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
=========================================
Testing single head enforcement: Case A-5
=========================================
A repository is set to only accept a single head per name (typically named
branch). However, obsolete changesets can make this enforcement more
complicated, because they can be kept visible by other changeset on other
branch.
This case is part of a series of tests checking this behavior.
Category A: Involving obsolescence
TestCase 5: Obsoleting a merge reveals two heads
.. old-state:
..
.. * 3 changesets on branch default (2 on their own branch + 1 merge)
.. * 1 changeset on branch Z (children of the merge)
..
.. new-state:
..
.. * 2 changesets on branch default (merge is obsolete) each a head
.. * 1 changeset on branch Z keeping the merge visible
..
.. expected-result:
..
.. * 2 heads detected (because we skip the merge)
..
.. graph-summary:
..
.. C ● (branch Z)
.. |
.. M ⊗
.. |\
.. A ● ● B
.. |/
.. ●
$ . $TESTDIR/testlib/push-checkheads-util.sh
$ cat >> $HGRCPATH << EOF
> [command-templates]
> log = "{node|short} [{branch}] ({phase}): {desc}\n"
> EOF
Test setup
----------
$ mkdir A5
$ cd A5
$ setuprepos single-head
creating basic server and client repo
updating to branch default
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cd client
$ hg up 0
0 files updated, 0 files merged, 1 files removed, 0 files unresolved
$ mkcommit B0
created new head
$ hg merge
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
(branch merge, don't forget to commit)
$ hg ci -m 'M0'
$ hg branch Z
marked working directory as branch Z
(branches are permanent and global, did you want a bookmark?)
$ mkcommit C0
$ hg push --new-branch
pushing to $TESTTMP/A5/server
searching for changes
adding changesets
adding manifests
adding file changes
added 3 changesets with 2 changes to 2 files
$ hg debugobsolete `getid "desc(M0)"` --record-parents
1 new obsolescence markers
obsoleted 1 changesets
1 new orphan changesets
$ hg heads
61c95483cc12 [Z] (draft): C0
74ff5441d343 [default] (draft): B0
8aaa48160adc [default] (draft): A0
$ hg log -G --hidden
@ 61c95483cc12 [Z] (draft): C0
|
x 14d3d4d41d1a [default] (draft): M0
|\
| o 74ff5441d343 [default] (draft): B0
| |
o | 8aaa48160adc [default] (draft): A0
|/
o 1e4be0697311 [default] (public): root
Actual testing
--------------
(force push to make sure we get the changeset on the remote)
$ hg push -r 'desc("C0")' --force
pushing to $TESTTMP/A5/server
searching for changes
no changes found
transaction abort!
rollback completed
abort: rejecting multiple heads on branch "default"
(2 heads: 8aaa48160adc 74ff5441d343)
[255]