PATCH, RFC: handle debug options for rawhide differently
Thorsten Leemhuis
fedora at leemhuis.info
Wed Nov 21 17:41:04 UTC 2012
Lo!
BTW: @dave, thx for reviewing.
On 20.11.2012 19:57, Josh Boyer wrote:
> On Sun, Oct 21, 2012 at 10:00:23PM +0200, Thorsten Leemhuis wrote:
>>
>> How does the patch (against master) look to you guys? Did I miss anything that is needed? Please note that all the changes the patch does to the config-foo files are just those that a "make release" does now. And a lot of stuff is removed from the Makefiles, hence the patch looks big.
> So... I think I might have found one thing. See below.
>
>> nodebug: release
>> @perl -pi -e 's/^%define debugbuildsenabled 1/%define debugbuildsenabled 0/' kernel.spec
>> -release: config-release
>> +release:
>> @perl -pi -e 's/^%define debugbuildsenabled 0/%define debugbuildsenabled 1/' kernel.spec
>> @perl -pi -e 's/^%define rawhide_skip_docs 1/%define rawhide_skip_docs 0/' kernel.spec
>> @rpmdev-bumpspec -c "Disable debugging options." kernel.spec
> Those two targets aren't mutually exclusive. So the following bit:
> [..]
> Basically turns a release kernel that doesn't have the debug flavour
> enabled back into a debug kernel. E.g. if you're in rawhide (or
> really, wherever) and you do:
>
> 'make release; make nodebug'
>
> you wind up with litterally no changes in the configs. However, that
> isn't what is expected. The 'make release' call turns off the debug
> options, and the 'make nodebug' call disables the debug flavour (so
> there's no kernel-debug) build.
Yeah, I think I see the problem. Sorry for missing this. If I got this right,
then this could be solved easily by adding this line to the release target:
@perl -pi -e 's/%define with_release.*/%define with_release %{?_without_release: 0} %{?!_without_release: 1}/' kernel.spec
The debug target then obviously needs the opposite:
@perl -pi -e 's/%define with_release.*/%define with_release %{?_with_release: 1} %{?!_with_release: 0}/' kernel.spec
Or am I missing something? Note, I hope I got the two calls above the right way
around, I haven't tested this properly, as I'm on the run, but wanted to answer
your mail.
Cu
knurd
More information about the kernel
mailing list