[Bug c++/11489] [3.4 Regression] ICE/segfault on legal code

["bangerth at dealii dot org" <gcc-bugzilla at gcc dot gnu dot org>]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11489



------- Additional Comments From bangerth at dealii dot org  2003-07-11 14:01 -------
Markus, there's not much I should say more, but the requirements for a check-in
are quite clear: run the testsuite and do not observe a regression. Other than
that, as I said before, some of the testing is relegated to other people with
their nightly builds. Sometimes they find bugs, then they report them. Given that
I am the single biggest reporter of C++ bugs (>100), I can only testify that Mark
et al have always been very helpful, timely, end efficient in fixing newly 
emerging regressions that weren't caught by gcc's own testsuite. Every fix of
a report includes adding a testcase to gcc's testsuite, so that this particular
problem doesn't reappear.

If you doubt the overall improvement in the C++ part: before the new parser was
merged, we had approximately 570 open C++ reports. We are now at 370. Of these
a significant number is new and thus regressions. I have no reason to doubt
that they will be fixed in time for 3.4, and that gcc3.4 will go out with less
than 350 open C++ reports. I personally consider this a) progress, and b) evidence
that the work those of us who are setting up nightly builds for their software
is doing good. If you keep testing your software, chances are good that 3.4
will be able to compile it. If you are not willing to, go spend money (instead
of time) on a compiler for which you can purchase premium support.

W.