[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Issue 1356 in lilypond: LilyPond-style comments embedded in a Scheme

From: Graham Percival
Subject: Re: Issue 1356 in lilypond: LilyPond-style comments embedded in a Scheme expression can't include special characters
Date: Mon, 28 Nov 2011 17:39:37 +0000
User-agent: Mutt/1.5.21 (2010-09-15)

On Mon, Nov 28, 2011 at 11:16:55AM +0100, David Kastrup wrote:
> Graham Percival <address@hidden> writes:
> > Too hard for people with less than 5 hours of training to
> > determine.
> I am not sure I understand that rationale.
> git log origin
> Is it there or not?  Can you think of a _simpler_ test that works for
> figuring out something is in _staging_?


I cannot teach somebody to use git in less than 5 hours.  All the
evidence from the past years of lilypond development suggests that
we cannot teach people to use git in less than 5 hours.

The idea behind the bug squad is that unskilled users -- no git,
no linux, no programming experience -- can help out with a minimal
time investment (20 minutes a week).  We require that they are
able to read+write emails in English, and use a web2.0 facebook on
par with the difficulty of using facebook.  That's it.  There's
also about 10 pages in the CG they need to read.  If training
takes more than, oh, say 3 sessions (i.e. 60 minutes total), then
most potential volunteers will give up.  Less than 50% of bug
squad volunteers are active after 4 weeks of initially
volunteering.  We should not be making it harder to do the bug
squad duties.

Now, maybe the best answer is "bug squad: don't try to verify
'patch' issues, leave all those for an experienced developer to
examine".  But I don't see the point of that.  It's pretty rare of
us to lose patches -- say, maybe 2% of patches get lost?  And that
was before all the Patchy stuff.  Having a simple check of "is it
in webgit or not" will probably catch 90-95% of the 2% of patches
that would have gotten lost.

> Note that it means _nothing_ at all if Savannah knows about the commit
> id.  There are a number of possible reasons for that (for example, it
> knows about something I accidentally pushed then removed again, and it
> knows about everything pushed to private branches).

That's a good point; I didn't realize about private branches.

Still, I think that to some extent this is a solution looking for
a problem.  (or maybe "a problem looking for a problem"?)
How much grief could we have avoided if we had a developer
checking for patches in the git log for master?  I can't think of
any grief this would have avoided.

Frankly, I don't think it would be horrible if nobody even _tried_
to verify patches and just automatically marked them as verified.
But if people want to do a bit more than that, I think that
checking for the commit in savannah is a decent trade-off.

- Graham

reply via email to

[Prev in Thread] Current Thread [Next in Thread]