bisect: avoid copying ancestor list for non-merge commits
During a bisection, hg needs to compute a list of all ancestors for every
candidate commit. This is accomplished via a bottom-up traversal of the set of
candidates, during which each revision's ancestor list is populated using the
ancestor list of its parent(s). Previously, this involved copying the entire
list, which could be very long in if the bisection range was large.
To help improve this, we can observe that each candidate commit is visited
exactly once, at which point its ancestor list is copied into its children's
lists and then dropped. In the case of non-merge commits, a commit's ancestor
list consists exactly of its parent's list plus itself. This means that we can
trivially reuse the parent's existing list for one of its non-merge children,
which avoids copying entirely if that commit is the parent's only child. This
makes bisections over linear ranges of commits much faster.
During some informal testing in the large publicly-available `mozilla-central`
repository, this noticeably sped up bisections over large ranges of history:
Setup:
$ cd mozilla-central
$ hg bisect --reset
$ hg bisect --good 0
$ hg log -r tip -T '{rev}\n'
628417
Test:
$ time hg bisect --bad tip --noupdate
Before:
real 3m35.927s
user 3m35.553s
sys 0m0.319s
After:
real 1m41.142s
user 1m40.810s
sys 0m0.285s
Octopus Merge Support
=====================
This will be moderately complicated, as we'll need to synthesize phony
changeset entries to explode the octopus into "revisions" that only
have two parents each. For today, we can probably just do something like
aaaaaaaaaaaaaaaaaaXX{20 bytes of exploded node's hex sha}
where XX is a counter (so we could have as many as 255 parents in a
git commit - more than I think we'd ever see.) That means that we can
install some check in this extension to disallow checking out or
otherwise interacting with the `aaaaaaaaaaaaaaaaaa` revisions.
Interface Creation
====================
We at least need an interface definition for `changelog` in core that
this extension can satisfy, and again for `basicstore`.
Reason About Locking
====================
We should spend some time thinking hard about locking, especially on
.git/index etc. We're probably adequately locking the _git_
repository, but may not have enough locking correctness in places
where hg does locking that git isn't aware of (notably the working
copy, which I believe Git does not lock.)