Can "bzr annotate" be sped up?

[Stefan Monnier]

>> So maybe a workaround for this lot-of-revisions case can be giving
>> annotate/qannotate a line number range to make it give answers faster
>> for the lines Eli cares about.

> Yes, that would be very good, as normally I need to annotate only a
> very small portion of a file.

Yes, yes, yes.  At the risk of repeating myself, here's how I see it
working:

   "bzr log --show-diff --line-range NN..MM FILE"

After all, I don't care about the "annotate output".  All I want is to
find the sequence of diff-hunks (and corresponding commit-logs) that
resulted in lines NN..MM of FILE.

Currently, to get that info I end up having to run "bzr annotate"+"bzr
log --diff" X times (where X is the number of hunks that affected the
lines I care about between the introduction of the problem I'm
investigating and the branch tip), which is *dog slow*.


        Stefan