Re: [patch] Recent findutils test suite failures

[Andreas Metzler]

On 2005-11-23 James Youngman <address@hidden> wrote:
[...]
> I had also been told that the many-short-arguments problem (supposedly
> fixed in 4.2.26) still persists on the hppa platform.  In the course
> of preparing the patch I discovered a possible cause for this (we were
> checking our argument count in two places, but the bugfix had only
> been applied in one).  I would like to know if this patch also fixes
> that problem.

> The patch should apply cleanly against findutils-4.2.26.

It does not: ,-)
address@hidden:/tmp/blah$ md5sum findutils-4.2.26.tar.gz
9ac4e62937b1fdc4eb643d1d4bf117d3  findutils-4.2.26.tar.gz
address@hidden:/tmp/blah$ tar xzf findutils-4.2.26.tar.gz
address@hidden:/tmp/blah$ cd findutils-4.2.26
address@hidden:/tmp/blah/findutils-4.2.26$ zcat 
~/xargs-ARG_MAX-variability-bug.patch.gz | patch --dry-run -p0
patching file NEWS
patching file lib/buildcmd.c
Hunk #2 FAILED at 230.
Hunk #3 FAILED at 269.
2 out of 3 hunks FAILED -- saving rejects to file lib/buildcmd.c.rej
patching file xargs/testsuite/Makefile.am
patching file xargs/testsuite/xargs.posix/arg_max_32bit_linux_bug.exp
patching file xargs/testsuite/xargs.posix/arg_max_32bit_linux_bug.xo
patching file xargs/testsuite/xargs.posix/arg_max_64bit_linux_bug.exp
patching file xargs/testsuite/xargs.posix/arg_max_64bit_linux_bug.xo

After applying manually the bug is still there, throwing in Tim
Waugh's one-line-change on top of it either fixes it or at least makes
it unreproducible with
yes|head -n 32766|tr '\n' ' ' |xargs/xargs echo > /dev/null
              cu andreas

-- 
The 'Galactic Cleaning' policy undertaken by Emperor Zhark is a personal
vision of the emperor's, and its inclusion in this work does not constitute
tacit approval by the author or the publisher for any such projects,
howsoever undertaken.                                (c) Jasper Ffforde