I have an important segfault when I create the index (via Ferret::Index::FieldInfos#create_index). I decided to run the tests, this is what I have : $> ruby test_all.rb Loading once Loaded suite test_all Started ....................EEEEEEEE./unit/../unit/index/../../unit/store/../../unit/analysis/../../unit/utils/../../unit/query_parser/../../unit/search/tc_filter.rb:11: [BUG] Segmentation fault ruby 1.8.5 (2006-12-04) [i686-linux] Aborted Dave, any idea ? -- Posted via http://www.ruby-forum.com/.
On 2/26/07, Florent Solt <florent at solt.biz> wrote:> I have an important segfault when I create the index (via > Ferret::Index::FieldInfos#create_index). > I decided to run the tests, this is what I have : > > $> ruby test_all.rb > Loading once > Loaded suite test_all > Started > ....................EEEEEEEE./unit/../unit/index/../../unit/store/../../unit/analysis/../../unit/utils/../../unit/query_parser/../../unit/search/tc_filter.rb:11: > [BUG] Segmentation fault > ruby 1.8.5 (2006-12-04) [i686-linux] > > Aborted > > > Dave, any idea ?I''ll try and get another version out tomorrow for people to try. Hopefully it will fix the problem. If anyone who is getting these bugs and is able to give me ssh access to their server, please let me know privately. Dave -- Dave Balmain http://www.davebalmain.com/
Florent Solt wrote:> I have an important segfault when I create the index (via > Ferret::Index::FieldInfos#create_index). > I decided to run the tests, this is what I have : > > $> ruby test_all.rb > Loading once > Loaded suite test_all > Started > ....................EEEEEEEE./unit/../unit/index/../../unit/store/../../unit/analysis/../../unit/utils/../../unit/query_parser/../../unit/search/tc_filter.rb:11: > [BUG] Segmentation fault > ruby 1.8.5 (2006-12-04) [i686-linux] > > Aborted > > > Dave, any idea ?After a few tests, it seems that this segfault is triggered when Ferret is compiled with GCC 4.1.*, _without_ -fstack-protector. Some Linux distributions (like Ubuntu) provide GCC with -fstack-protector enabled by default, hence the bug is triggered when compiling with -fno-stack-protector. This is why it looks like a distribution-dependant bug. I didn''t investigate further, but it looks like there''s a stack overflow somewhere. -- Maz -- Posted via http://www.ruby-forum.com/.
David Balmain wrote:> > I''ll try and get another version out tomorrow for people to try. > Hopefully it will fix the problem. If anyone who is getting these bugs > and is able to give me ssh access to their server, please let me know > privately. > > DaveI''m sorry Dave, I can''t :( The easiest way is maybe to test with in a VMWare or qemu. The linux I use is "Arch Linux", it''s a lightweigt one with rubygems packaged. As Maz said, you should not use Ubuntu because they add the "stack-protector". Florent. -- Posted via http://www.ruby-forum.com/.
On 2/27/07, maz <maz at rift.fr> wrote:> Florent Solt wrote: > > I have an important segfault when I create the index (via > > Ferret::Index::FieldInfos#create_index). > > I decided to run the tests, this is what I have : > > > > $> ruby test_all.rb > > Loading once > > Loaded suite test_all > > Started > > ....................EEEEEEEE./unit/../unit/index/../../unit/store/../../unit/analysis/../../unit/utils/../../unit/query_parser/../../unit/search/tc_filter.rb:11: > > [BUG] Segmentation fault > > ruby 1.8.5 (2006-12-04) [i686-linux] > > > > Aborted > > > > > > Dave, any idea ? > > After a few tests, it seems that this segfault is triggered when Ferret > is > compiled with GCC 4.1.*, _without_ -fstack-protector. > > Some Linux distributions (like Ubuntu) provide GCC with > -fstack-protector > enabled by default, hence the bug is triggered when compiling with > -fno-stack-protector. This is why it looks like a distribution-dependant > bug. > > I didn''t investigate further, but it looks like there''s a stack overflow > somewhere. > > -- > MazThanks Maz, that helps a lot. I wish I knew that earlier. -- Dave Balmain http://www.davebalmain.com/