Andre Oliveira da Costa
Fri Dec 31 13:28:00 GMT 1999
> I'd guess that this is a case-sensitivity (or non-sensitivity, in this
> case) issue. The tests were made to work under UNIX so if any of the
> filenames differ only in the case of the alpha characters, the test will
> fail on any Windows system.
I wouldn't expect the test scripters to write tests that involved files
which would only differ in the case of the alpha chars... however, if that
is indeed the case, you are right, the tests might have failed on cygwin,
and so I went to the 'same-file' test script to see if I could guess
something from it. As far as I could understand, the tests are failing on
the handling of symlinks/hardlinks. Any guess why is that happening?
I'm sending as an attachment the output of the error log, if anyone would
like to check it out. To obtain it, I uncommented the following line on the
# Uncomment this if you see a failure and want to try to diagnose it.
diff -u $expected $actual 1>&2
> You could change the test scripts or simply ignore any errors relating
> to "same-file" statements. Ignoring the errors may be difficult if the
> make exits for each error condition. You could try using the -i to
> ignore errors or a -k to abandon work on the current item but continue
> with the next. Check the man page for make for details on the options.
Thanks for the tips regarding the -i and -k flags, I didn't know about them.
However, I'm affraid of skipping these tests and running into problems later
(i.e. I don't know if it's safe to ignore them). I think I'll expect and see
if someone else makes a stand about it... otherwise I'll stick to the 3.16
version of fileutils that came with B20.1.
> Good luck.
Thank you very much for your help.
AndrÃÂ© Oliveira da Costa
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 5963 bytes
Desc: not available
More information about the Cygwin