> ... my only intention was to take what
> I had the feeling was the boring part of reading through the bugs and
> let us fix the problems later together.
I very much appreciate the effort you spent in doing bug triage. It
has been quite impressive.
Please don't stop if you're inclined to continue, but please be sure
you can build the library and the unit tests after each patch
application. I don't mind so much if occasionally a unit test breaks
at runtime as we can go back and fix it prior to release. But when
compilation breaks we lose the ability to do automated bisection to
identify the origin of regressions.
> Shall I branch my patches out and revert the trunk?
Nah, I got trunk compiling again last night. That said, I do now see
loggertestcase : -log4cxx: No appender could be found for logger (x).
log4cxx: Please initialize the log4cxx system properly.
-Line 396: Level::getError() == a0->getEffectiveLevel() was
expected to be true, was false.
FAILED 1 of 10
Failed Tests Total Fail Failed %
===================================================
loggertestcase 10 1 10.00%
which I'd not noticed before.
I am concerned that https://issues.apache.org/jira/browse/LOGCXX-337
is not correct. If anyone knows that section of the code and/or the
threading particulars, please review threadcxx.cpp after r1567521.
- Rhys
