[MPlayer-cvslog] r24941 - trunk/mplayer.c

[Michael Niedermayer]

On Mon, Nov 12, 2007 at 03:36:08PM +0200, Uoti Urpala wrote:
> On Mon, 2007-11-12 at 00:07 +0100, Michael Niedermayer wrote:
> > On Sat, Nov 10, 2007 at 07:50:55PM +0100, Diego Biurrun wrote:
> > > On Fri, Nov 09, 2007 at 12:21:53PM +0200, Uoti Urpala wrote:
> > > > - There are parts implying strict code ownership by maintainers and even
> > > > an explicit warning against against committing "trivial looking
> > > > fixes" (wtf?). Those should be replaced by something more flexible.
> > > 
> > > This is historic.  I think Arpi added it after I made some silly
> > > mistake, but I don't remember all the details.
> > 
> > many people made silly mistakes, it all could be described as people
> > commiting to code they dont maintain and dont fully understand
> > posting a patch which the maintainer of a specific part of the code
> > would review and comment on prevents that
> 
> People committing code without understanding what they are doing is bad.
> Avoiding that does not require strict code ownership.

what is this "strict code ownership" you speak about?


> 
> > so no doubt the policy can be worded better, but the spirit of it is
> > correct just the way its said is maybe not very clear but this falls more
> > in the nitpicking category IMHO than the "the policy is bad" one
> 
> The spirit of the text in svn-howto is wrong too. "You didn't understand
> what you were doing" is a valid complaint no matter what file you
> changed. "Your commit was correct but you should have asked the
> maintainer anyway" is not.

communication between developers working on the same part of the code is a
good thing, it prevents redundant work, it prevents people from causing each
other a lot of extra work with merging changes
and maybe most important it helps all developers who work on the same code to
keep track of what is changed, why it is and discuss such changes before they
are done

its a little bit like the differece between a stranger entering your
appartment through an open window and installing the lastest security fixes
on your computer vs. him knocking on the door and politely telling you
that he has found your system to be vulnerable and informing you how to
fix it and offering help to do so if you want ...
no doubt the end result is the same technically but absolutely not
socially
commit some good change to some code written and actively maintained by
someone else without any warning/discussion/notice/patch and the mainteiner
will often be angry and pissed, politely ask and its all fine
he also might have some suggestions on how to improve your change ...

also there is no such thing as a non trivial correct change, everything
larger than a few lines has a very good chance to contain a bug or 2 and
having other develpers (be it the maintainer or not) review and comment
on the changes definitly helps code quality


> 
> > the commit anything, anywhere system uoti seems to want (if i understood
> > him correctly) will lead to a mess i think theres no real disagreement here
> 
> Not "anything". Things you understand and can be reasonably certain are
> correct. If someone displays bad judgment about this then they can be
> told to limit their commits to the part of code they're familiar with.

who is going to tell them? and they almost certainly will feel singled out
and might leave the project as a result
just ask diego about the past when apri told various people who _did_ make
a mistake to keep their fingers off code except their own


> When someone does post a patch because he doesn't feel certain enough to
> commit directly it doesn't necessarily have to be the maintainer who
> verifies it. Usually a project should have multiple people who, even if
> not familiar enough with some parts of code, at least _understand_ when
> they're not familiar enough and so can be trusted to approve changes
> anywhere in the code.

we arent disagreeing here, if you think i have a different oppionion then
you missunderstood me

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

While the State exists there can be no freedom; when there is freedom there
will be no State. -- Vladimir Lenin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-cvslog/attachments/20071112/b707dd1a/attachment.pgp>