Re: GNU diff or not GNU diff, that is the question...

Karl Fogel <kfogel@newton.ch.collab.net> writes:

> What makes you so sure it's a regression? :-) Maybe it's always been
> there...

No it hasn't. My test used to be able to commit its test file this
way (and then it would fail later on of course), but now I had to
change it into committing the directory in order to reach the place
where it was supposed to fail.

> Can you package up your reproduction recipe as a new test (i.e., a
> patch to commit_tests.py) and post it here? (I mean, if you want to
> fix the bug too that's even better, but if not, just having the test
> case is more than half the battle for whoever does fix it.)

Roger Wilco. I don't have time to look for the bug at the mo, but a
patch to commit_tests.py is attached (good thing "svn diff" works
again... :). It's basically the same as the old "commit_one_file"
test, but with the file of choice changed to one of the newly added
ones. Hm, maybe there should be a test for commiting a single remove
as well? Maņana...

  // Marcus

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org