Le mercredi 17 juin 2015, 00:03:28 André Hentschel a écrit :
Am 16.06.2015 um 23:33 schrieb Nikolay Sivov:
On 17.06.2015 0:07, André Hentschel wrote:
Oh man, i didn't plan to jump into that "flamewar"...
But while I’m at it, Jeremy, the only thing about CW that sometimes bugs me, is that CW devs often send their patches without any description. (Maybe that was also what Theodore means?) Maybe i keep finding only the "bad" ones, but it's always with patches that catch my interest for some reason, and then, no info. And that looks a bit like "AJ already knows what this is about and what it is going to fix, no need to mention it". And of course i'm far from perfect, so maybe i just picked the "bad" ones and never had a look at the "good" ones, and of course i also don't always provide a good description... I just wanted to throw in the only thing that bugs me about CW, don't be offended :)>
That's a good question, thank you for sharing that, but I don't believe it works like that, there's no internal debates or intensive communication that results in patches stripped from text description. Sometimes when there's a bug report for it people mention it in mail body, sometimes they don't, but later mention a fix on a bug report. So it's up to submitter mostly.
I see, maybe I should simply ask the submitter next time...
Hello,
I understand that asking for a description is maybe the way forward. How I see commit message is very well described by a Xorg developer: http://who-t.blogspot.be/2009/12/on-commit-messages.html
But I see that often when the patch finally gets committed then only the title stays. The text of the commit message has been removed. I presume it is because the message often states something obvious for expert developers with a very good knowledge of the history of the project. But for a non expert who tries to go through the history of code in order to understand it, it doesn't help. So I think that I wouldn't personally dare asking the submitter next time for a better commit message...
Cheers, Marc