>> These days synchronizing with some public issue tracker (Git{Lab,Hub}
>> issue tracker or something of that sort) is probably essential.
> Which is what sd did and died, IIRC partially because of its
> complexity and huge dependency chain for all the different backends.
I wrote a conversion script to turn a Debbugs database into
a BuGit database. The purpose is just to generate a realistic BuGit
database (by converting debbugs.gnu.org) so as to see how it scales
(I'm not aiming to handle the million-or-so bugs of bugs.debian.org,
but if BuGit can't handle the 20K bugs of debbugs.gnu.org then
I consider it to be a problem).
Along the way I discovered that such conversions can be terribly
difficult if you want to do them well: my intention (tho currently the
code doesn't depend on it nor make use of it) is/was for BuGit's
comments to be written in some kind of "markdown", but that means that
Debbugs messages don't just need to be "de-MIME'd" but also converted
to markdown (and I probably don't want to allow embedded HTML in the
markdown since the messages should be readable as plain text).
For my current use case (generation of a realistic test database),
I just punted on those issues, but I can definitely see how you'd
quickly get "complexity and huge dependency chain for all the different
backends".
Stefan
_______________________________________________
dist-bugs mailing list
[email protected]
http://kitenet.net/cgi-bin/mailman/listinfo/dist-bugs
