gcl fails fricas tests

[Qian Yun]

Hi Camm,

Thanks for your recent GCL fixes for FriCAS.

Before your recent fixes, GCL fails to build fricas
on arm64 ubuntu-24.04,
(
https://github.com/oldk1331/fricas/actions/runs/13005737431/job/36272189339
when compiling OCAMON.spad:
>> System error:
Condition in SETQ [or a callee]: INTERNAL-SIMPLE-ERROR: Value stack
overflow.
)

But after your recent fixes, now the build passes on arm64,
but it fails at these tests:
(
https://github.com/oldk1331/fricas/actions/runs/13136618548/job/36653370009
https://github.com/fricas/fricas/actions/runs/13136970950/job/36654558469

src/input/files.input:
Aborted (core dumped)

check failed
2 failing files:

eltuniseg.output: 15
mantepse.output: 4
)

These tests also fail on x86_64, where they did not fail before.

- Best,
- Qian

[Camm Maguire]

Greetings, and thanks so much for your feedback!

Thank you so much for pointing out that these test failures are new, I
did not know that. I had been focusing on the 2.7 build, but will now
look at these so we have a clean build with both GCL versions.

Take care,

--
Camm Maguire ca...@maguirefamily.org
==========================================================================
"The earth is but one country, and mankind its citizens." -- Baha'u'llah

[Camm Maguire]

Greetings, and thanks again!

> eltuniseg.output: 15
> mantepse.output: 4
> )
>

I am having trouble deciphering these output files. They appear
identical to files which pass except for the error summary at the end.
Can you help me isolate a sample failing command?

Take care,

[Qian Yun]

I took a look and can give you the following minimal example:

(7) -> [x for x in 1..2]

(7) [1, 2]
Type:
List(PositiveInteger)
(8) -> [x for x in 2..1]

>> System error:
SIMPLE-ERROR: Cannot coerce #<vector 0000000000e80400> to LIST.

- Qian

[Camm Maguire]

Greetings, and thanks so much! Small fricas commit to fix this. Small
GCL 2.6.15pre commit to fix Waldek's circle print/read issue in
files.input. Current fricas master now builds against both current GCL
master and GCL Version_2_6_15pre without failures, AFAICS. Will update
Debian GCL packages in a few days.

Waldek, there is still the file flush/close issue you mentioned before.
Is there an example I can investigate?

Take care,

Qian Yun <oldk...@gmail.com> writes:

[Qian Yun]

Thanks, I can confirm that gcl can build fricas and pass
all tests on arm64 and x86_84 linux now.

- Qian

[Waldek Hebisch]

On Tue, Feb 04, 2025 at 10:34:46AM -0500, Camm Maguire wrote:
> Greetings, and thanks so much for your feedback!
>
> Thank you so much for pointing out that these test failures are new, I
> did not know that. I had been focusing on the 2.7 build, but will now
> look at these so we have a clean build with both GCL versions.

Is any of GCL versions supposed to run on X32.? I tried Version_2_6_15pre
in chroot with X32 gentoo (outside is kernel from Debian 12). Configure
reports 'x86_64-pc-linux-gnux32', build goes some time and later dies:

Compiling /home/hebisch/kompi/gcl-2.6.15pre/unixport/../pcl/gcl_pcl_defclass.lisp.

Error: ERROR "Caught fatal error [memory may be damaged]"
Fast links are on: do (si::use-fast-links nil) for debugging
Signalled by COMPILE-FILE.
ERROR "Caught fatal error [memory may be damaged]"

Broken at COMPILE-FILE. Type :H for Help.
1 Return to top level.
PCL>>make[1]: *** [makefile:35: gcl_pcl_boot.c] Error 255

--
Waldek Hebisch

[Camm Maguire]

Greetings, and thanks for your report!

Appears to be gentoo specific:

https://buildd.debian.org/status/package.php?p=gcl
https://buildd.debian.org/status/fetch.php?pkg=gcl&arch=x32&ver=2.6.14-12&stamp=1738620948&raw=0

If there is ssh access I'd be happy to take a look, otherwise I suggest
running the offending command in gdb and posting a backtrace.

Take care,