Might be nice to get a list of them and put it in the commit message. (e.g. if you commit the pt_BR translation, list all of the contributors in the commit message, that should be automatable.
On Wed, Mar 20, 2013 at 12:07 PM, Chip Childers <chip.child...@sungard.com> wrote: > On Wed, Mar 20, 2013 at 04:57:58AM -0400, Sebastien Goasguen wrote: >> Hi folks, >> >> I have been committing the translations to the 4.1 branch. >> >> There is an issue with authorship of the translations. >> >> Translators submit translations of strings via a site called transifex. A >> documentation file can therefore have multiple translators. >> The authorship appears in the .po files and looks like this: >> >> # Translators: >> # Daniela Sant Ana de Oliveira <daniela.sant...@ibge.gov.br>, 2013. >> # Edgard Balter Jr. <ebal...@gmail.com>, 2013. >> msgid "" >> msgstr "" >> "Project-Id-Version: Apache CloudStack Docs\n" >> "POT-Creation-Date: 2013-03-19T13:38:05\n" >> "PO-Revision-Date: 2013-03-19 14:13+0000\n" >> "Last-Translator: sebgoa <run...@gmail.com>\n" >> ….<snip> >> >> This example has two translators and I also appear there probably because I >> uploaded the empty strings. >> >> Those .po files are generated automatically and used by publican to build >> the guides in the different languages. >> >> I currently I committed all .po files and while putting the license header I >> mistakenly wiped out all authorship information, which is not good. >> >> However there are over 1000 .po files. If I want to commit with the proper >> authorship I need to automate it. Couple problems arise >> >> 1-The number of commit could be as large as 1000 :( >> 2-A single file can have multiple authors >> >> The easiest way is for me to commit all .po files in one commit and keep the >> authorship within the text of the po files. > > I'm OK with keeping the authorship in the .po files (assuming you were > to re-gen and re-fix license headers to not strip them). > > Generally, the ASF best practice (as I've observed on general@i.a.o and > infra-discuss@i.a.o) is to NOT include authors in the headers. However, > given the fact that transifex takes the time to deal with it, and given > the especially difficult workflow that individual commits would > introduce, I'd say that we're ok with the comment-based approach. > > Unless anyone else objects, go with it.
