[11:20] hi folks [11:21] if I want to recompile the ubuntu kernel with a different gccc variant than the one behind "gcc", what files do I need to change, or how would I generally achieve that goal? [11:43] Madkiss, HOSTCC and COMPILER define the compilers in use for the various parts [11:43] in Makefile [11:44] okay. [11:47] i presume you could change kmake in our packaging to pass those if you are using that [11:48] i am using whatever linux_3.13.0.orig.tar.gz comes with. [11:51] Madkiss: I thought you got the problem fixed properly? (I.e. in ways that don't require another compiler) [11:51] the one problem. [11:52] Madkiss: Ah :( [11:52] i suspect there are at least three other problems caused by compiler madness, and I simply want to check. [11:52] apw: Do modules generally have to be built with the same compiler as the kernel? [11:52] soren, yes, as you are relying on the calling conventions on internal functions be exactly the same [11:53] apw: Ok. I wasn't sure how well defined the ABI was. Thanks. [11:53] where does the Build-Depends line come from for christs sake? [11:53] I edited debian.master/control.stub to add stuff there, and in the .dsc file, my changes are gone [11:53] debian/control is built from debian.master/control.stub.in [11:53] wait. [11:54] so you are having a control.stub.in that creates a control.stub [11:54] and that creates a control file? [11:54] yes [11:54] are you kidding me? [11:54] no [11:54] who on earth would do that?! [11:54] who would guess how it works instead of reading debian/rules to find out [11:54] people following the, err, debian policy, I guess. [11:55] the stub.in has variable subs to .stub, stub is combined with the other stubs per flavour to make control [11:55] seems to make sense to me [11:55] Debian policy doesn't specify how those things are created. [11:55] fwiw [11:56] * soren wanders off again [12:00] apw: and what'S the rationale for copying the changelog around ? [12:01] Madkiss, it is part of the package meta data which is derivative specific, allowing rebased derivatives to use the same master unpinnings without collission [12:02] allowing both us and non-official kernels to base of our maintained base in a simple manner [12:03] aha. [12:13] /b #is [13:33] apw: any idea what is causing the problem at the end in https://launchpadlibrarian.net/192033308/buildlog_ubuntu-trusty-amd64.linux_3.13.0-40.69%2Bsyseleven2_FAILEDTOBUILD.txt.gz ? [14:52] apw: any chance you could help me out? :\ === txspud is now known as txspud|afk