On Mon, 21 Aug 2023, Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote: > On Mon, Aug 21, 2023 at 03:15:43PM +0300, Jani Nikula wrote: >> On Mon, 21 Aug 2023, Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote: >> > On Mon, Aug 21, 2023 at 11:29:27AM +0300, Jani Nikula wrote: >> >> On Fri, 18 Aug 2023, Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote: >> >> > but you might have (eg) >> >> > >> >> > Laurent Pinchard (Coca-Cola) <laurent.pinchart@xxxxxxxxxxxxxxxx> >> >> > >> >> > and then when working for another sponsor: >> >> > >> >> > Laurent Pinchard (Ford) <laurent.pinchart@xxxxxxxxxxxxxxxx> >> >> >> >> Just an observation, git shortlog -s/-se groups/distinguishes, >> >> respectively, the author and sponsor in Laurent's approach. Not so with >> >> Matthew's approach. >> > >> > Hm? >> > >> > $ git shortlog -s next-20230817.. >> > 1 Matthew Wilcox (Novartis) >> > 25 Matthew Wilcox (Oracle) >> > >> > $ git shortlog -se next-20230817.. >> > 1 Matthew Wilcox (Novartis) <willy@xxxxxxxxxxxxx> >> > 25 Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx> >> >> $ git shortlog v6.4.. -s --author="Laurent Pinchart" >> 12 Laurent Pinchart >> >> $ git shortlog v6.4.. -se --author="Laurent Pinchart" >> 2 Laurent Pinchart <laurent.pinchart+renesas@xxxxxxxxxxxxxxxx> >> 10 Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx> > > I must not be understanding your point correctly. That's probably a correct assesment, and not necessarily your fault. ;) My point was that git shortlog -s groups and adds up the commits together if the name is the same, but -se can still be used to distinguish the mbox+tag difference in email. 12 for -s, but 2 and 10 for -se in Laurent's stats above. If the names differ, even if just by the thing in parens, git shortlog won't combine them. > $ git shortlog v6.4.. -s --author="Matthew Wilcox" > 7 Matthew Wilcox > 1 Matthew Wilcox (Novartis) > 123 Matthew Wilcox (Oracle) > > That seems to me like it successfully distinguishes my fake commit on > behalf of Novartis (who I haven't actually worked for since 1997) from > my real commits on behalf of Oracle. It also shows a few places where > my commits weren't attributed to Oracle (I think this happens when I > send patches using mutt instead of git-send-email) I'm not saying either is right or wrong or has any shortcomings, I'm just saying they're *different* in this regard. BR, Jani. -- Jani Nikula, Intel Open Source Graphics Center
