[ANN] pidgin git import v6

[bjlockie at lockie.ca]

> Hi,
>
> Time for a new version. Other than the usual updates, this version now
> has a much faster (10x) script to fix commits, that parses properly
> the information 'mtn git_export' adds, and also converts *all* the mtn
> ids to git ones.
>
> I had to find some logic to properly traverse the history to find out
> the git ids, but it was not so hard after basically implementing a
> custom 'git filter-branch'.
>
> See for example this commit:
> http://gitorious.org/pidgin-clone/pidgin-clone-new/commit/e80524e5a80472ae1b481c5020246af29a270154
>
> The resulting repository with all the changes:
> http://gitorious.org/pidgin-clone/pidgin-clone-new
>
> I don't have any other tasks to do regarding the git conversion, so I
> say it's done.
>
> 1) Uses extensively tested and efficient 'mtn git_export'
>
> 2) Store important monotone information; ids, parents, author,
> branches, multiple changelogs, etc.
>
> 3) Automatically finds out which branches to split
>
> 4) Option to trim the repository (e.g after v2.7.0), and easily graft
> it back again
>
> 5) Convert *all* mtn ids in commit messages to git ids
>
> 6) Scrip to check if something is missing (authors, branches)
>
> 7) Extremely fast; full conversion takes less than an hour in my laptop
>
> 8) Incremental importing
>
> 9) Commit metadata and graph validated against pidgin's
> 'pidgin-mtn-conv-files'
>
> 10) No hacks or patches needed to git, or monotone
>
> 11) Fixes authors and commit messages (including SVN patch authors)
>
> 12) Maps tags and branches
>
> In addition, this can be easily converted to mercurial with a trivial
> hack to 'hg-git'. There would be no loss of information, except
> possibly renames, but this could be easily added.
>
> At this point in time I don't have any motivation to finalize the
> scripts for the proper git -> hg conversion, because I feel I'm doing
> everything myself; doing the mtn -> git -> hg conversion, and fixing
> 'pidgin-mtn-conv-files' mtn -> hg conversion, and possibly even the
> DVCS decision analysis without even a thank you[1].
>
> Cheers.
>
> [1] http://mid.gmane.org/4FC24341.80909@rekkanoryo.org
>
> --
> Felipe Contreras
>
> _______________________________________________
> Devel mailing list
> Devel at pidgin.im
> http://pidgin.im/cgi-bin/mailman/listinfo/devel
>

I haven't been following but thanks for the work Felipe.