[M3devel] minor problems on the release branch

[Jay K]

I believe the .M3SHIP tests have always failed because they look for Posix paths in the output.
You can likely now create correct stdout.pgm-NT386 and they will work.
It's not just the slashes, but also libfoo.a vs. foo.lib, foo.exe vs. foo.
I think.
I vaguely recall noticing that and that the tests not the functionality was broken.

To the extent that the tests are looking for "back substitution" to work and
don't care about the others, change the slashes, remove the last path element,
and declare that correct output?


The float<>string conversion worries me some.
I wish we had noticed it earlier.
At a certain level, it's not a big deal: the resulting strings are certainly "reasonable".
But programs could legitimately take a pretty close dependency on the behavior here. Right?
  Until/unless VAX support comes back, in which case we'll probably get varying results.


To debug this, a bit tedious, but someone might consider getting the release branch
to just before the failure and verifying it works, then just after the failure, verify it doens't work.
Then compare the two branches.

SOLsun fails for a minor reason:
+cc: Warning: -xarch=v8plus is deprecated, use -m32 -xarch=sparc instead
I probably just have to redirect it to >/dev/null.
I don't just change the config file. The config file works with a range of cc versions and probes their behavior.
  (something we might consider in a more autoconf-sh way, i.e. less often, though it isn't all that often, and
  someone can upgrade their cc at almost any time, so...)


Lots to work through..


   - Jay


----------------------------------------
> Date: Thu, 15 Jul 2010 14:43:43 +0200
> From: wagner at elegosoft.com
> To: m3devel at elegosoft.com
> Subject: [M3devel] minor problems on the release branch
>
> While checking the status of the final 5.8.6 release of CM3, I found
> several minor problems, which would benefit from attention nonetheless.
>
> Investigating these should also be a good starting point for newbies and
> others lurking on the list and waiting for something to do:
>
> Why did this change?
>
>
> http://hudson.modula3.com:8080/view/cm3/job/cm3-test-m3tests-AMD64_LINUX/lastBuild/testReport/(root)/m3tests/p040__binary_____ASCII_conversion_routines
> (failing since Build #556 (Apr 29, 2010 9:35:06 PM))
>
> same issue:
>
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-m3tests-I386_OPENBSD/lastBuild/testReport/(root)/m3tests/p040__binary_____ASCII_conversion_routines
> (failing since Build #28 (Nov 22, 2009 7:43:34 AM))
>
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-m3tests-LINUXLIBC6/lastBuild/testReport/(root)/m3tests/p040__binary_____ASCII_conversion_routines
> (failing since Build #276 (Jul 9, 2010 11:40:25 AM))
>
> The shipping tests on Nt386 seem to have never worked:
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-m3tests-NT386/374/testReport/
>
> Test reporting ist still more or less completely broken on SOLsun :-(
>
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-m3tests-SOLsun/lastBuild
>
> Failed packages on I386_OPENBSD:
>
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-all-pkgs-I386_OPENBSD/44/testReport/
>
> Failed packages on NT386:
>
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-all-pkgs-NT386/182/testReport/
>
> m3gdb build fails on PPC_LINUX and SPARC32_LINUX:
>
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-all-pkgs-PPC_LINUX/18/testReport/
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-all-pkgs-SPARC32_LINUX/lastBuild/testReport/
>
> Failed package tests on SOLgnu and SOLsun:
>
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-all-pkgs-SOLgnu/lastBuild/testReport/
>
> http://hudson.modula3.com:8080/view/cm3-test/job/cm3-test-all-pkgs-SOLsun/lastBuild/testReport/
>
> Regards,
>
> Olaf
> --
> Olaf Wagner -- elego Software Solutions GmbH
> Gustav-Meyer-Allee 25 / Gebäude 12, 13355 Berlin, Germany
> phone: +49 30 23 45 86 96 mobile: +49 177 2345 869 fax: +49 30 23 45 86 95
> http://www.elegosoft.com | Geschäftsführer: Olaf Wagner | Sitz: Berlin
> Handelregister: Amtsgericht Charlottenburg HRB 77719 | USt-IdNr: DE163214194
>