Re: df --output

[Pádraig Brady]

On 11/07/2012 11:48 AM, Bernhard Voelker wrote:
> On 11/07/2012 12:23 PM, Pádraig Brady wrote:
> > I've rolled the patch set up to 3 patches here:
> >     http://www.pixelbeat.org/patches/coreutils/df--output-merge.patch.bz2
> > which I'll merge later today.
>
> Hi Padraig,
>
> thanks - although I still don't see much gain to have such jumbo
> commits instead of ~20 little, on-topic ones which would also make
> it easier for future looks back in git history in order to find out
> why something has been changed the way it has been changed, see e.g.
> the build system rework of Stefano. Your choice.

Yes it's a trade off between short term and long term info.
There is no value for example in keeping my whitespace fixup
patch separate, nor fixes that were particular to the local branch
(like the mem leak fix) as they may confuse future investigations.
In future when doing `git blame` we can see immediately the change
related to df --output.  Also the shortlog generated in release
announcements is more feature focused than implementation focused.
So while I'm not convinced on the roll-up approach, it seems
like the better approach going forward.

> BTW: should we somewhere mention why we did all this?
> The discussion started with http://bugs.gnu.org/10915 .

I'll add that thanks.

> BTW2: I think you should mention yourself in a contribution
> line or alike, shouldn't you?

Committer field is fine for me :)

cheers,
Pádraig.