[17:21] <cjwatson> ogasawara: any updated plans for that kernel upload?
[17:21] <ogasawara> cjwatson: I'm still blocked on gcc-4.5
[17:22] <ogasawara> cjwatson: gcc-4.5 is still building on armel, and apparently looks like it's failed to build on amd64?
[17:22] <ogasawara> doko_: ^^
[17:22] <cjwatson> hm, yes
[17:22] <doko_> ogasawara: why do you wait? just upload
[17:22] <cjwatson> Function `GC_malloc_explicitly_typed' implicitly converted to pointer at /build/buildd/gcc-4.5-4.5.2/src/libobjc/objects.c:49
[17:23] <cjwatson> doko_: you know the thing where the kernel has an exact version of gcc encoded into it, and modules have trouble building if it mismatches ...
[17:23] <doko_> cjwatson: which "exact" version?
[17:24] <cjwatson> perhaps ogasawara can confirm, but from previous discussions I've been given to understand that the gcc-4.5 package version has essentially been made part of the kernel's module building interface
[17:24] <ogasawara> cjwatson, doko_: that was my understanding as well
[17:25] <doko_> then I can't help you, if you do this :-/
[17:26] <cjwatson> it's been that way for quite a while; I'm not exactly wild about it, but I believe that's why ogasawara wants gcc-4.5 to finish before uploading the kernel for beta-2
[17:26] <cjwatson> I can't find the mailing list discussions about it
[17:27] <ogasawara> doko_: we're planning to discuss this at UDS to see what can be done to not have this dependency, but is not something we can easily change now
[17:27] <doko_> cjwatson, lamont: so the previous build did succeed, with the very same warning, but the buildd did let it in the archive. is it really the time to change the buildd configuration at this time after the test rebuilds are done?
[17:28] <cjwatson> I acked that change and I thought it was before the test rebuilds
[17:28] <cjwatson> it's actually a restoration of previous buildd configuration that went missing by mistake, not an entirely new thing
[17:28] <cjwatson> it will probably be as fast to fix it as it would be to reconfigure the buildds and rebuild it
[17:29] <doko_> no, the gcc builds did all succeed in the test rebuild
[17:30] <doko_> except for getting the arm build in the archive only two days later
[17:31] <cjwatson> well, depends how quickly you think it will be possible to raise lamont
[17:31]  * cjwatson <- no buildd superpowers
[17:31] <doko_> did ping him today, no reply yet
[17:33] <cjwatson> personally, I'd say that having the armel build two days later isn't so bad; we don't need to wait for that in order to upload the kernel, since how many external module packages do we really care about on armel?
[17:33] <cjwatson> it'd be fine to upload the kernel once i386, amd64, and maybe powerpc are in
[17:37] <ogasawara> that's fine by me too
[17:50]  * ogasawara will check back in later.  cjwatson, I'll ping you when I've uploaded the kernel.
[18:15] <doko> lamont, cjwatson: I assume this check is not enabled on the ppa builders, so I didn't see it in the test rebuild
[19:44] <cjwatson> ogasawara: doko's uploaded a new gcc-4.5 (thanks!), which is building at the moment
[20:07] <ogasawara> cjwatson: ack, I'll keep an eye on the builds.