Re: [PATCH] .mailmap: Entries for Alex Bennée, Deskin Miller, Vitaly "_Vi" Shukela

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 21, 2010, 19:34 UTC
Message-ID: <7vk4t5qxii.fsf@alter.siamese.dyndns.org>
In-Reply-To: <4BA620CD.3060101@drmicha.warpmail.net>

Michael J Gruber <git@drmicha.warpmail.net> writes:

> Maybe I should have cc'ed Vitaly, the others are straight typo fixes.
> (One may even argue that Vitaly's change is, too.) Smothering on the
> backburner for a while, is fine, though ;)

Nico is right; what I wrote sounded that the patch will be forever on hold until I get all three acks, but that wasn't my intention.

Show 8 quoted lines

> Also, if used without "-e", shortlog coalesces authors by full name. I
> reckon this is the most common way it is used, and it's the mode for
> which I provided cleanup.
>
> If you distinguish by e-mail (-e) then there are many obvious multiple
> entries, but it's not up to me to decide which is the main e-mail
> address. One could take the latest one used, but some people distinguish
> between work for hire (maybe for several employers) and private submissions.

True.

We might want to allow projects to use different mode of operation if they so desire, though. Linus Torvalds at OSDL and Linus Torvalds at L-F may be one same person, and the person who is summarizing may want to coallesce his commmits like this, even under "-e" option:

    Linus Torvalds <torvalds@evo.osdl.org>,<torvalds@g5.osdl.org>,<torvalds@linux-foundation.org> (1093):
      Initial revision of "git", the information manager from hell
      Add copyright notices.
      ... (many more commits) ...

while there may be more than one John Smiths that we can differenciate only by email and they get their own separate group as different people (each of them may have more than one email addresses).

If a grouping of set of name-mail pairs per individual is kept somewhere, that can be used to enhance shortlog to support such a use case [*1*]. And the UUID thing could be one possible implementation (you make name-mail pair to first map to some third identifier, and treat name-mail pairs that map to the same identifier as "one person"). But I don't think it would work well if you forced the maintenance burden of such a mapping to individuals.

[Footnote]

*1* unless there is a situation where one person (say "Junio C Hamano") stopped using one e-mail address (say "junkio@cox.net") s/he has been using, and then a different person with the same name came and started using the same address, that is.