Git however.. Does not do $Id$ replacements, so we are back to manual
approach - like described in:
http://php.markmail.org/message/3t2pmgoq35h5ogep :)
I really like your idea. I can see only one problem connected with it. When revision ID from comment is SVN revision number at the same time it is pretty easy to check the diff (just
http://svn.php.net/viewvc/phpdoc/[file]?r1=[EN-Rev from translation]&r2=[newest rev]) and I know at least one tool using this. Probably vast majority of translators using it because is the logical way to check the differences between two revisions.
If we were to set revision="" manually, we lose this feature.
And the other problem I had noticed this night. As I wrote:
one cron script which is currently checking out SVN files.
All further operations are done with copy of these files.
To be precise: https://github.com/php/web-doc/blob/master/scripts/populatedocs.sh - this one. Is it not true that when we're cloning git repo, we're actually downloading full repository with all versions of all files? It is great feature because it allows DVCS to work but if we would be forced to download all these files the size would be enormous. I hope there is a way to get only newest versions of files.