Aleksandar Markovic <aleksandar.m.m...@gmail.com> writes:
> 23.08.2019. 08.13, "Markus Armbruster" <arm...@redhat.com> је написао/ла: >> >> Philippe Mathieu-Daudé <phi...@redhat.com> writes: >> >> > Trivial cleanup of .mailmap to have a nice 'git shortlog' output. >> > >> > Philippe Mathieu-Daudé (3): >> > mailmap: Reorder by sections >> > mailmap: Update philmd email address >> > mailmap: Add many entries to improve 'git shortlog' statistics >> > >> > .mailmap | 123 +++++++++++++++++++++++++++++++++++++++++++++++++++---- >> > 1 file changed, 115 insertions(+), 8 deletions(-) >> >> Series >> Reviewed-by: Markus Armbruster <arm...@redhat.com> >> >> However, it increases the difference to contrib/gitdm/aliases. > > Alex' initial gitdm effort, as I understood it, was not meant to cover all > history from 2007 or so, but just to give reasonable statistics for 2018 > (amd future years). > > In that light, .mailmap and gitdm aliases do not need to be equivalent. > > But perhaps Alex would now want gitdm to be used for all QEMU history? Is > this desirable? It would be of interest historically but not something I'd want to spend a lot of time adding code churn for. > > Aleksandar > >> I'm just >> as guilty; my recent "[PATCH 2/2] contrib/gitdm: Add arm...@pond.sub.org >> to group-map-redhat" updates only that. and not .mailmap. >> >> Perhaps we want to keep the two in sync manually. We should then add >> suitable comments to each file. >> >> Could we instead teach gitdm to use .mailmap, and ditch >> contrib/gitdm/aliases? >> >> aliases' format is documented in gitdm's README. Each line maps a >> non-canonical e-mail address to a canonical one. >> >> .mailmap's format is documented in git-shortlog(1). It can do a bit >> more. Even the common part differs: it has two addresses in different >> order *boggle*. >> -- Alex Bennée