[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Serious bug in GUILE rational handling.
From: |
Kevin Ryde |
Subject: |
Re: Serious bug in GUILE rational handling. |
Date: |
Sun, 24 Dec 2006 20:43:25 +1100 |
User-agent: |
Gnus/5.110006 (No Gnus v0.6) Emacs/21.4 (gnu/linux) |
Han-Wen Nienhuys <address@hidden> writes:
>
> I've added one (CVS HEAD)
You can put tests in the branch for bugs fixed in the branch.
Oh, and the popen.test was indeed my fault, indirectly. I stuck in a
configure test for pipe(), because it doesn't exist on mingw, and that
unintentionally enabled this bit of scm_display, write and write_char
scm_prin1 (obj, port, 0);
#ifdef HAVE_PIPE
# ifdef EPIPE
if (EPIPE == errno)
scm_close_port (port);
# endif
#endif
I've commented that out in the three places. I'm pretty sure it's not
reliable to test errno at that point. I think popen.test had an EPIPE
left from a long previous operation for instance, and it made
scm_display close the check-guile.log file port. :(
I can't tell where those lines came from, they're there in 1.6, but
disabled on account of no HAVE_PIPE. Shows what you get for changing
the configury I guess ...