On Mon, Jul 2, 2018 at 3:24 AM Greg KH gregkh@linuxfoundation.org wrote:
On Mon, Jul 02, 2018 at 11:57:36AM +0200, Geert Uytterhoeven wrote:
Pseudo-Shortlog of commits:
Given you do publish a git tree with these changes, is there any specific reason you're not using a real shortlog? Sorting and grouping would make it easier to see which patches of mine you have backported.
Many years ago Linus wanted it in this type of format as it was easier to see what was being applied to the trees.
I think you refer to an old email of mine where I said:
"I think Greg has a script he uses, so I'm cc'ing him here too. i'm not sure I really like his one-liner format, but I'm not sure what the right format would be. I *think* the format I'd prefer would be a two-liner one (with a newline in between), something like
git log --abbrev=12 --format="%h %aN <%aE>%n %s%n"
but that's more of a "maybe something along those lines.." than anything else"
Note that "maybe something along those lines", and in fact the format Greg uses doesn't have the commit ID in it, so it's more like
git log --format="%aN <%aE>%n %s%n"
I forget what the original format Greg used was, but I don't think it was "shortlog". This is from almost a decade ago, and I couldn't actually find any of Greg's emails with the original format that I found to be inconvenient. I think it was something really dense and incomprehensible based on my " i'm not sure I really like his one-liner format", but..
Maybe just "shortlog" would be better. I agree that it's likely easier for authors to see what commits got done from them, thanks to the author grouping.
Linus