DJ Lucas wrote: > On 11/26/2010 10:45 AM, Bruce Dubbs wrote: >> Matthew Burgess wrote: >>> On Wed, 24 Nov 2010 23:47:41 -0600, DJ Lucas <d...@linuxfromscratch.org> >>> wrote: >>> >>>> Can somebody else on a mult-core CPU, on 32-bit, confirm before I >>>> troubleshoot further? >>> This is also being tracked upstream and at RedHat: >>> >>> http://lists.gnu.org/archive/html/coreutils/2010-11/msg00083.html >>> https://bugzilla.redhat.com/show_bug.cgi?id=655096 >> Should we revert to .6 until this is fixed? >> >> -- Bruce > > My previous message did not get through for whatever reason. Quick hack > was to set nthreads=1 if using multibyte locales on or around line 4538 > in sort.c. I used getenv on LC_COLLATE, and compared only to C, but > there is certainly a better way to do that, I just don't know how...hack > works for me.
I saw the previous message. The idea of fixing packages in LFS/BLFS is not, IMO, setting a good precedent. Adding a patch or sed from upstream is one thing, but we shouldn't be taking over the developers job. That's not to say that you shouldn't contribute to upstream and incorporate that when it's accepted. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
