line"
    }
    puts "======"
}
proc nl {} {
}
proc section {text} {
    puts "'''$text'''"
    puts ""
}
proc subsection {text} {
    puts "''$text''"
    puts ""
}
proc bullet {text} {
    puts "   * [joinlines $text]"
}
proc indent {text} {
    puts "    :    [joinlines $text]"
}
proc defn {first args} {
    if {$first ne ""} {
        indent '''$first'''
    }

    foreach p $args {
        p $p
    }
}
}


##################################################################
#
# Entry/Exit
#
if {$autosetup(debug)} {
	main $argv
}
if {[catch {main $argv} msg opts] == 1} {
	show-notices
	autosetup-full-error [error-dump $msg $opts $::autosetup(debug)]
	if {!$autosetup(debug)} {
		puts stderr "Try: '[file tail $autosetup(exe)] --debug' for a full stack trace"
	}
	exit 1
}

# Copyright (c) 2011 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# @synopsis:
#
# The 'cc-db' module provides a knowledge based of system idiosyncrasies
# In general, this module can always be included

use cc

module-options {}

# openbsd needs sys/types.h to detect some system headers
cc-include-needs sys/socket.h sys/types.h
cc-include-needs netinet/in.h sys/types.h

# Copyright (c) 2011 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# @synopsis:
#
# Provides a library of common tests on top of the 'cc' module.

use cc

module-options {}

# @cc-check-lfs
#
# The equivalent of the AC_SYS_LARGEFILE macro
# 
# defines 'HAVE_LFS' if LFS is available,
# and defines '_FILE_OFFSET_BITS=64' if necessary
#
# Returns 1 if 'LFS' is available or 0 otherwise
#
proc cc-check-lfs {} {
	cc-check-includes sys/types.h
	msg-checking "Checking if -D_FILE_OFFSET_BITS=64 is needed..."
	set lfs 1
	if {[msg-quiet cc-with {-includes sys/types.h} {cc-check-sizeof off_t}] == 8} {
		msg-result no
	} elseif {[msg-quiet cc-with {-includes sys/types.h -cflags -D_FILE_OFFSET_BITS=64} {cc-check-sizeof off_t}] == 8} {
		define _FILE_OFFSET_BITS 64
		msg-result yes
	} else {
		set lfs 0
		msg-result none
	}
	define-feature lfs $lfs
	return $lfs
}

# @cc-check-endian
#
# The equivalent of the AC_C_BIGENDIAN macro
# 
# defines 'HAVE_BIG_ENDIAN' if endian is known to be big,
# or 'HAVE_LITTLE_ENDIAN' if endian is known to be little.
#
# Returns 1 if determined, or 0 if not.
#
proc cc-check-endian {} {
	cc-check-includes sys/types.h sys/param.h
	set rc 0
	msg-checking "Checking endian..."
	cc-with {-includes {sys/types.h sys/param.h}} {
		if {[cctest -code {
			#if !defined(BIG_ENDIAN) || !defined(BYTE_ORDER)
				#error unknown
			#elif BYTE_ORDER != BIG_ENDIAN
				#error little
			#endif
		}]} {
			define-feature big-endian
			msg-result "big"
			set rc 1
		} elseif {[cctest -code {
			#if !defined(LITTLE_ENDIAN) || !defined(BYTE_ORDER)
				#error unknown
			#elif BYTE_ORDER != LITTLE_ENDIAN
				#error big
			#endif
		}]} {
			define-feature little-endian
			msg-result "little"
			set rc 1
		} else {
			msg-result "unknown"
		}
	}
	return $rc
}

# @cc-check-flags flag ?...?
#
# Checks whether the given C/C++ compiler flags can be used. Defines feature
# names prefixed with 'HAVE_CFLAG' and 'HAVE_CXXFLAG' respectively, and
# appends working flags to '-cflags' and 'CFLAGS' or 'CXXFLAGS'.
proc cc-check-flags {args} {
    set result 1
    array set opts [cc-get-settings]
    switch -exact -- $opts(-lang) {
        c++ {
            set lang C++
            set prefix CXXFLAG
        }
        c {
            set lang C
            set prefix CFLAG
        }
        default {
            autosetup-error "cc-check-flags failed with unknown language: $opts(-lang)"
        }
    }
    foreach flag $args {
        msg-checking "Checking whether the $lang compiler accepts $flag..."
        if {[cctest -cflags $flag]} {
            msg-result yes
            define-feature $prefix$flag
            cc-with [list -cflags [list $flag]]
            define-append ${prefix}S $flag
        } else {
            msg-result no
            set result 0
        }
    }
    return $result
}

# @cc-check-standards ver ?...?
#
# Checks whether the C/C++ compiler accepts one of the specified '-std=$ver'
# options, and appends the first working one to '-cflags' and 'CFLAGS' or
# 'CXXFLAGS'.
proc cc-check-standards {args} {
    array set opts [cc-get-settings]
    foreach std $args {
        if {[cc-check-flags -std=$std]} {
            return $std
        }
    }
    return ""
}

# Checks whether $keyword is usable as alignof
proc cctest_alignof {keyword} {
    msg-checking "Checking for $keyword..."
    if {[cctest -code [subst -nobackslashes {
        printf("minimum alignment is %d == %d\n", ${keyword}(char), ${keyword}('x'));
    }]]} then {
        msg-result ok
        define-feature $keyword
    } else {
        msg-result "not found"
    }
}

# @cc-check-c11
#
# Checks for several C11/C++11 extensions and their alternatives. Currently
# checks for '_Static_assert', '_Alignof', '__alignof__', '__alignof'.
proc cc-check-c11 {} {
    msg-checking "Checking for _Static_assert..."
    if {[cctest -code {
        _Static_assert(1, "static assertions are available");
    }]} then {
        msg-result ok
        define-feature _Static_assert
    } else {
        msg-result "not found"
    }

    cctest_alignof _Alignof
    cctest_alignof __alignof__
    cctest_alignof __alignof
}

# @cc-check-alloca
#
# The equivalent of the AC_FUNC_ALLOCA macro
#
# Checks for the existence of alloca
# defines HAVE_ALLOCA and returns 1 if it exists
proc cc-check-alloca {} {
    cc-check-some-feature alloca {
        cctest -includes alloca.h -code { alloca (2 * sizeof (int)); }
    }
}

# @cc-signal-return-type
#
# The equivalent of the AC_TYPE_SIGNAL macro
#
# defines RETSIGTYPE to int or void
proc cc-signal-return-type {} {
    msg-checking "Checking return type of signal handlers..."
    cc-with {-includes {sys/types.h signal.h}} {
        if {[cctest -code {return *(signal (0, 0)) (0) == 1;}]} {
                set type int
        } else {
                set type void
        }
        define RETSIGTYPE $type
        msg-result $type
    }
}

# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# @synopsis:
#
# The 'cc-shared' module provides support for shared libraries and shared objects.
# It defines the following variables:
#
## SH_CFLAGS         Flags to use compiling sources destined for a shared library
## SH_LDFLAGS        Flags to use linking (creating) a shared library
## SH_SOPREFIX       Prefix to use to set the soname when creating a shared library
## SH_SOEXT          Extension for shared libs
## SH_SOEXTVER       Format for versioned shared libs - %s = version
## SHOBJ_CFLAGS      Flags to use compiling sources destined for a shared object
## SHOBJ_LDFLAGS     Flags to use linking a shared object, undefined symbols allowed
## SHOBJ_LDFLAGS_R   - as above, but all symbols must be resolved
## SH_LINKFLAGS      Flags to use linking an executable which will load shared objects
## LD_LIBRARY_PATH   Environment variable which specifies path to shared libraries
## STRIPLIBFLAGS     Arguments to strip a dynamic library

module-options {}

# Defaults: gcc on unix
define SHOBJ_CFLAGS -fpic
define SHOBJ_LDFLAGS -shared
define SH_CFLAGS -fpic
define SH_LDFLAGS -shared
define SH_LINKFLAGS -rdynamic
define SH_SOEXT .so
define SH_SOEXTVER .so.%s
define SH_SOPREFIX -Wl,-soname,
define LD_LIBRARY_PATH LD_LIBRARY_PATH
define STRIPLIBFLAGS --strip-unneeded

# Note: This is a helpful reference for identifying the toolchain
#       http://sourceforge.net/apps/mediawiki/predef/index.php?title=Compilers

switch -glob -- [get-define host] {
	*-*-darwin* {
		define SHOBJ_CFLAGS "-dynamic -fno-common"
		define SHOBJ_LDFLAGS "-bundle -undefined dynamic_lookup"
		define SHOBJ_LDFLAGS_R -bundle
		define SH_CFLAGS -dynamic
		define SH_LDFLAGS -dynamiclib
		define SH_LINKFLAGS ""
		define SH_SOEXT .dylib
		define SH_SOEXTVER .%s.dylib
		define SH_SOPREFIX -Wl,-install_name,
		define LD_LIBRARY_PATH DYLD_LIBRARY_PATH
		define STRIPLIBFLAGS -x
	}
	*-*-ming* - *-*-cygwin - *-*-msys {
		define SHOBJ_CFLAGS ""
		define SHOBJ_LDFLAGS -shared
		define SH_CFLAGS ""
		define SH_LDFLAGS -shared
		define SH_LINKFLAGS ""
		define SH_SOEXT .dll
		define SH_SOEXTVER .dll
		define SH_SOPREFIX ""
		define LD_LIBRARY_PATH PATH
	}
	sparc* {
		if {[msg-quiet cc-check-decls __SUNPRO_C]} {
			msg-result "Found sun stdio compiler"
			# sun stdio compiler
			# XXX: These haven't been fully tested. 
			define SHOBJ_CFLAGS -KPIC
			define SHOBJ_LDFLAGS "-G"
			define SH_CFLAGS -KPIC
			define SH_LINKFLAGS -Wl,-export-dynamic
			define SH_SOPREFIX -Wl,-h,
		} else {
			# sparc has a very small GOT table limit, so use -fPIC
			define SH_CFLAGS -fPIC
			define SHOBJ_CFLAGS -fPIC
		}
	}
	*-*-solaris* {
		if {[msg-quiet cc-check-decls __SUNPRO_C]} {
			msg-result "Found sun stdio compiler"
			# sun stdio compiler
			# XXX: These haven't been fully tested. 
			define SHOBJ_CFLAGS -KPIC
			define SHOBJ_LDFLAGS "-G"
			define SH_CFLAGS -KPIC
			define SH_LINKFLAGS -Wl,-export-dynamic
			define SH_SOPREFIX -Wl,-h,
		}
	}
	*-*-hpux {
		# XXX: These haven't been tested
		define SHOBJ_CFLAGS "+O3 +z"
		define SHOBJ_LDFLAGS -b
		define SH_CFLAGS +z
		define SH_LINKFLAGS -Wl,+s
		define LD_LIBRARY_PATH SHLIB_PATH
	}
	*-*-haiku {
		define SHOBJ_CFLAGS ""
		define SHOBJ_LDFLAGS -shared
		define SH_CFLAGS ""
		define SH_LDFLAGS -shared
		define SH_LINKFLAGS ""
		define SH_SOPREFIX ""
		define LD_LIBRARY_PATH LIBRARY_PATH
	}
	microblaze* {
		# Microblaze generally needs -fPIC rather than -fpic
		define SHOBJ_CFLAGS -fPIC
		define SH_CFLAGS -fPIC
	}
}

if {![is-defined SHOBJ_LDFLAGS_R]} {
	define SHOBJ_LDFLAGS_R [get-define SHOBJ_LDFLAGS]
}

# Copyright (c) 2010 WorkWare Systems http://www.workware.net.au/
# All rights reserved

# @synopsis:
#
# The 'cc' module supports checking various 'features' of the C or C++
# compiler/linker environment. Common commands are cc-check-includes,
# cc-check-types, cc-check-functions, cc-with, make-autoconf-h and make-template.
#
# The following environment variables are used if set:
#
## CC       - C compiler
## CXX      - C++ compiler
## CCACHE   - Set to "none" to disable automatic use of ccache
## CFLAGS   - Additional C compiler flags
## CXXFLAGS - Additional C++ compiler flags
## LDFLAGS  - Additional compiler flags during linking
## LIBS     - Additional libraries to use (for all tests)
## CROSS    - Tool prefix for cross compilation
#
# The following variables are defined from the corresponding
# environment variables if set.
#
## CPPFLAGS
## LINKFLAGS
## CC_FOR_BUILD
## LD

use system

module-options {}

# Note that the return code is not meaningful
proc cc-check-something {name code} {
	uplevel 1 $code
}

# Checks for the existence of the given function by linking
#
proc cctest_function {function} {
	cctest -link 1 -declare "extern void $function\(void);" -code "$function\();"
}

# Checks for the existence of the given type by compiling
proc cctest_type {type} {
	cctest -code "$type _x;"
}

# Checks for the existence of the given type/structure member.
# e.g. "struct stat.st_mtime"
proc cctest_member {struct_member} {
	# split at the first dot
	regexp {^([^.]+)[.](.*)$} $struct_member -> struct member
	cctest -code "static $struct _s; return sizeof(_s.$member);"
}

# Checks for the existence of the given define by compiling
#
proc cctest_define {name} {
	cctest -code "#ifndef $name\n#error not defined\n#endif"
}

# Checks for the existence of the given name either as
# a macro (#define) or an rvalue (such as an enum)
#
proc cctest_decl {name} {
	cctest -code "#ifndef $name\n(void)$name;\n#endif"
}

# @cc-check-sizeof type ...
#
# Checks the size of the given types (between 1 and 32, inclusive).
# Defines a variable with the size determined, or "unknown" otherwise.
# e.g. for type 'long long', defines SIZEOF_LONG_LONG.
# Returns the size of the last type.
#
proc cc-check-sizeof {args} {
	foreach type $args {
		msg-checking "Checking for sizeof $type..."
		set size unknown
		# Try the most common sizes first
		foreach i {4 8 1 2 16 32} {
			if {[cctest -code "static int _x\[sizeof($type) == $i ? 1 : -1\] = { 1 };"]} {
				set size $i
				break
			}
		}
		msg-result $size
		set define [feature-define-name $type SIZEOF_]
		define $define $size
	}
	# Return the last result
	get-define $define
}

# Checks for each feature in $list by using the given script.
#
# When the script is evaluated, $each is set to the feature
# being checked, and $extra is set to any additional cctest args.
#
# Returns 1 if all features were found, or 0 otherwise.
proc cc-check-some-feature {list script} {
	set ret 1
	foreach each $list {
		if {![check-feature $each $script]} {
			set ret 0
		}
	}
	return $ret
}

# @cc-check-includes includes ...
#
# Checks that the given include files can be used
proc cc-check-includes {args} {
	cc-check-some-feature $args {
		set with {}
		if {[dict exists $::autosetup(cc-include-deps) $each]} {
			set deps [dict keys [dict get $::autosetup(cc-include-deps) $each]]
			msg-quiet cc-check-includes {*}$deps
			foreach i $deps {
				if {[have-feature $i]} {
					lappend with $i
				}
			}
		}
		if {[llength $with]} {
			cc-with [list -includes $with] {
				cctest -includes $each
			}
		} else {
			cctest -includes $each
		}
	}
}

# @cc-include-needs include required ...
#
# Ensures that when checking for 'include', a check is first
# made for each 'required' file, and if found, it is #included
proc cc-include-needs {file args} {
	foreach depfile $args {
		dict set ::autosetup(cc-include-deps) $file $depfile 1
	}
}

# @cc-check-types type ...
#
# Checks that the types exist.
proc cc-check-types {args} {
	cc-check-some-feature $args {
		cctest_type $each
	}
}

# @cc-check-defines define ...
#
# Checks that the given preprocessor symbol is defined
proc cc-check-defines {args} {
	cc-check-some-feature $args {
		cctest_define $each
	}
}

# @cc-check-decls name ...
#
# Checks that each given name is either a preprocessor symbol or rvalue
# such as an enum. Note that the define used is HAVE_DECL_xxx
# rather than HAVE_xxx
proc cc-check-decls {args} {
	set ret 1
	foreach name $args {
		msg-checking "Checking for $name..."
		set r [cctest_decl $name]
		define-feature "decl $name" $r
		if {$r} {
			msg-result "ok"
		} else {
			msg-result "not found"
			set ret 0
		}
	}
	return $ret
}

# @cc-check-functions function ...
#
# Checks that the given functions exist (can be linked)
proc cc-check-functions {args} {
	cc-check-some-feature $args {
		cctest_function $each
	}
}

# @cc-check-members type.member ...
#
# Checks that the given type/structure members exist.
# A structure member is of the form "struct stat.st_mtime"
proc cc-check-members {args} {
	cc-check-some-feature $args {
		cctest_member $each
	}
}

# @cc-check-function-in-lib function libs ?otherlibs?
#
# Checks that the given function can be found in one of the libs.
#
# First checks for no library required, then checks each of the libraries
# in turn.
#
# If the function is found, the feature is defined and lib_$function is defined
# to -l$lib where the function was found, or "" if no library required.
# In addition, -l$lib is prepended to the LIBS define.
#
# If additional libraries may be needed for linking, they should be specified
# as $extralibs as "-lotherlib1 -lotherlib2".
# These libraries are not automatically added to LIBS.
#
# Returns 1 if found or 0 if not.
# 
proc cc-check-function-in-lib {function libs {otherlibs {}}} {
	msg-checking "Checking libs for $function..."
	set found 0
	cc-with [list -libs $otherlibs] {
		if {[cctest_function $function]} {
			msg-result "none needed"
			define lib_$function ""
			incr found
		} else {
			foreach lib $libs {
				cc-with [list -libs -l$lib] {
					if {[cctest_function $function]} {
						msg-result -l$lib
						define lib_$function -l$lib
						# prepend to LIBS
						define LIBS "-l$lib [get-define LIBS]"
						incr found
						break
					}
				}
			}
		}
	}
	if {$found} {
		define [feature-define-name $function]
	} else {
		msg-result "no"
	}
	return $found
}

# @cc-check-tools tool ...
#
# Checks for existence of the given compiler tools, taking
# into account any cross compilation prefix.
#
# For example, when checking for "ar", first AR is checked on the command
# line and then in the environment. If not found, "${host}-ar" or
# simply "ar" is assumed depending upon whether cross compiling.
# The path is searched for this executable, and if found AR is defined
# to the executable name.
# Note that even when cross compiling, the simple "ar" is used as a fallback,
# but a warning is generated. This is necessary for some toolchains.
#
# It is an error if the executable is not found.
#
proc cc-check-tools {args} {
	foreach tool $args {
		set TOOL [string toupper $tool]
		set exe [get-env $TOOL [get-define cross]$tool]
		if {[find-executable {*}$exe]} {
			define $TOOL $exe
			continue
		}
		if {[find-executable {*}$tool]} {
			msg-result "Warning: Failed to find $exe, falling back to $tool which may be incorrect"
			define $TOOL $tool
			continue
		}
		user-error "Failed to find $exe"
	}
}

# @cc-check-progs prog ...
#
# Checks for existence of the given executables on the path.
#
# For example, when checking for "grep", the path is searched for
# the executable, 'grep', and if found GREP is defined as "grep".
#
# If the executable is not found, the variable is defined as false.
# Returns 1 if all programs were found, or 0 otherwise.
#
proc cc-check-progs {args} {
	set failed 0
	foreach prog $args {
		set PROG [string toupper $prog]
		msg-checking "Checking for $prog..."
		if {![find-executable $prog]} {
			msg-result no
			define $PROG false
			incr failed
		} else {
			msg-result ok
			define $PROG $prog
		}
	}
	expr {!$failed}
}

# Adds the given settings to $::autosetup(ccsettings) and
# returns the old settings.
#
proc cc-add-settings {settings} {
	if {[llength $settings] % 2} {
		autosetup-error "settings list is missing a value: $settings"
	}

	set prev [cc-get-settings]
	# workaround a bug in some versions of jimsh by forcing
	# conversion of $prev to a list
	llength $prev

	array set new $prev

	foreach {name value} $settings {
		switch -exact -- $name {
			-cflags - -includes {
				# These are given as lists
				lappend new($name) {*}$value
			}
			-declare {
				lappend new($name) $value
			}
			-libs {
				# Note that new libraries are added before previous libraries
				set new($name) [list {*}$value {*}$new($name)]
			}
			-link - -lang - -nooutput {
				set new($name) $value
			}
			-source - -sourcefile - -code {
				# XXX: These probably are only valid directly from cctest
				set new($name) $value
			}
			default {
				autosetup-error "unknown cctest setting: $name"
			}
		}
	}

	cc-store-settings [array get new]

	return $prev
}

proc cc-store-settings {new} {
	set ::autosetup(ccsettings) $new
}

proc cc-get-settings {} {
	return $::autosetup(ccsettings)
}

# Similar to cc-add-settings, but each given setting
# simply replaces the existing value.
#
# Returns the previous settings
proc cc-update-settings {args} {
	set prev [cc-get-settings]
	cc-store-settings [dict merge $prev $args]
	return $prev
}

# @cc-with settings ?{ script }?
#
# Sets the given 'cctest' settings and then runs the tests in 'script'.
# Note that settings such as -lang replace the current setting, while
# those such as -includes are appended to the existing setting.
#
# If no script is given, the settings become the default for the remainder
# of the auto.def file.
#
## cc-with {-lang c++} {
##   # This will check with the C++ compiler
##   cc-check-types bool
##   cc-with {-includes signal.h} {
##     # This will check with the C++ compiler, signal.h and any existing includes.
##     ...
##   }
##   # back to just the C++ compiler
## }
#
# The -libs setting is special in that newer values are added *before* earlier ones.
#
## cc-with {-libs {-lc -lm}} {
##   cc-with {-libs -ldl} {
##     cctest -libs -lsocket ...
##     # libs will be in this order: -lsocket -ldl -lc -lm
##   }
## }
proc cc-with {settings args} {
	if {[llength $args] == 0} {
		cc-add-settings $settings
	} elseif {[llength $args] > 1} {
		autosetup-error "usage: cc-with settings ?script?"
	} else {
		set save [cc-add-settings $settings]
		set rc [catch {uplevel 1 [lindex $args 0]} result info]
		cc-store-settings $save
		if {$rc != 0} {
			return -code [dict get $info -code] $result
		}
		return $result
	}
}

# @cctest ?settings?
# 
# Low level C compiler checker. Compiles and or links a small C program
# according to the arguments and returns 1 if OK, or 0 if not.
#
# Supported settings are:
#
## -cflags cflags      A list of flags to pass to the compiler
## -includes list      A list of includes, e.g. {stdlib.h stdio.h}
## -declare code       Code to declare before main()
## -link 1             Don't just compile, link too
## -lang c|c++         Use the C (default) or C++ compiler
## -libs liblist       List of libraries to link, e.g. {-ldl -lm}
## -code code          Code to compile in the body of main()
## -source code        Compile a complete program. Ignore -includes, -declare and -code
## -sourcefile file    Shorthand for -source [readfile [get-define srcdir]/$file]
## -nooutput 1         Treat any compiler output (e.g. a warning) as an error
#
# Unless -source or -sourcefile is specified, the C program looks like:
#
## #include <firstinclude>   /* same for remaining includes in the list */
##
## declare-code              /* any code in -declare, verbatim */
##
## int main(void) {
##   code                    /* any code in -code, verbatim */
##   return 0;
## }
#
# Any failures are recorded in 'config.log'
#
proc cctest {args} {
	set src conftest__.c
	set tmp conftest__

	# Easiest way to merge in the settings
	cc-with $args {
		array set opts [cc-get-settings]
	}

	if {[info exists opts(-sourcefile)]} {
		set opts(-source) [readfile [get-define srcdir]/$opts(-sourcefile) "#error can't find $opts(-sourcefile)"]
	}
	if {[info exists opts(-source)]} {
		set lines $opts(-source)
	} else {
		foreach i $opts(-includes) {
			if {$opts(-code) ne "" && ![feature-checked $i]} {
				# Compiling real code with an unchecked header file
				# Quickly (and silently) check for it now

				# Remove all -includes from settings before checking
				set saveopts [cc-update-settings -includes {}]
				msg-quiet cc-check-includes $i
				cc-store-settings $saveopts
			}
			if {$opts(-code) eq "" || [have-feature $i]} {
				lappend source "#include <$i>"
			}
		}
		lappend source {*}$opts(-declare)
		lappend source "int main(void) {"
		lappend source $opts(-code)
		lappend source "return 0;"
		lappend source "}"

		set lines [join $source \n]
	}

	# Build the command line
	set cmdline {}
	lappend cmdline {*}[get-define CCACHE]
	switch -exact -- $opts(-lang) {
		c++ {
			lappend cmdline {*}[get-define CXX] {*}[get-define CXXFLAGS]
		}
		c {
			lappend cmdline {*}[get-define CC] {*}[get-define CFLAGS]
		}
		default {
			autosetup-error "cctest called with unknown language: $opts(-lang)"
		}
	}

	if {$opts(-link)} {
		lappend cmdline {*}[get-define LDFLAGS]
	} else {
		set tmp conftest__.o
		lappend cmdline -c
	}
	lappend cmdline {*}$opts(-cflags) {*}[get-define cc-default-debug ""]
	lappend cmdline $src -o $tmp {*}$opts(-libs)
	if {$opts(-link)} {
		lappend cmdline {*}[get-define LIBS]
	}

	# At this point we have the complete command line and the
	# complete source to be compiled. Get the result from cache if
	# we can
	if {[info exists ::cc_cache($cmdline,$lines)]} {
		msg-checking "(cached) "
		set ok $::cc_cache($cmdline,$lines)
		if {$::autosetup(debug)} {
			configlog "From cache (ok=$ok): [join $cmdline]"
			configlog "============"
			configlog $lines
			configlog "============"
		}
		return $ok
	}

	writefile $src $lines\n

	set ok 1
	set err [catch {exec-with-stderr {*}$cmdline} result errinfo]
	if {$err || ($opts(-nooutput) && [string length $result])} {
		configlog "Failed: [join $cmdline]"
		configlog $result
		configlog "============"
		configlog "The failed code was:"
		configlog $lines
		configlog "============"
		set ok 0
	} elseif {$::autosetup(debug)} {
		configlog "Compiled OK: [join $cmdline]"
		configlog "============"
		configlog $lines
		configlog "============"
	}
	file delete $src
	file delete $tmp

	# cache it
	set ::cc_cache($cmdline,$lines) $ok

	return $ok
}

# @make-autoconf-h outfile ?auto-patterns=HAVE_*? ?bare-patterns=SIZEOF_*?
#
# Deprecated - see make-config-header
proc make-autoconf-h {file {autopatterns {HAVE_*}} {barepatterns {SIZEOF_* HAVE_DECL_*}}} {
	user-notice "*** make-autoconf-h is deprecated -- use make-config-header instead"
	make-config-header $file -auto $autopatterns -bare $barepatterns
}

# @make-config-header outfile ?-auto patternlist? ?-bare patternlist? ?-none patternlist? ?-str patternlist? ...
#
# Examines all defined variables which match the given patterns
# and writes an include file, $file, which defines each of these.
# Variables which match '-auto' are output as follows:
# - defines which have the value "0" are ignored.
# - defines which have integer values are defined as the integer value.
# - any other value is defined as a string, e.g. "value"
# Variables which match '-bare' are defined as-is.
# Variables which match '-str' are defined as a string, e.g. "value"
# Variables which match '-none' are omitted.
#
# Note that order is important. The first pattern which matches is selected
# Default behaviour is:
#
#  -bare {SIZEOF_* HAVE_DECL_*} -auto HAVE_* -none *
#
# If the file would be unchanged, it is not written.
proc make-config-header {file args} {
	set guard _[string toupper [regsub -all {[^a-zA-Z0-9]} [file tail $file] _]]
	file mkdir [file dirname $file]
	set lines {}
	lappend lines "#ifndef $guard"
	lappend lines "#define $guard"

	# Add some defaults
	lappend args -bare {SIZEOF_* HAVE_DECL_*} -auto HAVE_*

	foreach n [lsort [dict keys [all-defines]]] {
		set value [get-define $n]
		set type [calc-define-output-type $n $args]
		switch -exact -- $type {
			-bare {
				# Just output the value unchanged
			}
			-none {
				continue
			}
			-str {
				set value \"[string map [list \\ \\\\ \" \\\"] $value]\"
			}
			-auto {
				# Automatically determine the type
				if {$value eq "0"} {
					lappend lines "/* #undef $n */"
					continue
				}
				if {![string is integer -strict $value]} {
					set value \"[string map [list \\ \\\\ \" \\\"] $value]\"
				}
			}
			"" {
				continue
			}
			default {
				autosetup-error "Unknown type in make-config-header: $type"
			}
		}
		lappend lines "#define $n $value"
	}
	lappend lines "#endif"
	set buf [join $lines \n]
	write-if-changed $file $buf {
		msg-result "Created $file"
	}
}

proc calc-define-output-type {name spec} {
	foreach {type patterns} $spec {
		foreach pattern $patterns {
			if {[string match $pattern $name]} {
				return $type
			}
		}
	}
	return ""
}

# Initialise some values from the environment or commandline or default settings
foreach i {LDFLAGS LIBS CPPFLAGS LINKFLAGS {CFLAGS "-g -O2"}} {
	lassign $i var default
	define $var [get-env $var $default]
}

if {[env-is-set CC]} {
	# Set by the user, so don't try anything else
	set try [list [get-env CC ""]]
} else {
	# Try some reasonable options
	set try [list [get-define cross]cc [get-define cross]gcc]
}
define CC [find-an-executable {*}$try]
if {[get-define CC] eq ""} {
	user-error "Could not find a C compiler. Tried: [join $try ", "]"
}

define CPP [get-env CPP "[get-define CC] -E"]

# XXX: Could avoid looking for a C++ compiler until requested
# Note that if CXX isn't found, we just set it to "false". It might not be needed.
if {[env-is-set CXX]} {
	define CXX [find-an-executable -required [get-env CXX ""]]
} else {
	define CXX [find-an-executable [get-define cross]c++ [get-define cross]g++ false]
}

# CXXFLAGS default to CFLAGS if not specified
define CXXFLAGS [get-env CXXFLAGS [get-define CFLAGS]]

# May need a CC_FOR_BUILD, so look for one
define CC_FOR_BUILD [find-an-executable [get-env CC_FOR_BUILD ""] cc gcc false]

if {[get-define CC] eq ""} {
	user-error "Could not find a C compiler. Tried: [join $try ", "]"
}

define CCACHE [find-an-executable [get-env CCACHE ccache]]

# Initial cctest settings
cc-store-settings {-cflags {} -includes {} -declare {} -link 0 -lang c -libs {} -code {} -nooutput 0}
set autosetup(cc-include-deps) {}

msg-result "C compiler...[get-define CCACHE] [get-define CC] [get-define CFLAGS]"
if {[get-define CXX] ne "false"} {
	msg-result "C++ compiler...[get-define CCACHE] [get-define CXX] [get-define CXXFLAGS]"
}
msg-result "Build C compiler...[get-define CC_FOR_BUILD]"

# On Darwin, we prefer to use -g0 to avoid creating .dSYM directories
# but some compilers may not support it, so test here.
switch -glob -- [get-define host] {
	*-*-darwin* {
		if {[cctest -cflags {-g0}]} {
			define cc-default-debug -g0
		}
	}
}

if {![cc-check-includes stdlib.h]} {
	user-error "Compiler does not work. See config.log"
}

#! /bin/sh
# Attempt to guess a canonical system name.
#   Copyright 1992-2014 Free Software Foundation, Inc.

timestamp='2014-11-04'

# This file is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
# General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, see <http://www.gnu.org/licenses/>.
#
# As a special exception to the GNU General Public License, if you
# distribute this file as part of a program that contains a
# configuration script generated by Autoconf, you may include it under
# the same distribution terms that you use for the rest of that
# program.  This Exception is an additional permission under section 7
# of the GNU General Public License, version 3 ("GPLv3").
#
# Originally written by Per Bothner; maintained since 2000 by Ben Elliston.
#
# You can get the latest version of this script from:
# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD
#
# Please send patches to <config-patches@gnu.org>.


me=`echo "$0" | sed -e 's,.*/,,'`

usage="\
Usage: $0 [OPTION]

Output the configuration name of the system \`$me' is run on.

Operation modes:
  -h, --help         print this help, then exit
  -t, --time-stamp   print date of last modification, then exit
  -v, --version      print version number, then exit

Report bugs and patches to <config-patches@gnu.org>."

version="\
GNU config.guess ($timestamp)

Originally written by Per Bothner.
Copyright 1992-2014 Free Software Foundation, Inc.

This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."

help="
Try \`$me --help' for more information."

# Parse command line
while test $# -gt 0 ; do
  case $1 in
    --time-stamp | --time* | -t )
       echo "$timestamp" ; exit ;;
    --version | -v )
       echo "$version" ; exit ;;
    --help | --h* | -h )
       echo "$usage"; exit ;;
    -- )     # Stop option processing
       shift; break ;;
    - )	# Use stdin as input.
       break ;;
    -* )
       echo "$me: invalid option $1$help" >&2
       exit 1 ;;
    * )
       break ;;
  esac
done

if test $# != 0; then
  echo "$me: too many arguments$help" >&2
  exit 1
fi

trap 'exit 1' 1 2 15

# CC_FOR_BUILD -- compiler used by this script. Note that the use of a
# compiler to aid in system detection is discouraged as it requires
# temporary files to be created and, as you can see below, it is a
# headache to deal with in a portable fashion.

# Historically, `CC_FOR_BUILD' used to be named `HOST_CC'. We still
# use `HOST_CC' if defined, but it is deprecated.

# Portable tmp directory creation inspired by the Autoconf team.

set_cc_for_build='
trap "exitcode=\$?; (rm -f \$tmpfiles 2>/dev/null; rmdir \$tmp 2>/dev/null) && exit \$exitcode" 0 ;
trap "rm -f \$tmpfiles 2>/dev/null; rmdir \$tmp 2>/dev/null; exit 1" 1 2 13 15 ;
: ${TMPDIR=/tmp} ;
 { tmp=`(umask 077 && mktemp -d "$TMPDIR/cgXXXXXX") 2>/dev/null` && test -n "$tmp" && test -d "$tmp" ; } ||
 { test -n "$RANDOM" && tmp=$TMPDIR/cg$$-$RANDOM && (umask 077 && mkdir $tmp) ; } ||
 { tmp=$TMPDIR/cg-$$ && (umask 077 && mkdir $tmp) && echo "Warning: creating insecure temp directory" >&2 ; } ||
 { echo "$me: cannot create a temporary directory in $TMPDIR" >&2 ; exit 1 ; } ;
dummy=$tmp/dummy ;
tmpfiles="$dummy.c $dummy.o $dummy.rel $dummy" ;
case $CC_FOR_BUILD,$HOST_CC,$CC in
 ,,)    echo "int x;" > $dummy.c ;
	for c in cc gcc c89 c99 ; do
	  if ($c -c -o $dummy.o $dummy.c) >/dev/null 2>&1 ; then
	     CC_FOR_BUILD="$c"; break ;
	  fi ;
	done ;
	if test x"$CC_FOR_BUILD" = x ; then
	  CC_FOR_BUILD=no_compiler_found ;
	fi
	;;
 ,,*)   CC_FOR_BUILD=$CC ;;
 ,*,*)  CC_FOR_BUILD=$HOST_CC ;;
esac ; set_cc_for_build= ;'

# This is needed to find uname on a Pyramid OSx when run in the BSD universe.
# (ghazi@noc.rutgers.edu 1994-08-24)
if (test -f /.attbin/uname) >/dev/null 2>&1 ; then
	PATH=$PATH:/.attbin ; export PATH
fi

UNAME_MACHINE=`(uname -m) 2>/dev/null` || UNAME_MACHINE=unknown
UNAME_RELEASE=`(uname -r) 2>/dev/null` || UNAME_RELEASE=unknown
UNAME_SYSTEM=`(uname -s) 2>/dev/null`  || UNAME_SYSTEM=unknown
UNAME_VERSION=`(uname -v) 2>/dev/null` || UNAME_VERSION=unknown

case "${UNAME_SYSTEM}" in
Linux|GNU|GNU/*)
	# If the system lacks a compiler, then just pick glibc.
	# We could probably try harder.
	LIBC=gnu

	eval $set_cc_for_build
	cat <<-EOF > $dummy.c
	#include <features.h>
	#if defined(__UCLIBC__)
	LIBC=uclibc
	#elif defined(__dietlibc__)
	LIBC=dietlibc
	#else
	LIBC=gnu
	#endif
	EOF
	eval `$CC_FOR_BUILD -E $dummy.c 2>/dev/null | grep '^LIBC' | sed 's, ,,g'`
	;;
esac

# Note: order is significant - the case branches are not exclusive.

case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
    *:NetBSD:*:*)
	# NetBSD (nbsd) targets should (where applicable) match one or
	# more of the tuples: *-*-netbsdelf*, *-*-netbsdaout*,
	# *-*-netbsdecoff* and *-*-netbsd*.  For targets that recently
	# switched to ELF, *-*-netbsd* would select the old
	# object file format.  This provides both forward
	# compatibility and a consistent mechanism for selecting the
	# object file format.
	#
	# Note: NetBSD doesn't particularly care about the vendor
	# portion of the name.  We always set it to "unknown".
	sysctl="sysctl -n hw.machine_arch"
	UNAME_MACHINE_ARCH=`(/sbin/$sysctl 2>/dev/null || \
	    /usr/sbin/$sysctl 2>/dev/null || echo unknown)`
	case "${UNAME_MACHINE_ARCH}" in
	    armeb) machine=armeb-unknown ;;
	    arm*) machine=arm-unknown ;;
	    sh3el) machine=shl-unknown ;;
	    sh3eb) machine=sh-unknown ;;
	    sh5el) machine=sh5le-unknown ;;
	    *) machine=${UNAME_MACHINE_ARCH}-unknown ;;
	esac
	# The Operating System including object format, if it has switched
	# to ELF recently, or will in the future.
	case "${UNAME_MACHINE_ARCH}" in
	    arm*|i386|m68k|ns32k|sh3*|sparc|vax)
		eval $set_cc_for_build
		if echo __ELF__ | $CC_FOR_BUILD -E - 2>/dev/null \
			| grep -q __ELF__
		then
		    # Once all utilities can be ECOFF (netbsdecoff) or a.out (netbsdaout).
		    # Return netbsd for either.  FIX?
		    os=netbsd
		else
		    os=netbsdelf
		fi
		;;
	    *)
		os=netbsd
		;;
	esac
	# The OS release
	# Debian GNU/NetBSD machines have a different userland, and
	# thus, need a distinct triplet. However, they do not need
	# kernel version information, so it can be replaced with a
	# suitable tag, in the style of linux-gnu.
	case "${UNAME_VERSION}" in
	    Debian*)
		release='-gnu'
		;;
	    *)
		release=`echo ${UNAME_RELEASE}|sed -e 's/[-_].*/\./'`
		;;
	esac
	# Since CPU_TYPE-MANUFACTURER-KERNEL-OPERATING_SYSTEM:
	# contains redundant information, the shorter form:
	# CPU_TYPE-MANUFACTURER-OPERATING_SYSTEM is used.
	echo "${machine}-${os}${release}"
	exit ;;
    *:Bitrig:*:*)
	UNAME_MACHINE_ARCH=`arch | sed 's/Bitrig.//'`
	echo ${UNAME_MACHINE_ARCH}-unknown-bitrig${UNAME_RELEASE}
	exit ;;
    *:OpenBSD:*:*)
	UNAME_MACHINE_ARCH=`arch | sed 's/OpenBSD.//'`
	echo ${UNAME_MACHINE_ARCH}-unknown-openbsd${UNAME_RELEASE}
	exit ;;
    *:ekkoBSD:*:*)
	echo ${UNAME_MACHINE}-unknown-ekkobsd${UNAME_RELEASE}
	exit ;;
    *:SolidBSD:*:*)
	echo ${UNAME_MACHINE}-unknown-solidbsd${UNAME_RELEASE}
	exit ;;
    macppc:MirBSD:*:*)
	echo powerpc-unknown-mirbsd${UNAME_RELEASE}
	exit ;;
    *:MirBSD:*:*)
	echo ${UNAME_MACHINE}-unknown-mirbsd${UNAME_RELEASE}
	exit ;;
    alpha:OSF1:*:*)
	case $UNAME_RELEASE in
	*4.0)
		UNAME_RELEASE=`/usr/sbin/sizer -v | awk '{print $3}'`
		;;
	*5.*)
		UNAME_RELEASE=`/usr/sbin/sizer -v | awk '{print $4}'`
		;;
	esac
	# According to Compaq, /usr/sbin/psrinfo has been available on
	# OSF/1 and Tru64 systems produced since 1995.  I hope that
	# covers most systems running today.  This code pipes the CPU
	# types through head -n 1, so we only detect the type of CPU 0.
	ALPHA_CPU_TYPE=`/usr/sbin/psrinfo -v | sed -n -e 's/^  The alpha \(.*\) processor.*$/\1/p' | head -n 1`
	case "$ALPHA_CPU_TYPE" in
	    "EV4 (21064)")
		UNAME_MACHINE="alpha" ;;
	    "EV4.5 (21064)")
		UNAME_MACHINE="alpha" ;;
	    "LCA4 (21066/21068)")
		UNAME_MACHINE="alpha" ;;
	    "EV5 (21164)")
		UNAME_MACHINE="alphaev5" ;;
	    "EV5.6 (21164A)")
		UNAME_MACHINE="alphaev56" ;;
	    "EV5.6 (21164PC)")
		UNAME_MACHINE="alphapca56" ;;
	    "EV5.7 (21164PC)")
		UNAME_MACHINE="alphapca57" ;;
	    "EV6 (21264)")
		UNAME_MACHINE="alphaev6" ;;
	    "EV6.7 (21264A)")
		UNAME_MACHINE="alphaev67" ;;
	    "EV6.8CB (21264C)")
		UNAME_MACHINE="alphaev68" ;;
	    "EV6.8AL (21264B)")
		UNAME_MACHINE="alphaev68" ;;
	    "EV6.8CX (21264D)")
		UNAME_MACHINE="alphaev68" ;;
	    "EV6.9A (21264/EV69A)")
		UNAME_MACHINE="alphaev69" ;;
	    "EV7 (21364)")
		UNAME_MACHINE="alphaev7" ;;
	    "EV7.9 (21364A)")
		UNAME_MACHINE="alphaev79" ;;
	esac
	# A Pn.n version is a patched version.
	# A Vn.n version is a released version.
	# A Tn.n version is a released field test version.
	# A Xn.n version is an unreleased experimental baselevel.
	# 1.2 uses "1.2" for uname -r.
	echo ${UNAME_MACHINE}-dec-osf`echo ${UNAME_RELEASE} | sed -e 's/^[PVTX]//' | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz'`
	# Reset EXIT trap before exiting to avoid spurious non-zero exit code.
	exitcode=$?
	trap '' 0
	exit $exitcode ;;
    Alpha\ *:Windows_NT*:*)
	# How do we know it's Interix rather than the generic POSIX subsystem?
	# Should we change UNAME_MACHINE based on the output of uname instead
	# of the specific Alpha model?
	echo alpha-pc-interix
	exit ;;
    21064:Windows_NT:50:3)
	echo alpha-dec-winnt3.5
	exit ;;
    Amiga*:UNIX_System_V:4.0:*)
	echo m68k-unknown-sysv4
	exit ;;
    *:[Aa]miga[Oo][Ss]:*:*)
	echo ${UNAME_MACHINE}-unknown-amigaos
	exit ;;
    *:[Mm]orph[Oo][Ss]:*:*)
	echo ${UNAME_MACHINE}-unknown-morphos
	exit ;;
    *:OS/390:*:*)
	echo i370-ibm-openedition
	exit ;;
    *:z/VM:*:*)
	echo s390-ibm-zvmoe
	exit ;;
    *:OS400:*:*)
	echo powerpc-ibm-os400
	exit ;;
    arm:RISC*:1.[012]*:*|arm:riscix:1.[012]*:*)
	echo arm-acorn-riscix${UNAME_RELEASE}
	exit ;;
    arm*:riscos:*:*|arm*:RISCOS:*:*)
	echo arm-unknown-riscos
	exit ;;
    SR2?01:HI-UX/MPP:*:* | SR8000:HI-UX/MPP:*:*)
	echo hppa1.1-hitachi-hiuxmpp
	exit ;;
    Pyramid*:OSx*:*:* | MIS*:OSx*:*:* | MIS*:SMP_DC-OSx*:*:*)
	# akee@wpdis03.wpafb.af.mil (Earle F. Ake) contributed MIS and NILE.
	if test "`(/bin/universe) 2>/dev/null`" = att ; then
		echo pyramid-pyramid-sysv3
	else
		echo pyramid-pyramid-bsd
	fi
	exit ;;
    NILE*:*:*:dcosx)
	echo pyramid-pyramid-svr4
	exit ;;
    DRS?6000:unix:4.0:6*)
	echo sparc-icl-nx6
	exit ;;
    DRS?6000:UNIX_SV:4.2*:7* | DRS?6000:isis:4.2*:7*)
	case `/usr/bin/uname -p` in
	    sparc) echo sparc-icl-nx7; exit ;;
	esac ;;
    s390x:SunOS:*:*)
	echo ${UNAME_MACHINE}-ibm-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
	exit ;;
    sun4H:SunOS:5.*:*)
	echo sparc-hal-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
	exit ;;
    sun4*:SunOS:5.*:* | tadpole*:SunOS:5.*:*)
	echo sparc-sun-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
	exit ;;
    i86pc:AuroraUX:5.*:* | i86xen:AuroraUX:5.*:*)
	echo i386-pc-auroraux${UNAME_RELEASE}
	exit ;;
    i86pc:SunOS:5.*:* | i86xen:SunOS:5.*:*)
	eval $set_cc_for_build
	SUN_ARCH="i386"
	# If there is a compiler, see if it is configured for 64-bit objects.
	# Note that the Sun cc does not turn __LP64__ into 1 like gcc does.
	# This test works for both compilers.
	if [ "$CC_FOR_BUILD" != 'no_compiler_found' ]; then
	    if (echo '#ifdef __amd64'; echo IS_64BIT_ARCH; echo '#endif') | \
		(CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) | \
		grep IS_64BIT_ARCH >/dev/null
	    then
		SUN_ARCH="x86_64"
	    fi
	fi
	echo ${SUN_ARCH}-pc-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
	exit ;;
    sun4*:SunOS:6*:*)
	# According to config.sub, this is the proper way to canonicalize
	# SunOS6.  Hard to guess exactly what SunOS6 will be like, but
	# it's likely to be more like Solaris than SunOS4.
	echo sparc-sun-solaris3`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
	exit ;;
    sun4*:SunOS:*:*)
	case "`/usr/bin/arch -k`" in
	    Series*|S4*)
		UNAME_RELEASE=`uname -v`
		;;
	esac
	# Japanese Language versions have a version number like `4.1.3-JL'.
	echo sparc-sun-sunos`echo ${UNAME_RELEASE}|sed -e 's/-/_/'`
	exit ;;
    sun3*:SunOS:*:*)
	echo m68k-sun-sunos${UNAME_RELEASE}
	exit ;;
    sun*:*:4.2BSD:*)
	UNAME_RELEASE=`(sed 1q /etc/motd | awk '{print substr($5,1,3)}') 2>/dev/null`
	test "x${UNAME_RELEASE}" = "x" && UNAME_RELEASE=3
	case "`/bin/arch`" in
	    sun3)
		echo m68k-sun-sunos${UNAME_RELEASE}
		;;
	    sun4)
		echo sparc-sun-sunos${UNAME_RELEASE}
		;;
	esac
	exit ;;
    aushp:SunOS:*:*)
	echo sparc-auspex-sunos${UNAME_RELEASE}
	exit ;;
    # The situation for MiNT is a little confusing.  The machine name
    # can be virtually everything (everything which is not
    # "atarist" or "atariste" at least should have a processor
    # > m68000).  The system name ranges from "MiNT" over "FreeMiNT"
    # to the lowercase version "mint" (or "freemint").  Finally
    # the system name "TOS" denotes a system which is actually not
    # MiNT.  But MiNT is downward compatible to TOS, so this should
    # be no problem.
    atarist[e]:*MiNT:*:* | atarist[e]:*mint:*:* | atarist[e]:*TOS:*:*)
	echo m68k-atari-mint${UNAME_RELEASE}
	exit ;;
    atari*:*MiNT:*:* | atari*:*mint:*:* | atarist[e]:*TOS:*:*)
	echo m68k-atari-mint${UNAME_RELEASE}
	exit ;;
    *falcon*:*MiNT:*:* | *falcon*:*mint:*:* | *falcon*:*TOS:*:*)
	echo m68k-atari-mint${UNAME_RELEASE}
	exit ;;
    milan*:*MiNT:*:* | milan*:*mint:*:* | *milan*:*TOS:*:*)
	echo m68k-milan-mint${UNAME_RELEASE}
	exit ;;
    hades*:*MiNT:*:* | hades*:*mint:*:* | *hades*:*TOS:*:*)
	echo m68k-hades-mint${UNAME_RELEASE}
	exit ;;
    *:*MiNT:*:* | *:*mint:*:* | *:*TOS:*:*)
	echo m68k-unknown-mint${UNAME_RELEASE}
	exit ;;
    m68k:machten:*:*)
	echo m68k-apple-machten${UNAME_RELEASE}
	exit ;;
    powerpc:machten:*:*)
	echo powerpc-apple-machten${UNAME_RELEASE}
	exit ;;
    RISC*:Mach:*:*)
	echo mips-dec-mach_bsd4.3
	exit ;;
    RISC*:ULTRIX:*:*)
	echo mips-dec-ultrix${UNAME_RELEASE}
	exit ;;
    VAX*:ULTRIX*:*:*)
	echo vax-dec-ultrix${UNAME_RELEASE}
	exit ;;
    2020:CLIX:*:* | 2430:CLIX:*:*)
	echo clipper-intergraph-clix${UNAME_RELEASE}
	exit ;;
    mips:*:*:UMIPS | mips:*:*:RISCos)
	eval $set_cc_for_build
	sed 's/^	//' << EOF >$dummy.c
#ifdef __cplusplus
#include <stdio.h>  /* for printf() prototype */
	int main (int argc, char *argv[]) {
#else
	int main (argc, argv) int argc; char *argv[]; {
#endif
	#if defined (host_mips) && defined (MIPSEB)
	#if defined (SYSTYPE_SYSV)
	  printf ("mips-mips-riscos%ssysv\n", argv[1]); exit (0);
	#endif
	#if defined (SYSTYPE_SVR4)
	  printf ("mips-mips-riscos%ssvr4\n", argv[1]); exit (0);
	#endif
	#if defined (SYSTYPE_BSD43) || defined(SYSTYPE_BSD)
	  printf ("mips-mips-riscos%sbsd\n", argv[1]); exit (0);
	#endif
	#endif
	  exit (-1);
	}
EOF
	$CC_FOR_BUILD -o $dummy $dummy.c &&
	  dummyarg=`echo "${UNAME_RELEASE}" | sed -n 's/\([0-9]*\).*/\1/p'` &&
	  SYSTEM_NAME=`$dummy $dummyarg` &&
	    { echo "$SYSTEM_NAME"; exit; }
	echo mips-mips-riscos${UNAME_RELEASE}
	exit ;;
    Motorola:PowerMAX_OS:*:*)
	echo powerpc-motorola-powermax
	exit ;;
    Motorola:*:4.3:PL8-*)
	echo powerpc-harris-powermax
	exit ;;
    Night_Hawk:*:*:PowerMAX_OS | Synergy:PowerMAX_OS:*:*)
	echo powerpc-harris-powermax
	exit ;;
    Night_Hawk:Power_UNIX:*:*)
	echo powerpc-harris-powerunix
	exit ;;
    m88k:CX/UX:7*:*)
	echo m88k-harris-cxux7
	exit ;;
    m88k:*:4*:R4*)
	echo m88k-motorola-sysv4
	exit ;;
    m88k:*:3*:R3*)
	echo m88k-motorola-sysv3
	exit ;;
    AViiON:dgux:*:*)
	# DG/UX returns AViiON for all architectures
	UNAME_PROCESSOR=`/usr/bin/uname -p`
	if [ $UNAME_PROCESSOR = mc88100 ] || [ $UNAME_PROCESSOR = mc88110 ]
	then
	    if [ ${TARGET_BINARY_INTERFACE}x = m88kdguxelfx ] || \
	       [ ${TARGET_BINARY_INTERFACE}x = x ]
	    then
		echo m88k-dg-dgux${UNAME_RELEASE}
	    else
		echo m88k-dg-dguxbcs${UNAME_RELEASE}
	    fi
	else
	    echo i586-dg-dgux${UNAME_RELEASE}
	fi
	exit ;;
    M88*:DolphinOS:*:*)	# DolphinOS (SVR3)
	echo m88k-dolphin-sysv3
	exit ;;
    M88*:*:R3*:*)
	# Delta 88k system running SVR3
	echo m88k-motorola-sysv3
	exit ;;
    XD88*:*:*:*) # Tektronix XD88 system running UTekV (SVR3)
	echo m88k-tektronix-sysv3
	exit ;;
    Tek43[0-9][0-9]:UTek:*:*) # Tektronix 4300 system running UTek (BSD)
	echo m68k-tektronix-bsd
	exit ;;
    *:IRIX*:*:*)
	echo mips-sgi-irix`echo ${UNAME_RELEASE}|sed -e 's/-/_/g'`
	exit ;;
    ????????:AIX?:[12].1:2)   # AIX 2.2.1 or AIX 2.1.1 is RT/PC AIX.
	echo romp-ibm-aix     # uname -m gives an 8 hex-code CPU id
	exit ;;               # Note that: echo "'`uname -s`'" gives 'AIX '
    i*86:AIX:*:*)
	echo i386-ibm-aix
	exit ;;
    ia64:AIX:*:*)
	if [ -x /usr/bin/oslevel ] ; then
		IBM_REV=`/usr/bin/oslevel`
	else
		IBM_REV=${UNAME_VERSION}.${UNAME_RELEASE}
	fi
	echo ${UNAME_MACHINE}-ibm-aix${IBM_REV}
	exit ;;
    *:AIX:2:3)
	if grep bos325 /usr/include/stdio.h >/dev/null 2>&1; then
		eval $set_cc_for_build
		sed 's/^		//' << EOF >$dummy.c
		#include <sys/systemcfg.h>

		main()
			{
			if (!__power_pc())
				exit(1);
			puts("powerpc-ibm-aix3.2.5");
			exit(0);
			}
EOF
		if $CC_FOR_BUILD -o $dummy $dummy.c && SYSTEM_NAME=`$dummy`
		then
			echo "$SYSTEM_NAME"
		else
			echo rs6000-ibm-aix3.2.5
		fi
	elif grep bos324 /usr/include/stdio.h >/dev/null 2>&1; then
		echo rs6000-ibm-aix3.2.4
	else
		echo rs6000-ibm-aix3.2
	fi
	exit ;;
    *:AIX:*:[4567])
	IBM_CPU_ID=`/usr/sbin/lsdev -C -c processor -S available | sed 1q | awk '{ print $1 }'`
	if /usr/sbin/lsattr -El ${IBM_CPU_ID} | grep ' POWER' >/dev/null 2>&1; then
		IBM_ARCH=rs6000
	else
		IBM_ARCH=powerpc
	fi
	if [ -x /usr/bin/lslpp ] ; then
		IBM_REV=`/usr/bin/lslpp -Lqc bos.rte.libc |
			   awk -F: '{ print $3 }' | sed s/[0-9]*$/0/`
	else
		IBM_REV=${UNAME_VERSION}.${UNAME_RELEASE}
	fi
	echo ${IBM_ARCH}-ibm-aix${IBM_REV}
	exit ;;
    *:AIX:*:*)
	echo rs6000-ibm-aix
	exit ;;
    ibmrt:4.4BSD:*|romp-ibm:BSD:*)
	echo romp-ibm-bsd4.4
	exit ;;
    ibmrt:*BSD:*|romp-ibm:BSD:*)            # covers RT/PC BSD and
	echo romp-ibm-bsd${UNAME_RELEASE}   # 4.3 with uname added to
	exit ;;                             # report: romp-ibm BSD 4.3
    *:BOSX:*:*)
	echo rs6000-bull-bosx
	exit ;;
    DPX/2?00:B.O.S.:*:*)
	echo m68k-bull-sysv3
	exit ;;
    9000/[34]??:4.3bsd:1.*:*)
	echo m68k-hp-bsd
	exit ;;
    hp300:4.4BSD:*:* | 9000/[34]??:4.3bsd:2.*:*)
	echo m68k-hp-bsd4.4
	exit ;;
    9000/[34678]??:HP-UX:*:*)
	HPUX_REV=`echo ${UNAME_RELEASE}|sed -e 's/[^.]*.[0B]*//'`
	case "${UNAME_MACHINE}" in
	    9000/31? )            HP_ARCH=m68000 ;;
	    9000/[34]?? )         HP_ARCH=m68k ;;
	    9000/[678][0-9][0-9])
		if [ -x /usr/bin/getconf ]; then
		    sc_cpu_version=`/usr/bin/getconf SC_CPU_VERSION 2>/dev/null`
		    sc_kernel_bits=`/usr/bin/getconf SC_KERNEL_BITS 2>/dev/null`
		    case "${sc_cpu_version}" in
		      523) HP_ARCH="hppa1.0" ;; # CPU_PA_RISC1_0
		      528) HP_ARCH="hppa1.1" ;; # CPU_PA_RISC1_1
		      532)                      # CPU_PA_RISC2_0
			case "${sc_kernel_bits}" in
			  32) HP_ARCH="hppa2.0n" ;;
			  64) HP_ARCH="hppa2.0w" ;;
			  '') HP_ARCH="hppa2.0" ;;   # HP-UX 10.20
			esac ;;
		    esac
		fi
		if [ "${HP_ARCH}" = "" ]; then
		    eval $set_cc_for_build
		    sed 's/^		//' << EOF >$dummy.c

		#define _HPUX_SOURCE
		#include <stdlib.h>
		#include <unistd.h>

		int main ()
		{
		#if defined(_SC_KERNEL_BITS)
		    long bits = sysconf(_SC_KERNEL_BITS);
		#endif
		    long cpu  = sysconf (_SC_CPU_VERSION);

		    switch (cpu)
			{
			case CPU_PA_RISC1_0: puts ("hppa1.0"); break;
			case CPU_PA_RISC1_1: puts ("hppa1.1"); break;
			case CPU_PA_RISC2_0:
		#if defined(_SC_KERNEL_BITS)
			    switch (bits)
				{
				case 64: puts ("hppa2.0w"); break;
				case 32: puts ("hppa2.0n"); break;
				default: puts ("hppa2.0"); break;
				} break;
		#else  /* !defined(_SC_KERNEL_BITS) */
			    puts ("hppa2.0"); break;
		#endif
			default: puts ("hppa1.0"); break;
			}
		    exit (0);
		}
EOF
		    (CCOPTS= $CC_FOR_BUILD -o $dummy $dummy.c 2>/dev/null) && HP_ARCH=`$dummy`
		    test -z "$HP_ARCH" && HP_ARCH=hppa
		fi ;;
	esac
	if [ ${HP_ARCH} = "hppa2.0w" ]
	then
	    eval $set_cc_for_build

	    # hppa2.0w-hp-hpux* has a 64-bit kernel and a compiler generating
	    # 32-bit code.  hppa64-hp-hpux* has the same kernel and a compiler
	    # generating 64-bit code.  GNU and HP use different nomenclature:
	    #
	    # $ CC_FOR_BUILD=cc ./config.guess
	    # => hppa2.0w-hp-hpux11.23
	    # $ CC_FOR_BUILD="cc +DA2.0w" ./config.guess
	    # => hppa64-hp-hpux11.23

	    if echo __LP64__ | (CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) |
		grep -q __LP64__
	    then
		HP_ARCH="hppa2.0w"
	    else
		HP_ARCH="hppa64"
	    fi
	fi
	echo ${HP_ARCH}-hp-hpux${HPUX_REV}
	exit ;;
    ia64:HP-UX:*:*)
	HPUX_REV=`echo ${UNAME_RELEASE}|sed -e 's/[^.]*.[0B]*//'`
	echo ia64-hp-hpux${HPUX_REV}
	exit ;;
    3050*:HI-UX:*:*)
	eval $set_cc_for_build
	sed 's/^	//' << EOF >$dummy.c
	#include <unistd.h>
	int
	main ()
	{
	  long cpu = sysconf (_SC_CPU_VERSION);
	  /* The order matters, because CPU_IS_HP_MC68K erroneously returns
	     true for CPU_PA_RISC1_0.  CPU_IS_PA_RISC returns correct
	     results, however.  */
	  if (CPU_IS_PA_RISC (cpu))
	    {
	      switch (cpu)
		{
		  case CPU_PA_RISC1_0: puts ("hppa1.0-hitachi-hiuxwe2"); break;
		  case CPU_PA_RISC1_1: puts ("hppa1.1-hitachi-hiuxwe2"); break;
		  case CPU_PA_RISC2_0: puts ("hppa2.0-hitachi-hiuxwe2"); break;
		  default: puts ("hppa-hitachi-hiuxwe2"); break;
		}
	    }
	  else if (CPU_IS_HP_MC68K (cpu))
	    puts ("m68k-hitachi-hiuxwe2");
	  else puts ("unknown-hitachi-hiuxwe2");
	  exit (0);
	}
EOF
	$CC_FOR_BUILD -o $dummy $dummy.c && SYSTEM_NAME=`$dummy` &&
		{ echo "$SYSTEM_NAME"; exit; }
	echo unknown-hitachi-hiuxwe2
	exit ;;
    9000/7??:4.3bsd:*:* | 9000/8?[79]:4.3bsd:*:* )
	echo hppa1.1-hp-bsd
	exit ;;
    9000/8??:4.3bsd:*:*)
	echo hppa1.0-hp-bsd
	exit ;;
    *9??*:MPE/iX:*:* | *3000*:MPE/iX:*:*)
	echo hppa1.0-hp-mpeix
	exit ;;
    hp7??:OSF1:*:* | hp8?[79]:OSF1:*:* )
	echo hppa1.1-hp-osf
	exit ;;
    hp8??:OSF1:*:*)
	echo hppa1.0-hp-osf
	exit ;;
    i*86:OSF1:*:*)
	if [ -x /usr/sbin/sysversion ] ; then
	    echo ${UNAME_MACHINE}-unknown-osf1mk
	else
	    echo ${UNAME_MACHINE}-unknown-osf1
	fi
	exit ;;
    parisc*:Lites*:*:*)
	echo hppa1.1-hp-lites
	exit ;;
    C1*:ConvexOS:*:* | convex:ConvexOS:C1*:*)
	echo c1-convex-bsd
	exit ;;
    C2*:ConvexOS:*:* | convex:ConvexOS:C2*:*)
	if getsysinfo -f scalar_acc
	then echo c32-convex-bsd
	else echo c2-convex-bsd
	fi
	exit ;;
    C34*:ConvexOS:*:* | convex:ConvexOS:C34*:*)
	echo c34-convex-bsd
	exit ;;
    C38*:ConvexOS:*:* | convex:ConvexOS:C38*:*)
	echo c38-convex-bsd
	exit ;;
    C4*:ConvexOS:*:* | convex:ConvexOS:C4*:*)
	echo c4-convex-bsd
	exit ;;
    CRAY*Y-MP:*:*:*)
	echo ymp-cray-unicos${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
	exit ;;
    CRAY*[A-Z]90:*:*:*)
	echo ${UNAME_MACHINE}-cray-unicos${UNAME_RELEASE} \
	| sed -e 's/CRAY.*\([A-Z]90\)/\1/' \
	      -e y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/ \
	      -e 's/\.[^.]*$/.X/'
	exit ;;
    CRAY*TS:*:*:*)
	echo t90-cray-unicos${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
	exit ;;
    CRAY*T3E:*:*:*)
	echo alphaev5-cray-unicosmk${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
	exit ;;
    CRAY*SV1:*:*:*)
	echo sv1-cray-unicos${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
	exit ;;
    *:UNICOS/mp:*:*)
	echo craynv-cray-unicosmp${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
	exit ;;
    F30[01]:UNIX_System_V:*:* | F700:UNIX_System_V:*:*)
	FUJITSU_PROC=`uname -m | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz'`
	FUJITSU_SYS=`uname -p | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/\///'`
	FUJITSU_REL=`echo ${UNAME_RELEASE} | sed -e 's/ /_/'`
	echo "${FUJITSU_PROC}-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
	exit ;;
    5000:UNIX_System_V:4.*:*)
	FUJITSU_SYS=`uname -p | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/\///'`
	FUJITSU_REL=`echo ${UNAME_RELEASE} | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/ /_/'`
	echo "sparc-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
	exit ;;
    i*86:BSD/386:*:* | i*86:BSD/OS:*:* | *:Ascend\ Embedded/OS:*:*)
	echo ${UNAME_MACHINE}-pc-bsdi${UNAME_RELEASE}
	exit ;;
    sparc*:BSD/OS:*:*)
	echo sparc-unknown-bsdi${UNAME_RELEASE}
	exit ;;
    *:BSD/OS:*:*)
	echo ${UNAME_MACHINE}-unknown-bsdi${UNAME_RELEASE}
	exit ;;
    *:FreeBSD:*:*)
	UNAME_PROCESSOR=`/usr/bin/uname -p`
	case ${UNAME_PROCESSOR} in
	    amd64)
		echo x86_64-unknown-freebsd`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'` ;;
	    *)
		echo ${UNAME_PROCESSOR}-unknown-freebsd`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'` ;;
	esac
	exit ;;
    i*:CYGWIN*:*)
	echo ${UNAME_MACHINE}-pc-cygwin
	exit ;;
    *:MINGW64*:*)
	echo ${UNAME_MACHINE}-pc-mingw64
	exit ;;
    *:MINGW*:*)
	echo ${UNAME_MACHINE}-pc-mingw32
	exit ;;
    *:MSYS*:*)
	echo ${UNAME_MACHINE}-pc-msys
	exit ;;
    i*:windows32*:*)
	# uname -m includes "-pc" on this system.
	echo ${UNAME_MACHINE}-mingw32
	exit ;;
    i*:PW*:*)
	echo ${UNAME_MACHINE}-pc-pw32
	exit ;;
    *:Interix*:*)
	case ${UNAME_MACHINE} in
	    x86)
		echo i586-pc-interix${UNAME_RELEASE}
		exit ;;
	    authenticamd | genuineintel | EM64T)
		echo x86_64-unknown-interix${UNAME_RELEASE}
		exit ;;
	    IA64)
		echo ia64-unknown-interix${UNAME_RELEASE}
		exit ;;
	esac ;;
    [345]86:Windows_95:* | [345]86:Windows_98:* | [345]86:Windows_NT:*)
	echo i${UNAME_MACHINE}-pc-mks
	exit ;;
    8664:Windows_NT:*)
	echo x86_64-pc-mks
	exit ;;
    i*:Windows_NT*:* | Pentium*:Windows_NT*:*)
	# How do we know it's Interix rather than the generic POSIX subsystem?
	# It also conflicts with pre-2.0 versions of AT&T UWIN. Should we
	# UNAME_MACHINE based on the output of uname instead of i386?
	echo i586-pc-interix
	exit ;;
    i*:UWIN*:*)
	echo ${UNAME_MACHINE}-pc-uwin
	exit ;;
    amd64:CYGWIN*:*:* | x86_64:CYGWIN*:*:*)
	echo x86_64-unknown-cygwin
	exit ;;
    p*:CYGWIN*:*)
	echo powerpcle-unknown-cygwin
	exit ;;
    prep*:SunOS:5.*:*)
	echo powerpcle-unknown-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
	exit ;;
    *:GNU:*:*)
	# the GNU system
	echo `echo ${UNAME_MACHINE}|sed -e 's,[-/].*$,,'`-unknown-${LIBC}`echo ${UNAME_RELEASE}|sed -e 's,/.*$,,'`
	exit ;;
    *:GNU/*:*:*)
	# other systems with GNU libc and userland
	echo ${UNAME_MACHINE}-unknown-`echo ${UNAME_SYSTEM} | sed 's,^[^/]*/,,' | tr '[A-Z]' '[a-z]'``echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`-${LIBC}
	exit ;;
    i*86:Minix:*:*)
	echo ${UNAME_MACHINE}-pc-minix
	exit ;;
    aarch64:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    aarch64_be:Linux:*:*)
	UNAME_MACHINE=aarch64_be
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    alpha:Linux:*:*)
	case `sed -n '/^cpu model/s/^.*: \(.*\)/\1/p' < /proc/cpuinfo` in
	  EV5)   UNAME_MACHINE=alphaev5 ;;
	  EV56)  UNAME_MACHINE=alphaev56 ;;
	  PCA56) UNAME_MACHINE=alphapca56 ;;
	  PCA57) UNAME_MACHINE=alphapca56 ;;
	  EV6)   UNAME_MACHINE=alphaev6 ;;
	  EV67)  UNAME_MACHINE=alphaev67 ;;
	  EV68*) UNAME_MACHINE=alphaev68 ;;
	esac
	objdump --private-headers /bin/sh | grep -q ld.so.1
	if test "$?" = 0 ; then LIBC="gnulibc1" ; fi
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    arc:Linux:*:* | arceb:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    arm*:Linux:*:*)
	eval $set_cc_for_build
	if echo __ARM_EABI__ | $CC_FOR_BUILD -E - 2>/dev/null \
	    | grep -q __ARM_EABI__
	then
	    echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	else
	    if echo __ARM_PCS_VFP | $CC_FOR_BUILD -E - 2>/dev/null \
		| grep -q __ARM_PCS_VFP
	    then
		echo ${UNAME_MACHINE}-unknown-linux-${LIBC}eabi
	    else
		echo ${UNAME_MACHINE}-unknown-linux-${LIBC}eabihf
	    fi
	fi
	exit ;;
    avr32*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    cris:Linux:*:*)
	echo ${UNAME_MACHINE}-axis-linux-${LIBC}
	exit ;;
    crisv32:Linux:*:*)
	echo ${UNAME_MACHINE}-axis-linux-${LIBC}
	exit ;;
    frv:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    hexagon:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    i*86:Linux:*:*)
	echo ${UNAME_MACHINE}-pc-linux-${LIBC}
	exit ;;
    ia64:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    m32r*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    m68*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    mips:Linux:*:* | mips64:Linux:*:*)
	eval $set_cc_for_build
	sed 's/^	//' << EOF >$dummy.c
	#undef CPU
	#undef ${UNAME_MACHINE}
	#undef ${UNAME_MACHINE}el
	#if defined(__MIPSEL__) || defined(__MIPSEL) || defined(_MIPSEL) || defined(MIPSEL)
	CPU=${UNAME_MACHINE}el
	#else
	#if defined(__MIPSEB__) || defined(__MIPSEB) || defined(_MIPSEB) || defined(MIPSEB)
	CPU=${UNAME_MACHINE}
	#else
	CPU=
	#endif
	#endif
EOF
	eval `$CC_FOR_BUILD -E $dummy.c 2>/dev/null | grep '^CPU'`
	test x"${CPU}" != x && { echo "${CPU}-unknown-linux-${LIBC}"; exit; }
	;;
    openrisc*:Linux:*:*)
	echo or1k-unknown-linux-${LIBC}
	exit ;;
    or32:Linux:*:* | or1k*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    padre:Linux:*:*)
	echo sparc-unknown-linux-${LIBC}
	exit ;;
    parisc64:Linux:*:* | hppa64:Linux:*:*)
	echo hppa64-unknown-linux-${LIBC}
	exit ;;
    parisc:Linux:*:* | hppa:Linux:*:*)
	# Look for CPU level
	case `grep '^cpu[^a-z]*:' /proc/cpuinfo 2>/dev/null | cut -d' ' -f2` in
	  PA7*) echo hppa1.1-unknown-linux-${LIBC} ;;
	  PA8*) echo hppa2.0-unknown-linux-${LIBC} ;;
	  *)    echo hppa-unknown-linux-${LIBC} ;;
	esac
	exit ;;
    ppc64:Linux:*:*)
	echo powerpc64-unknown-linux-${LIBC}
	exit ;;
    ppc:Linux:*:*)
	echo powerpc-unknown-linux-${LIBC}
	exit ;;
    ppc64le:Linux:*:*)
	echo powerpc64le-unknown-linux-${LIBC}
	exit ;;
    ppcle:Linux:*:*)
	echo powerpcle-unknown-linux-${LIBC}
	exit ;;
    s390:Linux:*:* | s390x:Linux:*:*)
	echo ${UNAME_MACHINE}-ibm-linux-${LIBC}
	exit ;;
    sh64*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    sh*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    sparc:Linux:*:* | sparc64:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    tile*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    vax:Linux:*:*)
	echo ${UNAME_MACHINE}-dec-linux-${LIBC}
	exit ;;
    x86_64:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    xtensa*:Linux:*:*)
	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
	exit ;;
    i*86:DYNIX/ptx:4*:*)
	# ptx 4.0 does uname -s correctly, with DYNIX/ptx in there.
	# earlier versions are messed up and put the nodename in both
	# sysname and nodename.
	echo i386-sequent-sysv4
	exit ;;
    i*86:UNIX_SV:4.2MP:2.*)
	# Unixware is an offshoot of SVR4, but it has its own version
	# number series starting with 2...
	# I am not positive that other SVR4 systems won't match this,
	# I just have to hope.  -- rms.
	# Use sysv4.2uw... so that sysv4* matches it.
	echo ${UNAME_MACHINE}-pc-sysv4.2uw${UNAME_VERSION}
	exit ;;
    i*86:OS/2:*:*)
	# If we were able to find `uname', then EMX Unix compatibility
	# is probably installed.
	echo ${UNAME_MACHINE}-pc-os2-emx
	exit ;;
    i*86:XTS-300:*:STOP)
	echo ${UNAME_MACHINE}-unknown-stop
	exit ;;
    i*86:atheos:*:*)
	echo ${UNAME_MACHINE}-unknown-atheos
	exit ;;
    i*86:syllable:*:*)
	echo ${UNAME_MACHINE}-pc-syllable
	exit ;;
    i*86:LynxOS:2.*:* | i*86:LynxOS:3.[01]*:* | i*86:LynxOS:4.[02]*:*)
	echo i386-unknown-lynxos${UNAME_RELEASE}
	exit ;;
    i*86:*DOS:*:*)
	echo ${UNAME_MACHINE}-pc-msdosdjgpp
	exit ;;
    i*86:*:4.*:* | i*86:SYSTEM_V:4.*:*)
	UNAME_REL=`echo ${UNAME_RELEASE} | sed 's/\/MP$//'`
	if grep Novell /usr/include/link.h >/dev/null 2>/dev/null; then
		echo ${UNAME_MACHINE}-univel-sysv${UNAME_REL}
	else
		echo ${UNAME_MACHINE}-pc-sysv${UNAME_REL}
	fi
	exit ;;
    i*86:*:5:[678]*)
	# UnixWare 7.x, OpenUNIX and OpenServer 6.
	case `/bin/uname -X | grep "^Machine"` in
	    *486*)	     UNAME_MACHINE=i486 ;;
	    *Pentium)	     UNAME_MACHINE=i586 ;;
	    *Pent*|*Celeron) UNAME_MACHINE=i686 ;;
	esac
	echo ${UNAME_MACHINE}-unknown-sysv${UNAME_RELEASE}${UNAME_SYSTEM}${UNAME_VERSION}
	exit ;;
    i*86:*:3.2:*)
	if test -f /usr/options/cb.name; then
		UNAME_REL=`sed -n 's/.*Version //p' </usr/options/cb.name`
		echo ${UNAME_MACHINE}-pc-isc$UNAME_REL
	elif /bin/uname -X 2>/dev/null >/dev/null ; then
		UNAME_REL=`(/bin/uname -X|grep Release|sed -e 's/.*= //')`
		(/bin/uname -X|grep i80486 >/dev/null) && UNAME_MACHINE=i486
		(/bin/uname -X|grep '^Machine.*Pentium' >/dev/null) \
			&& UNAME_MACHINE=i586
		(/bin/uname -X|grep '^Machine.*Pent *II' >/dev/null) \
			&& UNAME_MACHINE=i686
		(/bin/uname -X|grep '^Machine.*Pentium Pro' >/dev/null) \
			&& UNAME_MACHINE=i686
		echo ${UNAME_MACHINE}-pc-sco$UNAME_REL
	else
		echo ${UNAME_MACHINE}-pc-sysv32
	fi
	exit ;;
    pc:*:*:*)
	# Left here for compatibility:
	# uname -m prints for DJGPP always 'pc', but it prints nothing about
	# the processor, so we play safe by assuming i586.
	# Note: whatever this is, it MUST be the same as what config.sub
	# prints for the "djgpp" host, or else GDB configury will decide that
	# this is a cross-build.
	echo i586-pc-msdosdjgpp
	exit ;;
    Intel:Mach:3*:*)
	echo i386-pc-mach3
	exit ;;
    paragon:*:*:*)
	echo i860-intel-osf1
	exit ;;
    i860:*:4.*:*) # i860-SVR4
	if grep Stardent /usr/include/sys/uadmin.h >/dev/null 2>&1 ; then
	  echo i860-stardent-sysv${UNAME_RELEASE} # Stardent Vistra i860-SVR4
	else # Add other i860-SVR4 vendors below as they are discovered.
	  echo i860-unknown-sysv${UNAME_RELEASE}  # Unknown i860-SVR4
	fi
	exit ;;
    mini*:CTIX:SYS*5:*)
	# "miniframe"
	echo m68010-convergent-sysv
	exit ;;
    mc68k:UNIX:SYSTEM5:3.51m)
	echo m68k-convergent-sysv
	exit ;;
    M680?0:D-NIX:5.3:*)
	echo m68k-diab-dnix
	exit ;;
    M68*:*:R3V[5678]*:*)
	test -r /sysV68 && { echo 'm68k-motorola-sysv'; exit; } ;;
    3[345]??:*:4.0:3.0 | 3[34]??A:*:4.0:3.0 | 3[34]??,*:*:4.0:3.0 | 3[34]??/*:*:4.0:3.0 | 4400:*:4.0:3.0 | 4850:*:4.0:3.0 | SKA40:*:4.0:3.0 | SDS2:*:4.0:3.0 | SHG2:*:4.0:3.0 | S7501*:*:4.0:3.0)
	OS_REL=''
	test -r /etc/.relid \
	&& OS_REL=.`sed -n 's/[^ ]* [^ ]* \([0-9][0-9]\).*/\1/p' < /etc/.relid`
	/bin/uname -p 2>/dev/null | grep 86 >/dev/null \
	  && { echo i486-ncr-sysv4.3${OS_REL}; exit; }
	/bin/uname -p 2>/dev/null | /bin/grep entium >/dev/null \
	  && { echo i586-ncr-sysv4.3${OS_REL}; exit; } ;;
    3[34]??:*:4.0:* | 3[34]??,*:*:4.0:*)
	/bin/uname -p 2>/dev/null | grep 86 >/dev/null \
	  && { echo i486-ncr-sysv4; exit; } ;;
    NCR*:*:4.2:* | MPRAS*:*:4.2:*)
	OS_REL='.3'
	test -r /etc/.relid \
	    && OS_REL=.`sed -n 's/[^ ]* [^ ]* \([0-9][0-9]\).*/\1/p' < /etc/.relid`
	/bin/uname -p 2>/dev/null | grep 86 >/dev/null \
	    && { echo i486-ncr-sysv4.3${OS_REL}; exit; }
	/bin/uname -p 2>/dev/null | /bin/grep entium >/dev/null \
	    && { echo i586-ncr-sysv4.3${OS_REL}; exit; }
	/bin/uname -p 2>/dev/null | /bin/grep pteron >/dev/null \
	    && { echo i586-ncr-sysv4.3${OS_REL}; exit; } ;;
    m68*:LynxOS:2.*:* | m68*:LynxOS:3.0*:*)
	echo m68k-unknown-lynxos${UNAME_RELEASE}
	exit ;;
    mc68030:UNIX_System_V:4.*:*)
	echo m68k-atari-sysv4
	exit ;;
    TSUNAMI:LynxOS:2.*:*)
	echo sparc-unknown-lynxos${UNAME_RELEASE}
	exit ;;
    rs6000:LynxOS:2.*:*)
	echo rs6000-unknown-lynxos${UNAME_RELEASE}
	exit ;;
    PowerPC:LynxOS:2.*:* | PowerPC:LynxOS:3.[01]*:* | PowerPC:LynxOS:4.[02]*:*)
	echo powerpc-unknown-lynxos${UNAME_RELEASE}
	exit ;;
    SM[BE]S:UNIX_SV:*:*)
	echo mips-dde-sysv${UNAME_RELEASE}
	exit ;;
    RM*:ReliantUNIX-*:*:*)
	echo mips-sni-sysv4
	exit ;;
    RM*:SINIX-*:*:*)
	echo mips-sni-sysv4
	exit ;;
    *:SINIX-*:*:*)
	if uname -p 2>/dev/null >/dev/null ; then
		UNAME_MACHINE=`(uname -p) 2>/dev/null`
		echo ${UNAME_MACHINE}-sni-sysv4
	else
		echo ns32k-sni-sysv
	fi
	exit ;;
    PENTIUM:*:4.0*:*)	# Unisys `ClearPath HMP IX 4000' SVR4/MP effort
			# says <Richard.M.Bartel@ccMail.Census.GOV>
	echo i586-unisys-sysv4
	exit ;;
    *:UNIX_System_V:4*:FTX*)
	# From Gerald Hewes <hewes@openmarket.com>.
	# How about differentiating between stratus architectures? -djm
	echo hppa1.1-stratus-sysv4
	exit ;;
    *:*:*:FTX*)
	# From seanf@swdc.stratus.com.
	echo i860-stratus-sysv4
	exit ;;
    i*86:VOS:*:*)
	# From Paul.Green@stratus.com.
	echo ${UNAME_MACHINE}-stratus-vos
	exit ;;
    *:VOS:*:*)
	# From Paul.Green@stratus.com.
	echo hppa1.1-stratus-vos
	exit ;;
    mc68*:A/UX:*:*)
	echo m68k-apple-aux${UNAME_RELEASE}
	exit ;;
    news*:NEWS-OS:6*:*)
	echo mips-sony-newsos6
	exit ;;
    R[34]000:*System_V*:*:* | R4000:UNIX_SYSV:*:* | R*000:UNIX_SV:*:*)
	if [ -d /usr/nec ]; then
		echo mips-nec-sysv${UNAME_RELEASE}
	else
		echo mips-unknown-sysv${UNAME_RELEASE}
	fi
	exit ;;
    BeBox:BeOS:*:*)	# BeOS running on hardware made by Be, PPC only.
	echo powerpc-be-beos
	exit ;;
    BeMac:BeOS:*:*)	# BeOS running on Mac or Mac clone, PPC only.
	echo powerpc-apple-beos
	exit ;;
    BePC:BeOS:*:*)	# BeOS running on Intel PC compatible.
	echo i586-pc-beos
	exit ;;
    BePC:Haiku:*:*)	# Haiku running on Intel PC compatible.
	echo i586-pc-haiku
	exit ;;
    x86_64:Haiku:*:*)
	echo x86_64-unknown-haiku
	exit ;;
    SX-4:SUPER-UX:*:*)
	echo sx4-nec-superux${UNAME_RELEASE}
	exit ;;
    SX-5:SUPER-UX:*:*)
	echo sx5-nec-superux${UNAME_RELEASE}
	exit ;;
    SX-6:SUPER-UX:*:*)
	echo sx6-nec-superux${UNAME_RELEASE}
	exit ;;
    SX-7:SUPER-UX:*:*)
	echo sx7-nec-superux${UNAME_RELEASE}
	exit ;;
    SX-8:SUPER-UX:*:*)
	echo sx8-nec-superux${UNAME_RELEASE}
	exit ;;
    SX-8R:SUPER-UX:*:*)
	echo sx8r-nec-superux${UNAME_RELEASE}
	exit ;;
    Power*:Rhapsody:*:*)
	echo powerpc-apple-rhapsody${UNAME_RELEASE}
	exit ;;
    *:Rhapsody:*:*)
	echo ${UNAME_MACHINE}-apple-rhapsody${UNAME_RELEASE}
	exit ;;
    *:Darwin:*:*)
	UNAME_PROCESSOR=`uname -p` || UNAME_PROCESSOR=unknown
	eval $set_cc_for_build
	if test "$UNAME_PROCESSOR" = unknown ; then
	    UNAME_PROCESSOR=powerpc
	fi
	if test `echo "$UNAME_RELEASE" | sed -e 's/\..*//'` -le 10 ; then
	    if [ "$CC_FOR_BUILD" != 'no_compiler_found' ]; then
		if (echo '#ifdef __LP64__'; echo IS_64BIT_ARCH; echo '#endif') | \
		    (CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) | \
		    grep IS_64BIT_ARCH >/dev/null
		then
		    case $UNAME_PROCESSOR in
			i386) UNAME_PROCESSOR=x86_64 ;;
			powerpc) UNAME_PROCESSOR=powerpc64 ;;
		    esac
		fi
	    fi
	elif test "$UNAME_PROCESSOR" = i386 ; then
	    # Avoid executing cc on OS X 10.9, as it ships with a stub
	    # that puts up a graphical alert prompting to install
	    # developer tools.  Any system running Mac OS X 10.7 or
	    # later (Darwin 11 and later) is required to have a 64-bit
	    # processor. This is not true of the ARM version of Darwin
	    # that Apple uses in portable devices.
	    UNAME_PROCESSOR=x86_64
	fi
	echo ${UNAME_PROCESSOR}-apple-darwin${UNAME_RELEASE}
	exit ;;
    *:procnto*:*:* | *:QNX:[0123456789]*:*)
	UNAME_PROCESSOR=`uname -p`
	if test "$UNAME_PROCESSOR" = "x86"; then
		UNAME_PROCESSOR=i386
		UNAME_MACHINE=pc
	fi
	echo ${UNAME_PROCESSOR}-${UNAME_MACHINE}-nto-qnx${UNAME_RELEASE}
	exit ;;
    *:QNX:*:4*)
	echo i386-pc-qnx
	exit ;;
    NEO-?:NONSTOP_KERNEL:*:*)
	echo neo-tandem-nsk${UNAME_RELEASE}
	exit ;;
    NSE-*:NONSTOP_KERNEL:*:*)
	echo nse-tandem-nsk${UNAME_RELEASE}
	exit ;;
    NSR-?:NONSTOP_KERNEL:*:*)
	echo nsr-tandem-nsk${UNAME_RELEASE}
	exit ;;
    *:NonStop-UX:*:*)
	echo mips-compaq-nonstopux
	exit ;;
    BS2000:POSIX*:*:*)
	echo bs2000-siemens-sysv
	exit ;;
    DS/*:UNIX_System_V:*:*)
	echo ${UNAME_MACHINE}-${UNAME_SYSTEM}-${UNAME_RELEASE}
	exit ;;
    *:Plan9:*:*)
	# "uname -m" is not consistent, so use $cputype instead. 386
	# is converted to i386 for consistency with other x86
	# operating systems.
	if test "$cputype" = "386"; then
	    UNAME_MACHINE=i386
	else
	    UNAME_MACHINE="$cputype"
	fi
	echo ${UNAME_MACHINE}-unknown-plan9
	exit ;;
    *:TOPS-10:*:*)
	echo pdp10-unknown-tops10
	exit ;;
    *:TENEX:*:*)
	echo pdp10-unknown-tenex
	exit ;;
    KS10:TOPS-20:*:* | KL10:TOPS-20:*:* | TYPE4:TOPS-20:*:*)
	echo pdp10-dec-tops20
	exit ;;
    XKL-1:TOPS-20:*:* | TYPE5:TOPS-20:*:*)
	echo pdp10-xkl-tops20
	exit ;;
    *:TOPS-20:*:*)
	echo pdp10-unknown-tops20
	exit ;;
    *:ITS:*:*)
	echo pdp10-unknown-its
	exit ;;
    SEI:*:*:SEIUX)
	echo mips-sei-seiux${UNAME_RELEASE}
	exit ;;
    *:DragonFly:*:*)
	echo ${UNAME_MACHINE}-unknown-dragonfly`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`
	exit ;;
    *:*VMS:*:*)
	UNAME_MACHINE=`(uname -p) 2>/dev/null`
	case "${UNAME_MACHINE}" in
	    A*) echo alpha-dec-vms ; exit ;;
	    I*) echo ia64-dec-vms ; exit ;;
	    V*) echo vax-dec-vms ; exit ;;
	esac ;;
    *:XENIX:*:SysV)
	echo i386-pc-xenix
	exit ;;
    i*86:skyos:*:*)
	echo ${UNAME_MACHINE}-pc-skyos`echo ${UNAME_RELEASE}` | sed -e 's/ .*$//'
	exit ;;
    i*86:rdos:*:*)
	echo ${UNAME_MACHINE}-pc-rdos
	exit ;;
    i*86:AROS:*:*)
	echo ${UNAME_MACHINE}-pc-aros
	exit ;;
    x86_64:VMkernel:*:*)
	echo ${UNAME_MACHINE}-unknown-esx
	exit ;;
esac

cat >&2 <<EOF
$0: unable to guess system type

This script, last modified $timestamp, has failed to recognize
the operating system you are using. It is advised that you
download the most up to date version of the config scripts from

  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD
and
  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD

If the version you run ($0) is already up to date, please
send the following data and any information you think might be
pertinent to <config-patches@gnu.org> in order to provide the needed
information to handle your system.

config.guess timestamp = $timestamp

uname -m = `(uname -m) 2>/dev/null || echo unknown`
uname -r = `(uname -r) 2>/dev/null || echo unknown`
uname -s = `(uname -s) 2>/dev/null || echo unknown`
uname -v = `(uname -v) 2>/dev/null || echo unknown`

/usr/bin/uname -p = `(/usr/bin/uname -p) 2>/dev/null`
/bin/uname -X     = `(/bin/uname -X) 2>/dev/null`

hostinfo               = `(hostinfo) 2>/dev/null`
/bin/universe          = `(/bin/universe) 2>/dev/null`
/usr/bin/arch -k       = `(/usr/bin/arch -k) 2>/dev/null`
/bin/arch              = `(/bin/arch) 2>/dev/null`
/usr/bin/oslevel       = `(/usr/bin/oslevel) 2>/dev/null`
/usr/convex/getsysinfo = `(/usr/convex/getsysinfo) 2>/dev/null`

UNAME_MACHINE = ${UNAME_MACHINE}
UNAME_RELEASE = ${UNAME_RELEASE}
UNAME_SYSTEM  = ${UNAME_SYSTEM}
UNAME_VERSION = ${UNAME_VERSION}
EOF

exit 1

# Local variables:
# eval: (add-hook 'write-file-hooks 'time-stamp)
# time-stamp-start: "timestamp='"
# time-stamp-format: "%:y-%02m-%02d"
# time-stamp-end: "'"
# End:

#! /bin/sh
# Configuration validation subroutine script.
#   Copyright 1992-2014 Free Software Foundation, Inc.

timestamp='2014-12-03'

# This file is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
# General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, see <http://www.gnu.org/licenses/>.
#
# As a special exception to the GNU General Public License, if you
# distribute this file as part of a program that contains a
# configuration script generated by Autoconf, you may include it under
# the same distribution terms that you use for the rest of that
# program.  This Exception is an additional permission under section 7
# of the GNU General Public License, version 3 ("GPLv3").


# Please send patches to <config-patches@gnu.org>.
#
# Configuration subroutine to validate and canonicalize a configuration type.
# Supply the specified configuration type as an argument.
# If it is invalid, we print an error message on stderr and exit with code 1.
# Otherwise, we print the canonical config type on stdout and succeed.

# You can get the latest version of this script from:
# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD

# This file is supposed to be the same for all GNU packages
# and recognize all the CPU types, system types and aliases
# that are meaningful with *any* GNU software.
# Each package is responsible for reporting which valid configurations
# it does not support.  The user should be able to distinguish
# a failure to support a valid configuration from a meaningless
# configuration.

# The goal of this file is to map all the various variations of a given
# machine specification into a single specification in the form:
#	CPU_TYPE-MANUFACTURER-OPERATING_SYSTEM
# or in some cases, the newer four-part form:
#	CPU_TYPE-MANUFACTURER-KERNEL-OPERATING_SYSTEM
# It is wrong to echo any other type of specification.

me=`echo "$0" | sed -e 's,.*/,,'`

usage="\
Usage: $0 [OPTION] CPU-MFR-OPSYS
       $0 [OPTION] ALIAS

Canonicalize a configuration name.

Operation modes:
  -h, --help         print this help, then exit
  -t, --time-stamp   print date of last modification, then exit
  -v, --version      print version number, then exit

Report bugs and patches to <config-patches@gnu.org>."

version="\
GNU config.sub ($timestamp)

Copyright 1992-2014 Free Software Foundation, Inc.

This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."

help="
Try \`$me --help' for more information."

# Parse command line
while test $# -gt 0 ; do
  case $1 in
    --time-stamp | --time* | -t )
       echo "$timestamp" ; exit ;;
    --version | -v )
       echo "$version" ; exit ;;
    --help | --h* | -h )
       echo "$usage"; exit ;;
    -- )     # Stop option processing
       shift; break ;;
    - )	# Use stdin as input.
       break ;;
    -* )
       echo "$me: invalid option $1$help"
       exit 1 ;;

    *local*)
       # First pass through any local machine types.
       echo $1
       exit ;;

    * )
       break ;;
  esac
done

case $# in
 0) echo "$me: missing argument$help" >&2
    exit 1;;
 1) ;;
 *) echo "$me: too many arguments$help" >&2
    exit 1;;
esac

# Separate what the user gave into CPU-COMPANY and OS or KERNEL-OS (if any).
# Here we must recognize all the valid KERNEL-OS combinations.
maybe_os=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
case $maybe_os in
  nto-qnx* | linux-gnu* | linux-android* | linux-dietlibc | linux-newlib* | \
  linux-musl* | linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | \
  knetbsd*-gnu* | netbsd*-gnu* | \
  kopensolaris*-gnu* | \
  storm-chaos* | os2-emx* | rtmk-nova*)
    os=-$maybe_os
    basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`
    ;;
  android-linux)
    os=-linux-android
    basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`-unknown
    ;;
  *)
    basic_machine=`echo $1 | sed 's/-[^-]*$//'`
    if [ $basic_machine != $1 ]
    then os=`echo $1 | sed 's/.*-/-/'`
    else os=; fi
    ;;
esac

### Let's recognize common machines as not being operating systems so
### that things like config.sub decstation-3100 work.  We also
### recognize some manufacturers as not being operating systems, so we
### can provide default operating systems below.
case $os in
	-sun*os*)
		# Prevent following clause from handling this invalid input.
		;;
	-dec* | -mips* | -sequent* | -encore* | -pc532* | -sgi* | -sony* | \
	-att* | -7300* | -3300* | -delta* | -motorola* | -sun[234]* | \
	-unicom* | -ibm* | -next | -hp | -isi* | -apollo | -altos* | \
	-convergent* | -ncr* | -news | -32* | -3600* | -3100* | -hitachi* |\
	-c[123]* | -convex* | -sun | -crds | -omron* | -dg | -ultra | -tti* | \
	-harris | -dolphin | -highlevel | -gould | -cbm | -ns | -masscomp | \
	-apple | -axis | -knuth | -cray | -microblaze*)
		os=
		basic_machine=$1
		;;
	-bluegene*)
		os=-cnk
		;;
	-sim | -cisco | -oki | -wec | -winbond)
		os=
		basic_machine=$1
		;;
	-scout)
		;;
	-wrs)
		os=-vxworks
		basic_machine=$1
		;;
	-chorusos*)
		os=-chorusos
		basic_machine=$1
		;;
	-chorusrdb)
		os=-chorusrdb
		basic_machine=$1
		;;
	-hiux*)
		os=-hiuxwe2
		;;
	-sco6)
		os=-sco5v6
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-sco5)
		os=-sco3.2v5
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-sco4)
		os=-sco3.2v4
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-sco3.2.[4-9]*)
		os=`echo $os | sed -e 's/sco3.2./sco3.2v/'`
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-sco3.2v[4-9]*)
		# Don't forget version if it is 3.2v4 or newer.
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-sco5v6*)
		# Don't forget version if it is 3.2v4 or newer.
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-sco*)
		os=-sco3.2v2
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-udk*)
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-isc)
		os=-isc2.2
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-clix*)
		basic_machine=clipper-intergraph
		;;
	-isc*)
		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
		;;
	-lynx*178)
		os=-lynxos178
		;;
	-lynx*5)
		os=-lynxos5
		;;
	-lynx*)
		os=-lynxos
		;;
	-ptx*)
		basic_machine=`echo $1 | sed -e 's/86-.*/86-sequent/'`
		;;
	-windowsnt*)
		os=`echo $os | sed -e 's/windowsnt/winnt/'`
		;;
	-psos*)
		os=-psos
		;;
	-mint | -mint[0-9]*)
		basic_machine=m68k-atari
		os=-mint
		;;
esac

# Decode aliases for certain CPU-COMPANY combinations.
case $basic_machine in
	# Recognize the basic CPU types without company name.
	# Some are omitted here because they have special meanings below.
	1750a | 580 \
	| a29k \
	| aarch64 | aarch64_be \
	| alpha | alphaev[4-8] | alphaev56 | alphaev6[78] | alphapca5[67] \
	| alpha64 | alpha64ev[4-8] | alpha64ev56 | alpha64ev6[78] | alpha64pca5[67] \
	| am33_2.0 \
	| arc | arceb \
	| arm | arm[bl]e | arme[lb] | armv[2-8] | armv[3-8][lb] | armv7[arm] \
	| avr | avr32 \
	| be32 | be64 \
	| bfin \
	| c4x | c8051 | clipper \
	| d10v | d30v | dlx | dsp16xx \
	| epiphany \
	| fido | fr30 | frv \
	| h8300 | h8500 | hppa | hppa1.[01] | hppa2.0 | hppa2.0[nw] | hppa64 \
	| hexagon \
	| i370 | i860 | i960 | ia64 \
	| ip2k | iq2000 \
	| k1om \
	| le32 | le64 \
	| lm32 \
	| m32c | m32r | m32rle | m68000 | m68k | m88k \
	| maxq | mb | microblaze | microblazeel | mcore | mep | metag \
	| mips | mipsbe | mipseb | mipsel | mipsle \
	| mips16 \
	| mips64 | mips64el \
	| mips64octeon | mips64octeonel \
	| mips64orion | mips64orionel \
	| mips64r5900 | mips64r5900el \
	| mips64vr | mips64vrel \
	| mips64vr4100 | mips64vr4100el \
	| mips64vr4300 | mips64vr4300el \
	| mips64vr5000 | mips64vr5000el \
	| mips64vr5900 | mips64vr5900el \
	| mipsisa32 | mipsisa32el \
	| mipsisa32r2 | mipsisa32r2el \
	| mipsisa32r6 | mipsisa32r6el \
	| mipsisa64 | mipsisa64el \
	| mipsisa64r2 | mipsisa64r2el \
	| mipsisa64r6 | mipsisa64r6el \
	| mipsisa64sb1 | mipsisa64sb1el \
	| mipsisa64sr71k | mipsisa64sr71kel \
	| mipsr5900 | mipsr5900el \
	| mipstx39 | mipstx39el \
	| mn10200 | mn10300 \
	| moxie \
	| mt \
	| msp430 \
	| nds32 | nds32le | nds32be \
	| nios | nios2 | nios2eb | nios2el \
	| ns16k | ns32k \
	| open8 | or1k | or1knd | or32 \
	| pdp10 | pdp11 | pj | pjl \
	| powerpc | powerpc64 | powerpc64le | powerpcle \
	| pyramid \
	| riscv32 | riscv64 \
	| rl78 | rx \
	| score \
	| sh | sh[1234] | sh[24]a | sh[24]aeb | sh[23]e | sh[34]eb | sheb | shbe | shle | sh[1234]le | sh3ele \
	| sh64 | sh64le \
	| sparc | sparc64 | sparc64b | sparc64v | sparc86x | sparclet | sparclite \
	| sparcv8 | sparcv9 | sparcv9b | sparcv9v \
	| spu \
	| tahoe | tic4x | tic54x | tic55x | tic6x | tic80 | tron \
	| ubicom32 \
	| v850 | v850e | v850e1 | v850e2 | v850es | v850e2v3 \
	| visium \
	| we32k \
	| x86 | xc16x | xstormy16 | xtensa \
	| z8k | z80)
		basic_machine=$basic_machine-unknown
		;;
	c54x)
		basic_machine=tic54x-unknown
		;;
	c55x)
		basic_machine=tic55x-unknown
		;;
	c6x)
		basic_machine=tic6x-unknown
		;;
	leon|leon[3-9])
		basic_machine=sparc-$basic_machine
		;;
	m6811 | m68hc11 | m6812 | m68hc12 | m68hcs12x | nvptx | picochip)
		basic_machine=$basic_machine-unknown
		os=-none
		;;
	m88110 | m680[12346]0 | m683?2 | m68360 | m5200 | v70 | w65 | z8k)
		;;
	ms1)
		basic_machine=mt-unknown
		;;

	strongarm | thumb | xscale)
		basic_machine=arm-unknown
		;;
	xgate)
		basic_machine=$basic_machine-unknown
		os=-none
		;;
	xscaleeb)
		basic_machine=armeb-unknown
		;;

	xscaleel)
		basic_machine=armel-unknown
		;;

	# We use `pc' rather than `unknown'
	# because (1) that's what they normally are, and
	# (2) the word "unknown" tends to confuse beginning users.
	i*86 | x86_64)
	  basic_machine=$basic_machine-pc
	  ;;
	# Object if more than one company name word.
	*-*-*)
		echo Invalid configuration \`$1\': machine \`$basic_machine\' not recognized 1>&2
		exit 1
		;;
	# Recognize the basic CPU types with company name.
	580-* \
	| a29k-* \
	| aarch64-* | aarch64_be-* \
	| alpha-* | alphaev[4-8]-* | alphaev56-* | alphaev6[78]-* \
	| alpha64-* | alpha64ev[4-8]-* | alpha64ev56-* | alpha64ev6[78]-* \
	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* | arceb-* \
	| arm-*  | armbe-* | armle-* | armeb-* | armv*-* \
	| avr-* | avr32-* \
	| be32-* | be64-* \
	| bfin-* | bs2000-* \
	| c[123]* | c30-* | [cjt]90-* | c4x-* \
	| c8051-* | clipper-* | craynv-* | cydra-* \
	| d10v-* | d30v-* | dlx-* \
	| elxsi-* \
	| f30[01]-* | f700-* | fido-* | fr30-* | frv-* | fx80-* \
	| h8300-* | h8500-* \
	| hppa-* | hppa1.[01]-* | hppa2.0-* | hppa2.0[nw]-* | hppa64-* \
	| hexagon-* \
	| i*86-* | i860-* | i960-* | ia64-* \
	| ip2k-* | iq2000-* \
	| k1om-* \
	| le32-* | le64-* \
	| lm32-* \
	| m32c-* | m32r-* | m32rle-* \
	| m68000-* | m680[012346]0-* | m68360-* | m683?2-* | m68k-* \
	| m88110-* | m88k-* | maxq-* | mcore-* | metag-* \
	| microblaze-* | microblazeel-* \
	| mips-* | mipsbe-* | mipseb-* | mipsel-* | mipsle-* \
	| mips16-* \
	| mips64-* | mips64el-* \
	| mips64octeon-* | mips64octeonel-* \
	| mips64orion-* | mips64orionel-* \
	| mips64r5900-* | mips64r5900el-* \
	| mips64vr-* | mips64vrel-* \
	| mips64vr4100-* | mips64vr4100el-* \
	| mips64vr4300-* | mips64vr4300el-* \
	| mips64vr5000-* | mips64vr5000el-* \
	| mips64vr5900-* | mips64vr5900el-* \
	| mipsisa32-* | mipsisa32el-* \
	| mipsisa32r2-* | mipsisa32r2el-* \
	| mipsisa32r6-* | mipsisa32r6el-* \
	| mipsisa64-* | mipsisa64el-* \
	| mipsisa64r2-* | mipsisa64r2el-* \
	| mipsisa64r6-* | mipsisa64r6el-* \
	| mipsisa64sb1-* | mipsisa64sb1el-* \
	| mipsisa64sr71k-* | mipsisa64sr71kel-* \
	| mipsr5900-* | mipsr5900el-* \
	| mipstx39-* | mipstx39el-* \
	| mmix-* \
	| mt-* \
	| msp430-* \
	| nds32-* | nds32le-* | nds32be-* \
	| nios-* | nios2-* | nios2eb-* | nios2el-* \
	| none-* | np1-* | ns16k-* | ns32k-* \
	| open8-* \
	| or1k*-* \
	| orion-* \
	| pdp10-* | pdp11-* | pj-* | pjl-* | pn-* | power-* \
	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \
	| pyramid-* \
	| rl78-* | romp-* | rs6000-* | rx-* \
	| sh-* | sh[1234]-* | sh[24]a-* | sh[24]aeb-* | sh[23]e-* | sh[34]eb-* | sheb-* | shbe-* \
	| shle-* | sh[1234]le-* | sh3ele-* | sh64-* | sh64le-* \
	| sparc-* | sparc64-* | sparc64b-* | sparc64v-* | sparc86x-* | sparclet-* \
	| sparclite-* \
	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | sv1-* | sx?-* \
	| tahoe-* \
	| tic30-* | tic4x-* | tic54x-* | tic55x-* | tic6x-* | tic80-* \
	| tile*-* \
	| tron-* \
	| ubicom32-* \
	| v850-* | v850e-* | v850e1-* | v850es-* | v850e2-* | v850e2v3-* \
	| vax-* \
	| visium-* \
	| we32k-* \
	| x86-* | x86_64-* | xc16x-* | xps100-* \
	| xstormy16-* | xtensa*-* \
	| ymp-* \
	| z8k-* | z80-*)
		;;
	# Recognize the basic CPU types without company name, with glob match.
	xtensa*)
		basic_machine=$basic_machine-unknown
		;;
	# Recognize the various machine names and aliases which stand
	# for a CPU type and a company and sometimes even an OS.
	386bsd)
		basic_machine=i386-unknown
		os=-bsd
		;;
	3b1 | 7300 | 7300-att | att-7300 | pc7300 | safari | unixpc)
		basic_machine=m68000-att
		;;
	3b*)
		basic_machine=we32k-att
		;;
	a29khif)
		basic_machine=a29k-amd
		os=-udi
		;;
	abacus)
		basic_machine=abacus-unknown
		;;
	adobe68k)
		basic_machine=m68010-adobe
		os=-scout
		;;
	alliant | fx80)
		basic_machine=fx80-alliant
		;;
	altos | altos3068)
		basic_machine=m68k-altos
		;;
	am29k)
		basic_machine=a29k-none
		os=-bsd
		;;
	amd64)
		basic_machine=x86_64-pc
		;;
	amd64-*)
		basic_machine=x86_64-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	amdahl)
		basic_machine=580-amdahl
		os=-sysv
		;;
	amiga | amiga-*)
		basic_machine=m68k-unknown
		;;
	amigaos | amigados)
		basic_machine=m68k-unknown
		os=-amigaos
		;;
	amigaunix | amix)
		basic_machine=m68k-unknown
		os=-sysv4
		;;
	apollo68)
		basic_machine=m68k-apollo
		os=-sysv
		;;
	apollo68bsd)
		basic_machine=m68k-apollo
		os=-bsd
		;;
	aros)
		basic_machine=i386-pc
		os=-aros
		;;
	aux)
		basic_machine=m68k-apple
		os=-aux
		;;
	balance)
		basic_machine=ns32k-sequent
		os=-dynix
		;;
	blackfin)
		basic_machine=bfin-unknown
		os=-linux
		;;
	blackfin-*)
		basic_machine=bfin-`echo $basic_machine | sed 's/^[^-]*-//'`
		os=-linux
		;;
	bluegene*)
		basic_machine=powerpc-ibm
		os=-cnk
		;;
	c54x-*)
		basic_machine=tic54x-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	c55x-*)
		basic_machine=tic55x-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	c6x-*)
		basic_machine=tic6x-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	c90)
		basic_machine=c90-cray
		os=-unicos
		;;
	cegcc)
		basic_machine=arm-unknown
		os=-cegcc
		;;
	convex-c1)
		basic_machine=c1-convex
		os=-bsd
		;;
	convex-c2)
		basic_machine=c2-convex
		os=-bsd
		;;
	convex-c32)
		basic_machine=c32-convex
		os=-bsd
		;;
	convex-c34)
		basic_machine=c34-convex
		os=-bsd
		;;
	convex-c38)
		basic_machine=c38-convex
		os=-bsd
		;;
	cray | j90)
		basic_machine=j90-cray
		os=-unicos
		;;
	craynv)
		basic_machine=craynv-cray
		os=-unicosmp
		;;
	cr16 | cr16-*)
		basic_machine=cr16-unknown
		os=-elf
		;;
	crds | unos)
		basic_machine=m68k-crds
		;;
	crisv32 | crisv32-* | etraxfs*)
		basic_machine=crisv32-axis
		;;
	cris | cris-* | etrax*)
		basic_machine=cris-axis
		;;
	crx)
		basic_machine=crx-unknown
		os=-elf
		;;
	da30 | da30-*)
		basic_machine=m68k-da30
		;;
	decstation | decstation-3100 | pmax | pmax-* | pmin | dec3100 | decstatn)
		basic_machine=mips-dec
		;;
	decsystem10* | dec10*)
		basic_machine=pdp10-dec
		os=-tops10
		;;
	decsystem20* | dec20*)
		basic_machine=pdp10-dec
		os=-tops20
		;;
	delta | 3300 | motorola-3300 | motorola-delta \
	      | 3300-motorola | delta-motorola)
		basic_machine=m68k-motorola
		;;
	delta88)
		basic_machine=m88k-motorola
		os=-sysv3
		;;
	dicos)
		basic_machine=i686-pc
		os=-dicos
		;;
	djgpp)
		basic_machine=i586-pc
		os=-msdosdjgpp
		;;
	dpx20 | dpx20-*)
		basic_machine=rs6000-bull
		os=-bosx
		;;
	dpx2* | dpx2*-bull)
		basic_machine=m68k-bull
		os=-sysv3
		;;
	ebmon29k)
		basic_machine=a29k-amd
		os=-ebmon
		;;
	elxsi)
		basic_machine=elxsi-elxsi
		os=-bsd
		;;
	encore | umax | mmax)
		basic_machine=ns32k-encore
		;;
	es1800 | OSE68k | ose68k | ose | OSE)
		basic_machine=m68k-ericsson
		os=-ose
		;;
	fx2800)
		basic_machine=i860-alliant
		;;
	genix)
		basic_machine=ns32k-ns
		;;
	gmicro)
		basic_machine=tron-gmicro
		os=-sysv
		;;
	go32)
		basic_machine=i386-pc
		os=-go32
		;;
	h3050r* | hiux*)
		basic_machine=hppa1.1-hitachi
		os=-hiuxwe2
		;;
	h8300hms)
		basic_machine=h8300-hitachi
		os=-hms
		;;
	h8300xray)
		basic_machine=h8300-hitachi
		os=-xray
		;;
	h8500hms)
		basic_machine=h8500-hitachi
		os=-hms
		;;
	harris)
		basic_machine=m88k-harris
		os=-sysv3
		;;
	hp300-*)
		basic_machine=m68k-hp
		;;
	hp300bsd)
		basic_machine=m68k-hp
		os=-bsd
		;;
	hp300hpux)
		basic_machine=m68k-hp
		os=-hpux
		;;
	hp3k9[0-9][0-9] | hp9[0-9][0-9])
		basic_machine=hppa1.0-hp
		;;
	hp9k2[0-9][0-9] | hp9k31[0-9])
		basic_machine=m68000-hp
		;;
	hp9k3[2-9][0-9])
		basic_machine=m68k-hp
		;;
	hp9k6[0-9][0-9] | hp6[0-9][0-9])
		basic_machine=hppa1.0-hp
		;;
	hp9k7[0-79][0-9] | hp7[0-79][0-9])
		basic_machine=hppa1.1-hp
		;;
	hp9k78[0-9] | hp78[0-9])
		# FIXME: really hppa2.0-hp
		basic_machine=hppa1.1-hp
		;;
	hp9k8[67]1 | hp8[67]1 | hp9k80[24] | hp80[24] | hp9k8[78]9 | hp8[78]9 | hp9k893 | hp893)
		# FIXME: really hppa2.0-hp
		basic_machine=hppa1.1-hp
		;;
	hp9k8[0-9][13679] | hp8[0-9][13679])
		basic_machine=hppa1.1-hp
		;;
	hp9k8[0-9][0-9] | hp8[0-9][0-9])
		basic_machine=hppa1.0-hp
		;;
	hppa-next)
		os=-nextstep3
		;;
	hppaosf)
		basic_machine=hppa1.1-hp
		os=-osf
		;;
	hppro)
		basic_machine=hppa1.1-hp
		os=-proelf
		;;
	i370-ibm* | ibm*)
		basic_machine=i370-ibm
		;;
	i*86v32)
		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
		os=-sysv32
		;;
	i*86v4*)
		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
		os=-sysv4
		;;
	i*86v)
		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
		os=-sysv
		;;
	i*86sol2)
		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
		os=-solaris2
		;;
	i386mach)
		basic_machine=i386-mach
		os=-mach
		;;
	i386-vsta | vsta)
		basic_machine=i386-unknown
		os=-vsta
		;;
	iris | iris4d)
		basic_machine=mips-sgi
		case $os in
		    -irix*)
			;;
		    *)
			os=-irix4
			;;
		esac
		;;
	isi68 | isi)
		basic_machine=m68k-isi
		os=-sysv
		;;
	leon-*|leon[3-9]-*)
		basic_machine=sparc-`echo $basic_machine | sed 's/-.*//'`
		;;
	m68knommu)
		basic_machine=m68k-unknown
		os=-linux
		;;
	m68knommu-*)
		basic_machine=m68k-`echo $basic_machine | sed 's/^[^-]*-//'`
		os=-linux
		;;
	m88k-omron*)
		basic_machine=m88k-omron
		;;
	magnum | m3230)
		basic_machine=mips-mips
		os=-sysv
		;;
	merlin)
		basic_machine=ns32k-utek
		os=-sysv
		;;
	microblaze*)
		basic_machine=microblaze-xilinx
		;;
	mingw64)
		basic_machine=x86_64-pc
		os=-mingw64
		;;
	mingw32)
		basic_machine=i686-pc
		os=-mingw32
		;;
	mingw32ce)
		basic_machine=arm-unknown
		os=-mingw32ce
		;;
	miniframe)
		basic_machine=m68000-convergent
		;;
	*mint | -mint[0-9]* | *MiNT | *MiNT[0-9]*)
		basic_machine=m68k-atari
		os=-mint
		;;
	mips3*-*)
		basic_machine=`echo $basic_machine | sed -e 's/mips3/mips64/'`
		;;
	mips3*)
		basic_machine=`echo $basic_machine | sed -e 's/mips3/mips64/'`-unknown
		;;
	monitor)
		basic_machine=m68k-rom68k
		os=-coff
		;;
	morphos)
		basic_machine=powerpc-unknown
		os=-morphos
		;;
	moxiebox)
		basic_machine=moxie-unknown
		os=-moxiebox
		;;
	msdos)
		basic_machine=i386-pc
		os=-msdos
		;;
	ms1-*)
		basic_machine=`echo $basic_machine | sed -e 's/ms1-/mt-/'`
		;;
	msys)
		basic_machine=i686-pc
		os=-msys
		;;
	mvs)
		basic_machine=i370-ibm
		os=-mvs
		;;
	nacl)
		basic_machine=le32-unknown
		os=-nacl
		;;
	ncr3000)
		basic_machine=i486-ncr
		os=-sysv4
		;;
	netbsd386)
		basic_machine=i386-unknown
		os=-netbsd
		;;
	netwinder)
		basic_machine=armv4l-rebel
		os=-linux
		;;
	news | news700 | news800 | news900)
		basic_machine=m68k-sony
		os=-newsos
		;;
	news1000)
		basic_machine=m68030-sony
		os=-newsos
		;;
	news-3600 | risc-news)
		basic_machine=mips-sony
		os=-newsos
		;;
	necv70)
		basic_machine=v70-nec
		os=-sysv
		;;
	next | m*-next )
		basic_machine=m68k-next
		case $os in
		    -nextstep* )
			;;
		    -ns2*)
		      os=-nextstep2
			;;
		    *)
		      os=-nextstep3
			;;
		esac
		;;
	nh3000)
		basic_machine=m68k-harris
		os=-cxux
		;;
	nh[45]000)
		basic_machine=m88k-harris
		os=-cxux
		;;
	nindy960)
		basic_machine=i960-intel
		os=-nindy
		;;
	mon960)
		basic_machine=i960-intel
		os=-mon960
		;;
	nonstopux)
		basic_machine=mips-compaq
		os=-nonstopux
		;;
	np1)
		basic_machine=np1-gould
		;;
	neo-tandem)
		basic_machine=neo-tandem
		;;
	nse-tandem)
		basic_machine=nse-tandem
		;;
	nsr-tandem)
		basic_machine=nsr-tandem
		;;
	op50n-* | op60c-*)
		basic_machine=hppa1.1-oki
		os=-proelf
		;;
	openrisc | openrisc-*)
		basic_machine=or32-unknown
		;;
	os400)
		basic_machine=powerpc-ibm
		os=-os400
		;;
	OSE68000 | ose68000)
		basic_machine=m68000-ericsson
		os=-ose
		;;
	os68k)
		basic_machine=m68k-none
		os=-os68k
		;;
	pa-hitachi)
		basic_machine=hppa1.1-hitachi
		os=-hiuxwe2
		;;
	paragon)
		basic_machine=i860-intel
		os=-osf
		;;
	parisc)
		basic_machine=hppa-unknown
		os=-linux
		;;
	parisc-*)
		basic_machine=hppa-`echo $basic_machine | sed 's/^[^-]*-//'`
		os=-linux
		;;
	pbd)
		basic_machine=sparc-tti
		;;
	pbb)
		basic_machine=m68k-tti
		;;
	pc532 | pc532-*)
		basic_machine=ns32k-pc532
		;;
	pc98)
		basic_machine=i386-pc
		;;
	pc98-*)
		basic_machine=i386-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	pentium | p5 | k5 | k6 | nexgen | viac3)
		basic_machine=i586-pc
		;;
	pentiumpro | p6 | 6x86 | athlon | athlon_*)
		basic_machine=i686-pc
		;;
	pentiumii | pentium2 | pentiumiii | pentium3)
		basic_machine=i686-pc
		;;
	pentium4)
		basic_machine=i786-pc
		;;
	pentium-* | p5-* | k5-* | k6-* | nexgen-* | viac3-*)
		basic_machine=i586-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	pentiumpro-* | p6-* | 6x86-* | athlon-*)
		basic_machine=i686-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	pentiumii-* | pentium2-* | pentiumiii-* | pentium3-*)
		basic_machine=i686-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	pentium4-*)
		basic_machine=i786-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	pn)
		basic_machine=pn-gould
		;;
	power)	basic_machine=power-ibm
		;;
	ppc | ppcbe)	basic_machine=powerpc-unknown
		;;
	ppc-* | ppcbe-*)
		basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	ppcle | powerpclittle | ppc-le | powerpc-little)
		basic_machine=powerpcle-unknown
		;;
	ppcle-* | powerpclittle-*)
		basic_machine=powerpcle-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	ppc64)	basic_machine=powerpc64-unknown
		;;
	ppc64-*) basic_machine=powerpc64-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	ppc64le | powerpc64little | ppc64-le | powerpc64-little)
		basic_machine=powerpc64le-unknown
		;;
	ppc64le-* | powerpc64little-*)
		basic_machine=powerpc64le-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	ps2)
		basic_machine=i386-ibm
		;;
	pw32)
		basic_machine=i586-unknown
		os=-pw32
		;;
	rdos | rdos64)
		basic_machine=x86_64-pc
		os=-rdos
		;;
	rdos32)
		basic_machine=i386-pc
		os=-rdos
		;;
	rom68k)
		basic_machine=m68k-rom68k
		os=-coff
		;;
	rm[46]00)
		basic_machine=mips-siemens
		;;
	rtpc | rtpc-*)
		basic_machine=romp-ibm
		;;
	s390 | s390-*)
		basic_machine=s390-ibm
		;;
	s390x | s390x-*)
		basic_machine=s390x-ibm
		;;
	sa29200)
		basic_machine=a29k-amd
		os=-udi
		;;
	sb1)
		basic_machine=mipsisa64sb1-unknown
		;;
	sb1el)
		basic_machine=mipsisa64sb1el-unknown
		;;
	sde)
		basic_machine=mipsisa32-sde
		os=-elf
		;;
	sei)
		basic_machine=mips-sei
		os=-seiux
		;;
	sequent)
		basic_machine=i386-sequent
		;;
	sh)
		basic_machine=sh-hitachi
		os=-hms
		;;
	sh5el)
		basic_machine=sh5le-unknown
		;;
	sh64)
		basic_machine=sh64-unknown
		;;
	sparclite-wrs | simso-wrs)
		basic_machine=sparclite-wrs
		os=-vxworks
		;;
	sps7)
		basic_machine=m68k-bull
		os=-sysv2
		;;
	spur)
		basic_machine=spur-unknown
		;;
	st2000)
		basic_machine=m68k-tandem
		;;
	stratus)
		basic_machine=i860-stratus
		os=-sysv4
		;;
	strongarm-* | thumb-*)
		basic_machine=arm-`echo $basic_machine | sed 's/^[^-]*-//'`
		;;
	sun2)
		basic_machine=m68000-sun
		;;
	sun2os3)
		basic_machine=m68000-sun
		os=-sunos3
		;;
	sun2os4)
		basic_machine=m68000-sun
		os=-sunos4
		;;
	sun3os3)
		basic_machine=m68k-sun
		os=-sunos3
		;;
	sun3os4)
		basic_machine=m68k-sun
		os=-sunos4
		;;
	sun4os3)
		basic_machine=sparc-sun
		os=-sunos3
		;;
	sun4os4)
		basic_machine=sparc-sun
		os=-sunos4
		;;
	sun4sol2)
		basic_machine=sparc-sun
		os=-solaris2
		;;
	sun3 | sun3-*)
		basic_machine=m68k-sun
		;;
	sun4)
		basic_machine=sparc-sun
		;;
	sun386 | sun386i | roadrunner)
		basic_machine=i386-sun
		;;
	sv1)
		basic_machine=sv1-cray
		os=-unicos
		;;
	symmetry)
		basic_machine=i386-sequent
		os=-dynix
		;;
	t3e)
		basic_machine=alphaev5-cray
		os=-unicos
		;;
	t90)
		basic_machine=t90-cray
		os=-unicos
		;;
	tile*)
		basic_machine=$basic_machine-unknown
		os=-linux-gnu
		;;
	tx39)
		basic_machine=mipstx39-unknown
		;;
	tx39el)
		basic_machine=mipstx39el-unknown
		;;
	toad1)
		basic_machine=pdp10-xkl
		os=-tops20
		;;
	tower | tower-32)
		basic_machine=m68k-ncr
		;;
	tpf)
		basic_machine=s390x-ibm
		os=-tpf
		;;
	udi29k)
		basic_machine=a29k-amd
		os=-udi
		;;
	ultra3)
		basic_machine=a29k-nyu
		os=-sym1
		;;
	v810 | necv810)
		basic_machine=v810-nec
		os=-none
		;;
	vaxv)
		basic_machine=vax-dec
		os=-sysv
		;;
	vms)
		basic_machine=vax-dec
		os=-vms
		;;
	vpp*|vx|vx-*)
		basic_machine=f301-fujitsu
		;;
	vxworks960)
		basic_machine=i960-wrs
		os=-vxworks
		;;
	vxworks68)
		basic_machine=m68k-wrs
		os=-vxworks
		;;
	vxworks29k)
		basic_machine=a29k-wrs
		os=-vxworks
		;;
	w65*)
		basic_machine=w65-wdc
		os=-none
		;;
	w89k-*)
		basic_machine=hppa1.1-winbond
		os=-proelf
		;;
	xbox)
		basic_machine=i686-pc
		os=-mingw32
		;;
	xps | xps100)
		basic_machine=xps100-honeywell
		;;
	xscale-* | xscalee[bl]-*)
		basic_machine=`echo $basic_machine | sed 's/^xscale/arm/'`
		;;
	ymp)
		basic_machine=ymp-cray
		os=-unicos
		;;
	z8k-*-coff)
		basic_machine=z8k-unknown
		os=-sim
		;;
	z80-*-coff)
		basic_machine=z80-unknown
		os=-sim
		;;
	none)
		basic_machine=none-none
		os=-none
		;;

# Here we handle the default manufacturer of certain CPU types.  It is in
# some cases the only manufacturer, in others, it is the most popular.
	w89k)
		basic_machine=hppa1.1-winbond
		;;
	op50n)
		basic_machine=hppa1.1-oki
		;;
	op60c)
		basic_machine=hppa1.1-oki
		;;
	romp)
		basic_machine=romp-ibm
		;;
	mmix)
		basic_machine=mmix-knuth
		;;
	rs6000)
		basic_machine=rs6000-ibm
		;;
	vax)
		basic_machine=vax-dec
		;;
	pdp10)
		# there are many clones, so DEC is not a safe bet
		basic_machine=pdp10-unknown
		;;
	pdp11)
		basic_machine=pdp11-dec
		;;
	we32k)
		basic_machine=we32k-att
		;;
	sh[1234] | sh[24]a | sh[24]aeb | sh[34]eb | sh[1234]le | sh[23]ele)
		basic_machine=sh-unknown
		;;
	sparc | sparcv8 | sparcv9 | sparcv9b | sparcv9v)
		basic_machine=sparc-sun
		;;
	cydra)
		basic_machine=cydra-cydrome
		;;
	orion)
		basic_machine=orion-highlevel
		;;
	orion105)
		basic_machine=clipper-highlevel
		;;
	mac | mpw | mac-mpw)
		basic_machine=m68k-apple
		;;
	pmac | pmac-mpw)
		basic_machine=powerpc-apple
		;;
	*-unknown)
		# Make sure to match an already-canonicalized machine name.
		;;
	*)
		echo Invalid configuration \`$1\': machine \`$basic_machine\' not recognized 1>&2
		exit 1
		;;
esac

# Here we canonicalize certain aliases for manufacturers.
case $basic_machine in
	*-digital*)
		basic_machine=`echo $basic_machine | sed 's/digital.*/dec/'`
		;;
	*-commodore*)
		basic_machine=`echo $basic_machine | sed 's/commodore.*/cbm/'`
		;;
	*)
		;;
esac

# Decode manufacturer-specific aliases for certain operating systems.

if [ x"$os" != x"" ]
then
case $os in
	# First match some system type aliases
	# that might get confused with valid system types.
	# -solaris* is a basic system type, with this one exception.
	-auroraux)
		os=-auroraux
		;;
	-solaris1 | -solaris1.*)
		os=`echo $os | sed -e 's|solaris1|sunos4|'`
		;;
	-solaris)
		os=-solaris2
		;;
	-svr4*)
		os=-sysv4
		;;
	-unixware*)
		os=-sysv4.2uw
		;;
	-gnu/linux*)
		os=`echo $os | sed -e 's|gnu/linux|linux-gnu|'`
		;;
	# First accept the basic system types.
	# The portable systems comes first.
	# Each alternative MUST END IN A *, to match a version number.
	# -sysv* is not here because it comes later, after sysvr4.
	-gnu* | -bsd* | -mach* | -minix* | -genix* | -ultrix* | -irix* \
	      | -*vms* | -sco* | -esix* | -isc* | -aix* | -cnk* | -sunos | -sunos[34]*\
	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -auroraux* | -solaris* \
	      | -sym* | -kopensolaris* | -plan9* \
	      | -amigaos* | -amigados* | -msdos* | -newsos* | -unicos* | -aof* \
	      | -aos* | -aros* \
	      | -nindy* | -vxsim* | -vxworks* | -ebmon* | -hms* | -mvs* \
	      | -clix* | -riscos* | -uniplus* | -iris* | -rtu* | -xenix* \
	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* \
	      | -bitrig* | -openbsd* | -solidbsd* \
	      | -ekkobsd* | -kfreebsd* | -freebsd* | -riscix* | -lynxos* \
	      | -bosx* | -nextstep* | -cxux* | -aout* | -elf* | -oabi* \
	      | -ptx* | -coff* | -ecoff* | -winnt* | -domain* | -vsta* \
	      | -udi* | -eabi* | -lites* | -ieee* | -go32* | -aux* \
	      | -chorusos* | -chorusrdb* | -cegcc* \
	      | -cygwin* | -msys* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
	      | -mingw32* | -mingw64* | -linux-gnu* | -linux-android* \
	      | -linux-newlib* | -linux-musl* | -linux-uclibc* \
	      | -uxpv* | -beos* | -mpeix* | -udk* | -moxiebox* \
	      | -interix* | -uwin* | -mks* | -rhapsody* | -darwin* | -opened* \
	      | -openstep* | -oskit* | -conix* | -pw32* | -nonstopux* \
	      | -storm-chaos* | -tops10* | -tenex* | -tops20* | -its* \
	      | -os2* | -vos* | -palmos* | -uclinux* | -nucleus* \
	      | -morphos* | -superux* | -rtmk* | -rtmk-nova* | -windiss* \
	      | -powermax* | -dnix* | -nx6 | -nx7 | -sei* | -dragonfly* \
	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es* | -tirtos*)
	# Remember, each alternative MUST END IN *, to match a version number.
		;;
	-qnx*)
		case $basic_machine in
		    x86-* | i*86-*)
			;;
		    *)
			os=-nto$os
			;;
		esac
		;;
	-nto-qnx*)
		;;
	-nto*)
		os=`echo $os | sed -e 's|nto|nto-qnx|'`
		;;
	-sim | -es1800* | -hms* | -xray | -os68k* | -none* | -v88r* \
	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* | -haiku* \
	      | -macos* | -mpw* | -magic* | -mmixware* | -mon960* | -lnews*)
		;;
	-mac*)
		os=`echo $os | sed -e 's|mac|macos|'`
		;;
	-linux-dietlibc)
		os=-linux-dietlibc
		;;
	-linux*)
		os=`echo $os | sed -e 's|linux|linux-gnu|'`
		;;
	-sunos5*)
		os=`echo $os | sed -e 's|sunos5|solaris2|'`
		;;
	-sunos6*)
		os=`echo $os | sed -e 's|sunos6|solaris3|'`
		;;
	-opened*)
		os=-openedition
		;;
	-os400*)
		os=-os400
		;;
	-wince*)
		os=-wince
		;;
	-osfrose*)
		os=-osfrose
		;;
	-osf*)
		os=-osf
		;;
	-utek*)
		os=-bsd
		;;
	-dynix*)
		os=-bsd
		;;
	-acis*)
		os=-aos
		;;
	-atheos*)
		os=-atheos
		;;
	-syllable*)
		os=-syllable
		;;
	-386bsd)
		os=-bsd
		;;
	-ctix* | -uts*)
		os=-sysv
		;;
	-nova*)
		os=-rtmk-nova
		;;
	-ns2 )
		os=-nextstep2
		;;
	-nsk*)
		os=-nsk
		;;
	# Preserve the version number of sinix5.
	-sinix5.*)
		os=`echo $os | sed -e 's|sinix|sysv|'`
		;;
	-sinix*)
		os=-sysv4
		;;
	-tpf*)
		os=-tpf
		;;
	-triton*)
		os=-sysv3
		;;
	-oss*)
		os=-sysv3
		;;
	-svr4)
		os=-sysv4
		;;
	-svr3)
		os=-sysv3
		;;
	-sysvr4)
		os=-sysv4
		;;
	# This must come after -sysvr4.
	-sysv*)
		;;
	-ose*)
		os=-ose
		;;
	-es1800*)
		os=-ose
		;;
	-xenix)
		os=-xenix
		;;
	-*mint | -mint[0-9]* | -*MiNT | -MiNT[0-9]*)
		os=-mint
		;;
	-aros*)
		os=-aros
		;;
	-zvmoe)
		os=-zvmoe
		;;
	-dicos*)
		os=-dicos
		;;
	-nacl*)
		;;
	-none)
		;;
	*)
		# Get rid of the `-' at the beginning of $os.
		os=`echo $os | sed 's/[^-]*-//'`
		echo Invalid configuration \`$1\': system \`$os\' not recognized 1>&2
		exit 1
		;;
esac
else

# Here we handle the default operating systems that come with various machines.
# The value should be what the vendor currently ships out the door with their
# machine or put another way, the most popular os provided with the machine.

# Note that if you're going to try to match "-MANUFACTURER" here (say,
# "-sun"), then you have to tell the case statement up towards the top
# that MANUFACTURER isn't an operating system.  Otherwise, code above
# will signal an error saying that MANUFACTURER isn't an operating
# system, and we'll never get to this point.

case $basic_machine in
	score-*)
		os=-elf
		;;
	spu-*)
		os=-elf
		;;
	*-acorn)
		os=-riscix1.2
		;;
	arm*-rebel)
		os=-linux
		;;
	arm*-semi)
		os=-aout
		;;
	c4x-* | tic4x-*)
		os=-coff
		;;
	c8051-*)
		os=-elf
		;;
	hexagon-*)
		os=-elf
		;;
	tic54x-*)
		os=-coff
		;;
	tic55x-*)
		os=-coff
		;;
	tic6x-*)
		os=-coff
		;;
	# This must come before the *-dec entry.
	pdp10-*)
		os=-tops20
		;;
	pdp11-*)
		os=-none
		;;
	*-dec | vax-*)
		os=-ultrix4.2
		;;
	m68*-apollo)
		os=-domain
		;;
	i386-sun)
		os=-sunos4.0.2
		;;
	m68000-sun)
		os=-sunos3
		;;
	m68*-cisco)
		os=-aout
		;;
	mep-*)
		os=-elf
		;;
	mips*-cisco)
		os=-elf
		;;
	mips*-*)
		os=-elf
		;;
	or32-*)
		os=-coff
		;;
	*-tti)	# must be before sparc entry or we get the wrong os.
		os=-sysv3
		;;
	sparc-* | *-sun)
		os=-sunos4.1.1
		;;
	*-be)
		os=-beos
		;;
	*-haiku)
		os=-haiku
		;;
	*-ibm)
		os=-aix
		;;
	*-knuth)
		os=-mmixware
		;;
	*-wec)
		os=-proelf
		;;
	*-winbond)
		os=-proelf
		;;
	*-oki)
		os=-proelf
		;;
	*-hp)
		os=-hpux
		;;
	*-hitachi)
		os=-hiux
		;;
	i860-* | *-att | *-ncr | *-altos | *-motorola | *-convergent)
		os=-sysv
		;;
	*-cbm)
		os=-amigaos
		;;
	*-dg)
		os=-dgux
		;;
	*-dolphin)
		os=-sysv3
		;;
	m68k-ccur)
		os=-rtu
		;;
	m88k-omron*)
		os=-luna
		;;
	*-next )
		os=-nextstep
		;;
	*-sequent)
		os=-ptx
		;;
	*-crds)
		os=-unos
		;;
	*-ns)
		os=-genix
		;;
	i370-*)
		os=-mvs
		;;
	*-next)
		os=-nextstep3
		;;
	*-gould)
		os=-sysv
		;;
	*-highlevel)
		os=-bsd
		;;
	*-encore)
		os=-bsd
		;;
	*-sgi)
		os=-irix
		;;
	*-siemens)
		os=-sysv4
		;;
	*-masscomp)
		os=-rtu
		;;
	f30[01]-fujitsu | f700-fujitsu)
		os=-uxpv
		;;
	*-rom68k)
		os=-coff
		;;
	*-*bug)
		os=-coff
		;;
	*-apple)
		os=-macos
		;;
	*-atari*)
		os=-mint
		;;
	*)
		os=-none
		;;
esac
fi

# Here we handle the case where we know the os, and the CPU type, but not the
# manufacturer.  We pick the logical manufacturer.
vendor=unknown
case $basic_machine in
	*-unknown)
		case $os in
			-riscix*)
				vendor=acorn
				;;
			-sunos*)
				vendor=sun
				;;
			-cnk*|-aix*)
				vendor=ibm
				;;
			-beos*)
				vendor=be
				;;
			-hpux*)
				vendor=hp
				;;
			-mpeix*)
				vendor=hp
				;;
			-hiux*)
				vendor=hitachi
				;;
			-unos*)
				vendor=crds
				;;
			-dgux*)
				vendor=dg
				;;
			-luna*)
				vendor=omron
				;;
			-genix*)
				vendor=ns
				;;
			-mvs* | -opened*)
				vendor=ibm
				;;
			-os400*)
				vendor=ibm
				;;
			-ptx*)
				vendor=sequent
				;;
			-tpf*)
				vendor=ibm
				;;
			-vxsim* | -vxworks* | -windiss*)
				vendor=wrs
				;;
			-aux*)
				vendor=apple
				;;
			-hms*)
				vendor=hitachi
				;;
			-mpw* | -macos*)
				vendor=apple
				;;
			-*mint | -mint[0-9]* | -*MiNT | -MiNT[0-9]*)
				vendor=atari
				;;
			-vos*)
				vendor=stratus
				;;
		esac
		basic_machine=`echo $basic_machine | sed "s/unknown/$vendor/"`
		;;
esac

echo $basic_machine$os
exit

# Local variables:
# eval: (add-hook 'write-file-hooks 'time-stamp)
# time-stamp-start: "timestamp='"
# time-stamp-format: "%:y-%02m-%02d"
# time-stamp-end: "'"
# End: