Florian Weimer <f...@deneb.enyo.de>:
> Thanks for your offer for working on this. (I actually brought up
> your name at the Cauldron. :)
>
> Can your tools handle changes in preferred email addresses over time?
It would in theory be possible to do this with multiple applications
of the reposurgeon "authors read" command using different map files
and applied to date-bounded spans of revisions. So, yes, my tools
could handle this case just fine.
In practice, nobody has ever wanted to actually try this. The
complexity overhead of deciding when an address should change would be
high, and doing so would destroy a useful property these user IDs
normally have, which is to imply a 1-1 map between IDs and committers
(for searchability).
In fact I have sometimes been asked to correct older email addresses to
newer preferred ones in order to make a 1-1 mapping where it didn't exist
before.
The best way to fill in that map probably starts with grepping all the
>From lines out of your mail archives. Once you sort | uniq those and
weed out older addresses you want to ignore, a script to fill in the map using
that data should not be difficult to write.
--
<a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>
