On 26 June 2010 20:38, DJ Lucas <[email protected]> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > The last decision was that we would prepare for BLFS-6.5. I personally > have no interest in working towards a BLFS-6.5 release, but others may > differ. We could create a branch if there is still interest in > continuing along that goal, and now would be a good time that Gnome is > completed and deemed stable. >
Having been burned in 6.4, and lacking a 6.5 system, I too have no interest in 6.5. But perhaps the current svn could be released "warts and all" for 6.5 with a note that packages NOT labelled as known to build with 6.5 are left as an exercise for the reader ? It would be educational, maybe we could add a list of places to look for non-BLFS patches and build instructions. I've got a 6.6 system that I was intending to keep as a sort of "long term support" version (because it uses a long-term stable kernel), but I'm going to have to rebuild that because I can't seem to get html5 working on youtube with midori on gnome-2.28 [ and I just drop everything into /usr on my own builds - that might be about to change :) ] But equally, current LFS-svn looks interesting : the x86_64 kernel problem looks as if the fix should hit a released kernel soon, and the glibc build fixes work (it always scares me when current glibc has to be patched to build). But at the moment I don't have any experience of running gcc-4.5. I expect some degree of pain, just like for every previous minor gcc version increase. > I'd like to start working on current LFS-SVN. I haven't committed > anything in so long because of this and it is about time for a new > desktop build. Still working on my testing server which is somewhere > between 6.6 and current (some package updates to 6.6) so I can commit to > cyrus-sasl, postfix, openldap, samba, php, mysql, openssl, httpd, > dovecot, etc. I do not see that anything will conflict with current SVN > and I'll want to do a rebuild before it goes live anyway. > I'm not commenting on the $GNOME_CONFIG question, it's not something I've ever looked at, and I'm happy to go with whatever is decided. If we do go for either 6.6 or svn, the SBU/space measurements become a problem - x86_64 is fully supported in LFS. I happen to only build x86_64 these days, and I'm fairly sure one or two packages need special action to build correctly (e.g. passing -fPIC). I can't see any editor feeling disposed to build on both arches as a matter of course. If we move to LFS-svn, the tag "known to build against LFS-svn" is not particularly useful because things change. Arguably, a pair of "built on x86/x86_64 using LFS-svn-ccyymmdd" tags might have a little purpose, but doing that would only really show when a package was last tested. After Alexander has long departed, I'm in agreement that there is *too much* in the book - it can't all be kept up to date for lack of interest. Meanwhile, I'm off to update my buildscripts for icecat-3.6.4 (trying this instead of firefox for political reasons ) now that I've finally got it building on 32-bit ppc64. Only took 94 minutes to build, and that was after I'd found the magic words to get through xptc and then ipc. I think I might be a bit "the worse for wear" by the time I've built and installed it using my script ;-) Whatever is decided, I'll personally be trying to build something based on LFS with gcc-4.4.3 on my ppc64 before I move on (already done it on the ppc laptop, but ppc64 needs a full biarch toolchain so it can build its kernel - I'm expecting fun). I'm also committed to many non-computing things, so don' get the impression I'll suddenly jump back in. ĸen -- We will sternly refuse to take of strong waters, will we not ? Yes, we will not. By no manner of means. (Joyce, "Ulysses") -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
