[Development] Qt Coding Guidelines

[André Somers]

Op 16/03/2016 om 20:47 schreef Ziller Eike:
>> On Mar 16, 2016, at 20:33, André Somers <andre at familiesomers.nl> wrote:
>>
>>
>>
>> Op 16/03/2016 om 16:14 schreef Koehne Kai:
>>> Hi there,
>>>
>>> We have had quite some discussions about the use of C++11 features and right API in the past on this mailing list - but if there has been a consensus (which is sometimes hard to find out), it was often buried pretty deep in the mailing thread. IMO it would be good to make decisions more explicit, and write them down also somewhere outside of this list.
>>>
>>> We already have the coding conventions page: https://wiki.qt.io/Coding_Conventions . But we haven't done a good job at keeping it up to date - and one reason is IMO that, given that it's a wiki everybody can edit, people in a twist of irony stay away from editing it to avoid editing wars.
>>>
>>> I've been contemplating whether we should instead use some more formalized decision process. We could have a document uploaded in git, and every change needs to be reviewed and approved by Lars. While at it, this fresh start would also be a good opportunity to check whether all the rules in above wiki page, and the structure of the document in general, can be improved.
>>>
>>> As sort of a demo I created
>>>
>>> https://github.com/kkoehne/qt-coding-guidelines/blob/master/qt-coding-guidelines.md
>>>
>>> What do you think? If nobody sees the value in this, I'll refrain from sinking more time into it.
>>>
>> Could work I think. But how do you propose these changes get announced? Who will be added to review such changes?
> Changes should still be discussed on this mailing list first (if they are not cosmetic). These discussions can result with a change on code review (also posted here).
> As something that effects the whole of Qt, it must be reviewed by the Chief Maintainer == Lars.
>
> (My 2c, we do it similar with Qt Creator, seemed to have worked fine enough so far.)
>
Doesn't that lead to having the discussion twice: first once in the ML, 
then again in the MR for the change?

André