Re: 'make check' from system() under Python

[Eric Blake]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

According to Albert Chin on 2/4/2007 7:55 AM:
> 'make check' works fine when run from a shell. It works fine when run
> from system() under Perl. However, when run under Python (2.2.3,
> 2.3.4, 2.4.2), "os.system('make check')" fails with:
>   Checking ./134.sysval
>   @ ../doc/m4.texinfo:4812: Origin of test
>   ./134.sysval: stdout mismatch
>   2c2
>   < 3328
>   ---
>   > 0
>   ...
>   Failed checks were:
>     ./134.sysval:out
> 
> Any ideas as to why?

Probably something to do with inherited signal handling and python's
choice of what signals to ignore.  I don't know whether the bug is in
python for ignoring a signal that the child process expects to normally be
fatal, or in m4 for not trying to detect the situation of a parent process
starting it with an unusual signal mask.  But since only python tickles
the bug, that's where I would send the bug report.

> 
> The latest Fedora m4.spec has:
>   * Sun Nov 26 2006 Miloslav Trmac <address@hidden> - 1.4.8-1
>   - Update to m4-1.4.8

Is there a web page to any of the Fedora m4 pages?  I would be interested
in tracking if they have any other useful things that should be added
upstream.

> I tried removing SIGPIPE from lib/fatal-signal.c but that didn't help.

lib/fatal-signal.c is maintained in gnulib; it would be worth asking there
for more on what behavior should be expected.

- --
Don't work too hard, make some time for fun as well!

Eric Blake             address@hidden
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFxrIA84KuGfSFAYARAp99AJ9XB7QdlflLjbZRgBPIWvcvZyv9XACZAUbn
jum3xIDdbbxzz3SQ2hQlSEg=
=CRiH
-----END PGP SIGNATURE-----