obsolete: order of magnitude speedup in _computebumpedset
Reminder: a changeset is said "bumped" if it tries to obsolete a immutable
changeset.
The previous algorithm for computing bumped changeset was:
1) Get all public changesets
2) Find all they successors
3) Search for stuff that are eligible for being "bumped"
(mutable and non obsolete)
The entry size of this algorithm is `O(len(public))` which is mostly the same as
`O(len(repo))`. Even this this approach mean fewer obsolescence marker are
traveled, this is not very scalable.
The new algorithm is:
1) For each potential bumped changesets (non obsolete mutable)
2) iterate over precursors
3) if a precursors is public. changeset is bumped
We travel more obsolescence marker, but the entry size is much smaller since
the amount of potential bumped should remains mostly stable with time `O(1)`.
On some confidential gigantic repo this move bumped computation from 15.19s to
0.46s (×33 speedup…). On "smaller" repo (mercurial, cubicweb's review) no
significant gain were seen. The additional traversal of obsolescence marker is
probably probably counter balance the advantage of it.
Other optimisation could be done in the future (eg: sharing precursors cache
for divergence detection)
Test alignment of multibyte characters
$ HGENCODING=utf-8
$ export HGENCODING
$ hg init t
$ cd t
$ python << EOF
> # (byte, width) = (6, 4)
> s = "\xe7\x9f\xad\xe5\x90\x8d"
> # (byte, width) = (7, 7): odd width is good for alignment test
> m = "MIDDLE_"
> # (byte, width) = (18, 12)
> l = "\xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d"
> f = file('s', 'w'); f.write(s); f.close()
> f = file('m', 'w'); f.write(m); f.close()
> f = file('l', 'w'); f.write(l); f.close()
> # instant extension to show list of options
> f = file('showoptlist.py', 'w'); f.write("""# encoding: utf-8
> def showoptlist(ui, repo, *pats, **opts):
> '''dummy command to show option descriptions'''
> return 0
> cmdtable = {
> 'showoptlist':
> (showoptlist,
> [('s', 'opt1', '', 'short width' + ' %(s)s' * 8, '%(s)s'),
> ('m', 'opt2', '', 'middle width' + ' %(m)s' * 8, '%(m)s'),
> ('l', 'opt3', '', 'long width' + ' %(l)s' * 8, '%(l)s')
> ],
> ""
> )
> }
> """ % globals())
> f.close()
> EOF
$ S=`cat s`
$ M=`cat m`
$ L=`cat l`
alignment of option descriptions in help
$ cat <<EOF > .hg/hgrc
> [extensions]
> ja_ext = `pwd`/showoptlist.py
> EOF
check alignment of option descriptions in help
$ hg help showoptlist
hg showoptlist
dummy command to show option descriptions
options:
-s --opt1 \xe7\x9f\xad\xe5\x90\x8d short width \xe7\x9f\xad\xe5\x90\x8d \xe7\x9f\xad\xe5\x90\x8d \xe7\x9f\xad\xe5\x90\x8d \xe7\x9f\xad\xe5\x90\x8d \xe7\x9f\xad\xe5\x90\x8d \xe7\x9f\xad\xe5\x90\x8d \xe7\x9f\xad\xe5\x90\x8d \xe7\x9f\xad\xe5\x90\x8d (esc)
-m --opt2 MIDDLE_ middle width MIDDLE_ MIDDLE_ MIDDLE_ MIDDLE_ MIDDLE_
MIDDLE_ MIDDLE_ MIDDLE_
-l --opt3 \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d long width \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d (esc)
\xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d (esc)
\xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d (esc)
use "hg -v help showoptlist" to show the global options
$ rm -f s; touch s
$ rm -f m; touch m
$ rm -f l; touch l
add files
$ cp s $S
$ hg add $S
$ cp m $M
$ hg add $M
$ cp l $L
$ hg add $L
commit(1)
$ echo 'first line(1)' >> s; cp s $S
$ echo 'first line(2)' >> m; cp m $M
$ echo 'first line(3)' >> l; cp l $L
$ hg commit -m 'first commit' -u $S
commit(2)
$ echo 'second line(1)' >> s; cp s $S
$ echo 'second line(2)' >> m; cp m $M
$ echo 'second line(3)' >> l; cp l $L
$ hg commit -m 'second commit' -u $M
commit(3)
$ echo 'third line(1)' >> s; cp s $S
$ echo 'third line(2)' >> m; cp m $M
$ echo 'third line(3)' >> l; cp l $L
$ hg commit -m 'third commit' -u $L
check alignment of user names in annotate
$ hg annotate -u $M
\xe7\x9f\xad\xe5\x90\x8d: first line(2) (esc)
MIDDLE_: second line(2)
\xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d: third line(2) (esc)
check alignment of filenames in diffstat
$ hg diff -c tip --stat
MIDDLE_ | 1 +
\xe7\x9f\xad\xe5\x90\x8d | 1 + (esc)
\xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d | 1 + (esc)
3 files changed, 3 insertions(+), 0 deletions(-)
add branches/tags
$ hg branch $S
marked working directory as branch \xe7\x9f\xad\xe5\x90\x8d (esc)
(branches are permanent and global, did you want a bookmark?)
$ hg tag $S
$ hg branch $M
marked working directory as branch MIDDLE_
(branches are permanent and global, did you want a bookmark?)
$ hg tag $M
$ hg branch $L
marked working directory as branch \xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d (esc)
(branches are permanent and global, did you want a bookmark?)
$ hg tag $L
check alignment of branches
$ hg tags
tip 5:d745ff46155b
\xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d 4:9259be597f19 (esc)
MIDDLE_ 3:b06c5b6def9e
\xe7\x9f\xad\xe5\x90\x8d 2:64a70663cee8 (esc)
check alignment of tags
$ hg tags
tip 5:d745ff46155b
\xe9\x95\xb7\xe3\x81\x84\xe9\x95\xb7\xe3\x81\x84\xe5\x90\x8d\xe5\x89\x8d 4:9259be597f19 (esc)
MIDDLE_ 3:b06c5b6def9e
\xe7\x9f\xad\xe5\x90\x8d 2:64a70663cee8 (esc)
$ cd ..