Re: [PATCH] .mailmap: Update mailmap
From: Ævar Arnfjörð Bjarmason <hidden>
Date: 2021-09-11 01:41:06
On Fri, Sep 10 2021, Junio C Hamano wrote:
quoted hunk ↗ jump to hunk
Jeff King [off-list ref] writes:quoted
On Fri, Sep 10, 2021 at 03:22:47PM +0000, Gwyneth Morgan wrote:quoted
The line for Jessica Clarke should probably just be Jessica Clarke [off-list ref] That works the same and doesn't put a reference to an old name.Thanks, that's a good suggestion. I kind of wonder if these mass mailmap-cleanup patches are a good idea in general. They are making assumptions about how people want their names to be represented, and whether and how they want any mappings to appear. Maybe that's something we should be leaving to people to propose for their own identities. Of course people who aren't active in the project anymore may not bother to do the cleanup, and of course messy data makes me sad. But on the whole, I'm not sure it's that big a deal either way.I am not enthused by the idea of replying to this thread, knowing that many of the CC'ed addresses will bounce X-<, but I agree with you on all three counts. Even for those who are no longer active, it makes sense to unify multiple idents that are spelled differently to help "git shortlog", but which one to unify to is not something we can decide without their input. Which leads me to suggest something like the attached patch. I wrote "Please notify us" for those who are no longer active and forgot how .mailmap entries are spelled to ask for help correcting. Of course, the updated instruction does not prevent a motivated volunteer to contact the people _individually_ and then send in a patch with entries that the volunteer secured consent, perhaps in the form of Acked-by ;-) .mailmap | 5 +++++ 1 file changed, 5 insertions(+)diff --git i/.mailmap w/.mailmap index 9c6a446bdf..20b581c879 100644 --- i/.mailmap +++ w/.mailmap@@ -4,6 +4,11 @@ # and/or not always written the same way, making contributions from the # same person appearing not to be so. # +# If you find an incorrect entry that affects yourself, please notify us +# at <git@vger.kernel.org> and suggest corrections. Because the way people +# want their names to be represented varies, please refrain from touching +# entries for other people unless you positively know that the updated +# entries are what they want. <nico@fluxnic.net> <nico@cam.org> Alejandro R. Sedeño <asedeno@MIT.EDU> <asedeno@mit.edu>
That seems too high a burden for edits that are likely to be
uncontroversial.
I.e. this seems targeted at someone who's changing someone's name
without their approval after the fact and when they can't be contacted.
Whereas most if not all edits to this file are likely to be janitorial
work such as de-duplicating E-Mail addresses, or cases where the people
involved even if they can't be contacted have already shared this
information unambiguously with the world, it just happens to be in the
mailing list archive, not in .mailmap.
E.g. I'd think these were fine, even assuming you can't contact the
parties involved:
* The ML history reveals someone who's clearly the same person was
using N E-Mail addresses in succession, we should probably map it all
to the latest one, unclutters shortlog and the like.
* Ditto even for their name in some cases. Is someone's commit history
200 commits of abbreviating their middle name, with 1-2 commits at
the start where they don't? Seems fine to just normalize that.
* Is the E-Mail they last used bouncing their E-Mails ("this person
doesn't work here anymore"), domain expired etc?
It would be useful to other contributors to map that to
this-is-unreachable@example.org or whatever so they won't waste time
CC-ing a bad address.
Which doesn't mean that there aren't things we shouldn't be doing
without asking:
* Changing a name from Alice to Bob? Yes, ask and have the person ack
it.
* Found the current E-Mail address someone who contributed years ago
but not under that address to git.git? Ask them first, they may not
wish to be contacted at all.
etc.