On 28.07.2015 15:48, Ivan Zhakov wrote: > On 28 July 2015 at 00:15, Stefan Fuhrmann <stefan.fuhrm...@wandisco.com> > wrote: >> On Mon, Jul 27, 2015 at 8:02 AM, Branko Čibej <br...@apache.org> wrote: >>> The 1.8.14 release artefacts are now available for testing/signing. >>> >>> Please get the tarballs from >>> >>> https://dist.apache.org/repos/dist/dev/subversion/ >>> >>> and add your signatures there. >>> >>> Thanks! >> >> Summary: >> >> +1 to release >> > [...] > >> Results: >> >> (Not a regression) The text delta code for old-style deltas segfaults >> when compiled with GCC 4.9 and -O3, e.g. with >> --disable-debug --enable-optimize. This failes >> lt-random-test 2: random combine delta test >> 1.9 does not have that problem and a fix has been >> proposed for backport. >> > Should we consider this problem as a blocker for the release? > > I agree that this is not a regression, but is it OK to make a release > that does not build properly with the *actual* compiler?
Since it's not a regression, I wouldn't treat this as a release blocker. --enable-optimize is not the default, and the obvious workaround is simply to not use it. Code built with the default configure options works, so it's not a question of not building properly. -- Brane
