> On 8 Apr 2022, at 00:07, Matt Turner <matts...@gentoo.org> wrote:
> 
> On Thu, Apr 7, 2022 at 11:42 AM Michał Górny <mgo...@gentoo.org> wrote:
>> 
>> Hello,
>> 
>> Right now we're keeping both email addresses (obligatory) and names
>> (optional) for downstream maintainers in metadata.xml. The way I see
>> it, there are three problems with that:
>> 
>> 1. As noticed on IRC lately, a few devs haven't been listing their names
>> at all, resulting in these names being missing from packages.g.o.
>> 
>> 2. Not all names are listed consistently. This is especially the case
>> for projects. When you want to group everything by maintainer, which
>> name should be used?
>> 
>> 3. In the end, listing the same names all over the place is a lot of
>> redundancy.
>> 
>> 
>> I'd like to propose that we deprecate <name/> for downstream
>> maintainers, and instead work towards using an additional mapping from
>> maintainer email addresses to their names.
>> 
>> a. For projects, we can simply use projects.xml. We already require
>> that all type="project" maintainers correspond to entries
>> in projects.xml, so we should be good here.
>> 
>> b. For human maintainers, I think we can use metadata/AUTHORS. This is
>> pretty much killing two birds with one stone -- we could finally getting
>> the file more complete, and at the same time use it to provide names for
>> maintainers.
>> 
>> While keeping names in metadata.xml has the advantage that they are
>> immediately available (provided that they are actually listed there),
>> I don't think this is really a show-stopper.
> 
> Sounds like a good plan to me.

Yep. It also has a nice consequence of allowing AUTHORS to be used as a mailmap
for git (although git doesn't respect symlinks for mailmap, so we'd need to tell
people to set it with the config option, but still.)

The main value for me is in making AUTHORS more useful. If it has to exist,
we should use it properly.

But I love a bit of deduplication too.

Best,
sam

Attachment: signature.asc
Description: Message signed with OpenPGP

Reply via email to