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
#require no-windows
$ . "$TESTDIR/remotefilelog-library.sh"
$ hg init master
$ cd master
$ cat >> .hg/hgrc <<EOF
> [remotefilelog]
> server=True
> EOF
$ echo x > x
$ hg commit -qAm x
$ cd ..
$ hgcloneshallow ssh://user@dummy/master shallow -q
1 files fetched over 1 fetches - (1 misses, 0.00% hit ratio) over *s (glob)
$ cd shallow
$ cat >> $TESTTMP/get_file_linknode.py <<EOF
> from mercurial import node, registrar, scmutil
> cmdtable = {}
> command = registrar.command(cmdtable)
> @command(b'debug-file-linknode', [(b'r', b'rev', b'.', b'rev')], b'hg debug-file-linknode FILE')
> def debug_file_linknode(ui, repo, file, **opts):
> rflctx = scmutil.revsingle(repo.unfiltered(), opts['rev']).filectx(file)
> ui.status(b'%s\n' % node.hex(rflctx.ancestormap()[rflctx._filenode][2]))
> EOF
$ cat >> .hg/hgrc <<EOF
> [ui]
> interactive=1
> [extensions]
> strip=
> get_file_linknode=$TESTTMP/get_file_linknode.py
> [experimental]
> evolution=createmarkers,allowunstable
> EOF
$ echo a > a
$ hg commit -qAm msg1
$ hg commit --amend 're:^$' -m msg2
$ hg commit --amend 're:^$' -m msg3
$ hg --hidden log -G -T '{rev} {node|short}'
@ 3 df91f74b871e
|
| x 2 70494d7ec5ef
|/
| x 1 1e423846dde0
|/
o 0 b292c1e3311f
$ hg debug-file-linknode -r 70494d a
df91f74b871e064c89afa1fe9e2f66afa2c125df
$ hg --hidden strip -r 1 3
0 files updated, 0 files merged, 1 files removed, 0 files unresolved
saved backup bundle to $TESTTMP/shallow/.hg/strip-backup/df91f74b871e-c94d67be-backup.hg
$ hg --hidden log -G -T '{rev} {node|short}'
o 1 70494d7ec5ef
|
@ 0 b292c1e3311f
Demonstrate that the linknode points to a commit that is actually in the repo
after the strip operation. Otherwise remotefilelog has to search every commit in
the repository looking for a valid linkrev every time it's queried, such as
during push.
$ hg debug-file-linknode -r 70494d a
70494d7ec5ef6cd3cd6939a9fd2812f9956bf553