discovery: compute newly discovered missing in a more efficient way
Calling "descendants" is expensive, instead, we bound the walk inside the know set
of undecided revision.
This help with discovery performance:
# without the revset '%ld' improvement
$ hg perfdiscovery -R pypy-left pypy-right
before: wall 0.675631 comb 0.680000 user 0.670000 sys 0.010000 (median of 15)
after: wall 0.520145 comb 0.530000 user 0.510000 sys 0.020000 (median of 19)
There is another series in flight that greatly improves performances of "%ld"
substitution in `repo.revs` call. If this changeset is applied above it, we
see a similar performance boost.
# with the revset '%ld' improvement
$ hg perfdiscovery -R pypy-left pypy-right
before: wall 0.477848 comb 0.480000 user 0.480000 sys 0.000000 (median of 22)
after: wall 0.404163 comb 0.400000 user 0.400000 sys 0.000000 (median of 24)
#require no-symlink
# The following script was used to create the bundle:
#
# hg init symlinks
# cd symlinks
# echo a > a
# mkdir d
# echo b > d/b
# ln -s a a.lnk
# ln -s d/b d/b.lnk
# hg ci -Am t
# hg bundle --base null ../test-no-symlinks.hg
Extract a symlink on a platform not supporting them
$ hg init t
$ cd t
$ hg pull -q "$TESTDIR/bundles/test-no-symlinks.hg"
$ hg update
4 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat a.lnk && echo
a
$ cat d/b.lnk && echo
d/b
Copy a symlink and move another
$ hg copy a.lnk d/a2.lnk
$ hg mv d/b.lnk b2.lnk
$ hg ci -Am copy
$ cat d/a2.lnk && echo
a
$ cat b2.lnk && echo
d/b
Bundle and extract again
$ hg bundle --base null ../symlinks.hg
2 changesets found
$ cd ..
$ hg init t2
$ cd t2
$ hg pull ../symlinks.hg
pulling from ../symlinks.hg
requesting all changes
adding changesets
adding manifests
adding file changes
added 2 changesets with 6 changes to 6 files
new changesets d326ae2d01ee:71d85cf3ba90 (2 drafts)
(run 'hg update' to get a working copy)
$ hg update
5 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ cat a.lnk && echo
a
$ cat d/a2.lnk && echo
a
$ cat b2.lnk && echo
d/b