On Sun, Jan 15, 2012 at 10:39 PM, Uli Schlachter <[email protected]> wrote:
> Heh, I make other people send in bugs. :-)

This was fun overall. Thanks, Uli :).

> Yeah. This feels messy. Why does all the code everywhere use a screen index
> instead of screen objects?

No clue. I would have though performance, but Lua passes arguments as
a reference.

>> git fetch origin isn't returning anything. You messaged too early?
>> Or server issues?
>
> I just said "merged", not "pushed". ;-)

Bah :-|.

[...]
>> This is something I picked from git.git (no, I am not a git
>> contributor, but I do read the code and git log and git mailing list
>> sometimes):
[...]
> I never saw such a commit, I think. However, I don't read any git mailing 
> lists
> either.

It is fun to observe different communities. Particularly if there are
'big shots' in there. You get to learn some new cool things often. A
lot of time things go over my head also.

> So if I edit your patch and introduce a bug, wouldn't that mean that you
> get blamed for it? Or what if I change your patch in a way that you don't 
> like?

A patch series can go through a lot of iterations. Such iterations
need not only involve the author and the maintainer. Anyone on the
list, can participate in the ensuing discussion. Sometimes it makes
more sense to just send in a review. For example, when you spot a
mistake even without applying and testing the patch. At other times,
it makes sense to amend the patch and re-send it on the list for
further reviews (of everyone participating in that patch series). For
example, when you apply a patch and spot a typo in a variable name
when compiling/running/some other way of testing. While the patch is
on the list, maintainer is just a reviewer with more experience and
influence. Once the series is finalized on the list, the maintainer
applies the series.

Yes, I get blamed for introducing a bug. Because if you introduce a
'bad' change in my patch, as the author, I am responsible to protest
such a change, no? Besides, git blame is not a punishment tool.
Something broke, and you just want to find a person to contact to get
it fixed or ask questions. In the above style, everyone involved in
_evolving_ the series is logged. If the author is unreachable, there
are others you can get in touch with.

Of course, these is nothing sacrosanct about this strategy. Tens of
such iterations probably don't make sense for smaller, less complex
patches. You use your better judgement and evolve good practices over
time.

There are other interesting things in Git's git log. See the signed
off section of this commit:
https://github.com/git/git/commit/024c843d472658a3dabb344c843c9242c2f7e4de

> I don't like editing other people's patches and then pretending they did 
> that...

I understand. But it is more like, 'Hey, the author made this mistake,
but I fixed it. Everyone cool? I will merge it then.'

P.S: I hope it all makes sense.

-- 
Anurag Priyam

-- 
To unsubscribe, send mail to [email protected].

Reply via email to