bug#13524: Improving user experience for non-recursive builds

[Stefano Lattarini]

On 02/08/2013 10:11 AM, Peter Rosin wrote:
> On 2013-02-08 09:45, Peter Rosin wrote:
>> Stefano Lattarini wrote:
>>> Fine as well.  And of curse, if you want to speed thing up and have more
>>> control on the final result, feel free to shepherd the pending patches to
>>> the agreed form ;-) -- which if I'm not mistaken is:
>>>
>>>   - make the series consist of only two patches, one introducing the
>>>     feature (complete with documentation and NEWS additions, plus your
>>>     original test case), and one follow-up patch implementing my
>>>     testsuite enhancement;
>>>
>>>   - use the '%...%' form, and prefer lower-case for long names (so,
>>>     '%reldir%' a.k.a. '%D%' and '%canon-reldir%' a.k.a. '%C%').
>>
>> Done. I didn't address the canonicalization concern raised by Miles. That
>> seems like a bigger issue than this patch series. However, If that naming
>> is changing, it must be done before this series lands in a released
>> version, or we'll end up with bad comptibility hacks...
> 
> Errm, compatibility...
> 
> Forgot to ask, but should I push out this non-fast-forward to the
> experimental/preproc branch?
>
Yes please.  I will rewind it again once I have, as you suggest, fixed
my "demo test" not to needlessly fiddle with '/dev/full'.

> That would be 'git checkout experimental/preproc' followed by
> 'git push -f origin', right? I did rewrite that branch locally
> assuming the old branch was destined for the bin, but I have
> never actually done such a rewind of upstream before...
>
I fear that Savannah forbid such simple for of non-fast-forward pushes.
Here is what you can do to still force the rewind:

  $ git branch -m experimental/preproc x
  $ git push origin :experimental/preproc
  $ git branch -m x experimental/preproc
  $ git push origin experimental/preproc

Thanks,
  Stefano