On Jan 16, 2006, at 6:50 AM, Steck, David wrote:
I second this request. I've been watching for this to be committed
since it was mentioned on the mailing list, since I think I'm
affected by it, too.
Thanks,
-David Steck
[EMAIL PROTECTED]
Looked at this today. Yep, looks like an ancient transcription
error, so the fix looks good. The hard part is trying to figure out
how to right a unit test that checks the bad behavior. I modified
log4j to misbehave and then ran its unit test suite and there were a
few tests that behaved oddly, but none that specifically tested that
piece of functionality. I wrote a unit test that passed with the
good code and failed with the bad for log4j, but when I transcribed
it to log4cxx it would pass regardless of whether the fix was in.
I'll spend some more time on it tomorrow and will just commit the fix
without tests if necessary, but I won't feel good about it.
