GSoC Progress: Configuring Dive Computers via Subsurface
Dirk Hohndel
dirk at hohndel.org
Tue Jul 29 09:17:35 PDT 2014
On Mon, Jul 28, 2014 at 01:09:46PM -0700, Thiago Macieira wrote:
> On Monday 28 July 2014 08:06:58 Dirk Hohndel wrote:
> > I don’t have a firm rule for that.
> > I usually ask that commits don’t break the build (because that makes
> > bisecting so much more painful). I don’t like it when people send me
> > sequences like this:
> >
> > A
> > B
> > revert A
> > C
> >
> > (just rebase and drop A and revertA)
> >
> > Or if two consecutive commits are
> >
> > New code plus debug output
> > Remove debug output
> >
> > (just rebase and squash them together)
>
> It's more like:
>
> Add some code
> Add some more code related to the first
> Do something else
> Replace constants in commits #1 and 2 with enums
>
> It reads as a train of thought too: Josh started doing something one way and
> then realised he needed to refactor in order to be more flexible for different
> dive computers.
That seems perfectly valid to me.
> > But especially in longer series it seems to make little sense to try and
> > make them look perfect. Basically apply good taste and do what seems right.
> > If I hate it, I’ll yell :-)
>
> Understood.
>
> > And yes, I would love a copy of the repository where you have signed off the
> > commits that you have reviewed, Thiago. This will document that this was
> > GSOC work where a mentor spent the time and effort to review the code
> > before I pulled it.
>
> Will do and will try to squash things together where it makes sense.
Thanks
> By the way, I also told Josh that I would count the code he developed for
> libdivecomputer as part of his GSoC project, though I haven't seen any yet.
Yes, Jef and I had discussed that briefly - this only seems to make sense
as the two projects are very closely related and Josh's project kind of
straddles the border of both.
/D
More information about the subsurface
mailing list