Re: [edk2-devel] [PATCH 0/8] .mailmap: Add mailmap file to have a cleaner git history

["Philippe Mathieu-Daudé" <philmd@redhat.com>]

On 12/3/19 7:51 PM, Laszlo Ersek wrote:
> Hi Phil,
> 
> (heavily trimmer CC list)

Good idea :)

> On 11/26/19 16:08, Philippe Mathieu-Daudé via Groups.Io wrote:
>> Hi,
>>
>> The .mailmap git feature helps fixing commit mistakes (in name/email).
>>
>> The easiest way to use it is with the --use-mailmap flag:
>>
>>    $ git log --use-mailmap
>>
>> See:
>> https://git-scm.com/docs/git-shortlog#_mapping_authors
>> https://git-scm.com/docs/git-check-mailmap#_mapping_authors
>>
>> Also interesting:
>> https://github.com/sympy/sympy/wiki/Using-.mailmap#making-mailmap-entries
>>
>> Regards,
>>
>> Phil.
>>
>> Philippe Mathieu-Daudé (8):
>>    .mailmap: Convert emails generated by the Subversion original import
>>    .mailmap: Fix emails from the Subversion era
>>    .mailmap: Fix Intel emails rewritten by Microsoft Exchange Server
>>    .mailmap: Fix emails rewritten by lists DMARC / DKIM / SPF
>>    .mailmap: Fix incorrect email formats
>>    .mailmap: Unify Intel email addresses format
>>    .mailmap: Fix UTF-8 mojibaked names
>>    .mailmap: Miscellaneous fixes on various emails
>>
>>   .mailmap | 221 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>   1 file changed, 221 insertions(+)
>>   create mode 100644 .mailmap
>>
> 
> our suggestion from the stewards' confcall is to avoid cross-domain
> mappings, unless a contributor explicitly confirms (in an email sent
> from the mapped-to, i.e. "new" address) that they are OK with the
> mapped-from (i.e. "old") address being "hidden" by the mapping.
> 
> Due to two reasons:
> - mappings across domains could hide employer changes,

Hmm OK.

> - even the "mapped-to" (i.e. "new") address may not be up-to-date.

Agreed.

> Stewards may be able ACK typo fixes and simple name order changes even
> in the absence of the affected (historical) contributor, but
> cross-domain mappings are not something they can ACK on their own.

Understood.

> (At least this is my understanding from the meeting.)
> 
> Personal note in the end: can we perhaps structure this patch set per
> email adress / "liveliness groups"? I think most people will ignore a
> patch that has tens of CC's, even if their personal ACK is needed for
> it. On the other hand, if we have 60-100 patches, but each patch is just
> emailed to 2-3 people (and any given person only gets a very low number
> of personal patch emails), such a series has a better chance at
> succeeding. (We could even commit such a set piece-meal -- commit the
> most recently ACKed subset every week or every two weeks.) Just some
> speculation on my part.

Yes, I'll try to find a clever way.

(I had the v2 ready, but I was waiting to figure out how to restrict to 
100 recipients).