Re: new coreutils snapshot; are they useful?

[Matthew Woehlke]

Matthew Woehlke wrote:
> [blah blah] it built clean, which is an improvement 
> over last time when as you'll recall I had problems with the strstr 
> declaration. I'll be pleasantly surprised if it passes 'make check' 
> cleanly, as that tends to hit false positives and/or kernel bugs, but 
> it's looking good!

Wow! Well, I am pleasantly shocked :-). 'make clean' did indeed pass 
clean. I even managed[1] to run check-root and (as root) the chgrp tests.

Ultimately, this is what I actually ran:
'make check' w/o chgrp on NFS
'make check' w/o chgrp on local
'make check' w/ chgrp as root on local (rm/fail-eaccess fails[2])
'make check-root' on local

[1] Because a: I can't create root-owned files on NFS, and b: my AIX box 
literally has almost no local HD space, I had to copy the tests 
directory to local and fill the parent with symlinks to the NFS where 
the rest of the real files are. This brings up an interesting point; it 
might be useful to have some way to run the test suite in a different 
directory, so that it is easy to run once on NFS and once on local. In 
particular I remember that on Darwin this was important.

[2] This isn't surprising, as the test *should* be skipped when run as 
root. See the other e-mail I just sent here.

--
Matthew
"I like to think of it as 'unplanned dissonance'" -- A fellow chorister