On Tue, 1 Aug 2006, Steve Kargl wrote: > On Tue, Aug 01, 2006 at 04:23:57PM -0500, Matthew D. Fuller wrote: >> On Tue, Aug 01, 2006 at 02:16:57PM -0700 I heard the voice of >> Steve Kargl, and lo! it spake thus: >>> >>> If this change would have been noted in UPDATING, I would have >>> chosen a different day to deal with the mess that David Xu has >>> caused. >> >> Considering how recently changes in libpthread in libc 6 vs 7 were >> causing every threaded program to just dump core, this seems a rather >> harsh indictment. > > I've switch to libpthread and everything is working, again. > If I use libmap.conf to map libpthread to libthr, everything > is broken. David Xu committed a change that uses a libc.so.7 > function where libthr previously functioned fine with libc.so.6. > This change should be noted in UPDATING. If you don't like the > words I chose to use, then write whatever political correct > message you want and add it to UPDATING. The problem is really that libthr (and all other necessary libraries) need to have their versions bumped. It's not just thr_getscheduler. There were networking changes added that required the libc version bump. If there had not been a version bump to libc, adding thr_getscheduler() and friends to libc.so.6 would not have caused you or anyone else any grief. We add functions and system calls all the time to libc without any problems. It is only when library versions are bumped that cause problems. -- DEReceived on Tue Aug 01 2006 - 19:44:15 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:58 UTC