Build failure in helpcontent2

[Norbert Thiebaud]

On Tue, Mar 27, 2012 at 6:29 AM, Michael Meeks <michael.meeks at suse.com> wrote:
>
> On Tue, 2012-03-27 at 09:00 +0100, Mark Stanton wrote:
>> Thanks for your help.  Looks like you're right.  The branch output is
>
>        This happens quite a lot, even to me ;-)
>
>        I wonder - how hard would it be to add a feature to core/g to check
> that the checkout is consistent, and add that to configure.in :-)

That would be just another kind of pita...

1/  the additional git repos are not there at initial configure...
they get cloned during make fetch. so you can't check them. and if
they were there ./g would have worked to start with.
2/ it is quite common for people to only branch core when then do
feature branch or even local branch to test some patch or the like...

That problem would be solved by git submodule btw. since checking out
core would also select the 'right' commit on all submodules....

I have not taken the time to play more with that, but their maybe a
way to have it do the right thing from git by just tweaking
.git/config (*)
in which case all that would be needed is a 'setup' script to set the
right value in .git/config once an for all
(of course the main pita would be the transition before-after
submodule... I haven't found a satisfactory solution to allow for
checking-out to a point prior to the 'sub-modulization' or
vice-versa... but that is a 'temporary pain.. just like onegit was...

Norbert

(*) submodules support would also require gerrit 2.3 which is still  in Beta.