RE: [Monotone-devel] Solaris breakage

[Kelly F. Hickel]

> 
> On 7/24/07, Patrick Georgi <address@hidden> wrote:
> > truss log attached. If there is anything more you need to know, I can
> > try to coerce dtrace into telling me.
> 
> Thanks.  That's suggestive... but I think you have something else
> wrong. You're getting a boost::filesystem exception thrown, which
> should never happen anymore.  Can you repeat the test with a brand new
> checkout of the tip of net.venge.monotone, please?
> 
> Having said that, the actual failure is ...
> 
> rmdir("/home/patrick/work/monotone-
> build/tester_dir/clone_warning_with_multiple_heads/test_dir1/_MTN")
> Err#17 EEXIST
> 
> ... which is not EINVAL, but I still have no idea how that could be
> happening.  Could you please try
> 
> truss rm /home/patrick/work/monotone-
> build/tester_dir/clone_warning_with_multiple_heads/test_dir1/_MTN
> 2>&1 | grep rmdir
> 

[Kelly F. Hickel] if you meant "truss rm" then the output is "is a directory".  
Assuming you meant "truss rmdir", the output is below (from my system), but I 
suspect it isn't going to help you because the rmdir succeeded.

If you suspect ZFS, I can try to run the tests on a non zfs filesyetm, if I can 
find some disk space to do so....

execve("/usr/bin/rmdir", 0xFFBFFBC4, 0xFFBFFBD0)  argc = 2
resolvepath("/usr/bin/rmdir", "/usr/bin/rmdir", 1023) = 14
stat("/usr/bin/rmdir", 0xFFBFF9A0)              = 0
rmdir("/export/home/mtbldbot/slave-dir/full-sparc-solaris10-sunstudio11/build/tester_dir/clone_warning_with_multiple_heads/test_dirA")
 = 0

-Kelly