Opened 19 years ago

Closed 19 years ago

#1062 closed defect (duplicate)

GCC 4.0.0 branch

Reported by: opslynx@… Owned by: lfs-book@…
Priority: lowest Milestone:
Component: Book Version: SVN
Severity: normal Keywords:
Cc:

Description

GCC 4 has been branched, therefore it's ready to be included in SVN

Change History (5)

comment:1 by Matthew Burgess, 19 years ago

I don't think it's quite ready for SVN yet. glibc currently doesn't build with it, and won't until 2.3.6 (see http://sourceware.org/ml/libc-hacker/2005-03/msg00022.html and http://sources.redhat.com/bugzilla/show_bug.cgi?id=721). Experience has shown that gcc-x.0.0 releases typically are less stable than future versions and therefore may not be suitable for an LFS release. However, once glibc-2.3.6 is out (and perhaps shortly before) I see no problem creating a proof-of-concept branch to test things out with it. I'd reckon on such a branch being of use in around 4-6 weeks (e.g. mid-April).

comment:2 by opslynx@…, 19 years ago

It seems that there's already a patch for glibc to address the issue by Mark Loeser, although I don't know how reliable it is, given the fact that he said it himself that he's new to glibc hacking. I expect the issue to be fixed in glibc head within the next few days. But I don't think that you guys are willing to go back to glibc from cvs. So the whole thing might be too sketchy to be included even in unstable, especially given the new change in the development model.

comment:3 by Matthew Burgess, 19 years ago

http://sourceware.org/ml/libc-hacker/2005-03/msg00008.html has the patch for HEAD, though Roland McGrath isn't seeing the breakage that myself and Andreas have. Personally, getting glibc and gcc to play ball has given me enough headaches for a while, so I'm going to just wait until they produce an official release that cooperates nicely. As you've seen, upstream are aware of the issues, and there's plenty of time before we need to consider having gcc-4.0.x in the book.

Added to all this is the fact that gcc-4.0.0 snapshots don't build on a glibc-2.3.4 based machine (due to it pulling in pthread.h which has invalid C in it). This means that gcc-4.0.0 will have to wait at least 1 LFS release anyway, so we can put a fixed glibc in either by way of a new upstream release (2.3.6) or by patching it up ourselves (the patch is trivial, justifying it's inclusion is another matter).

comment:4 by jim@…, 19 years ago

GCC 4.0.x will not be added until it's released and the development team has done some testing. We don't know what impact this packages will have on other packages during the build. We know that glibc will not build, and that's expected to be fixed in 2.3.6.

comment:5 by Matthew Burgess, 19 years ago

Resolution: duplicate
Status: newclosed

* This bug has been marked as a duplicate of 33 *

Note: See TracTickets for help on using tickets.