Re: Uncommented code in LilyPond

[Graham Percival]

On Tue, Sep 04, 2012 at 01:46:28PM +0200, Janek Warchoł wrote:
> On Tue, Sep 4, 2012 at 1:24 PM, Graham Percival
> <address@hidden> wrote:
> > I think that being pickier with the reviews, possibly including
> > David's nice idea about only "replying" with updated patches, is
> > the best direction to move in.
> 
> Looks like i missed that part of the discussion...  could you clarify
> what do you mean by "replying only with updated patches"?

Sure.  Let's say that Mike posts a new patch that fixes something.
David adds a comment saying "why are you pushing this onto a
vector instead of using an array index?"

Currently, Mike would reply to the email or log in to Rietveld and
say something like "it's simple and speed isn't a problem here".

If we try the experiment of "replying only with updated patches",
then Mike would not be allowed to reply to the email.  Instead, he
would have to upload a new patch containing
    // push to a vector instead of using an array because
    // speed isn't a problem here

That way, no knowledge is lost through comments on rietveld that
nobody looks at after a patch is pushed.  This is also a good
exercise for writing readable patches -- if a developer has a
question about the patch you wrote, then it's not as readable as
you think it is!


I don't mean to pick on Mike here, so perhaps you'd volunteer to
spend a week doing the "reply via patches only" experiment.
(of course if there's a different discussion on -devel, then you'd
reply to those normally via email.  It's only the discussion for
your particular patch(es) that you would avoid using email (or
chats) for)

- Graham