Mercurial > hg
view tests/test-pathconflicts-merge.t @ 38732:be4984261611
merge: mark file gets as not thread safe (issue5933)
In default installs, this has the effect of disabling the thread-based
worker on Windows when manifesting files in the working directory. My
measurements have shown that with revlog-based repositories, Mercurial
spends a lot of CPU time in revlog code resolving file data. This ends
up incurring a lot of context switching across threads and slows down
`hg update` operations when going from an empty working directory to
the tip of the repo.
On mozilla-unified (246,351 files) on an i7-6700K (4+4 CPUs):
before: 487s wall
after: 360s wall (equivalent to worker.enabled=false)
cpus=2: 379s wall
Even with only 2 threads, the thread pool is still slower.
The introduction of the thread-based worker (02b36e860e0b) states that
it resulted in a "~50%" speedup for `hg sparse --enable-profile` and
`hg sparse --disable-profile`. This disagrees with my measurement
above. I theorize a few reasons for this:
1) Removal of files from the working directory is I/O - not CPU - bound
and should benefit from a thread pool (unless I/O is insanely fast
and the GIL release is near instantaneous). So tests like `hg sparse
--enable-profile` may exercise deletion throughput and aren't good
benchmarks for worker tasks that are CPU heavy.
2) The patch was authored by someone at Facebook. The results were
likely measured against a repository using remotefilelog. And I
believe that revision retrieval during working directory updates with
remotefilelog will often use a remote store, thus being I/O and not
CPU bound. This probably resulted in an overstated performance gain.
Since there appears to be a need to enable the thread-based worker with
some stores, I've made the flagging of file gets as thread safe
configurable. I've made it experimental because I don't want to formalize
a boolean flag for this option and because this attribute is best
captured against the store implementation. But we don't have a proper
store API for this yet. I'd rather cross this bridge later.
It is possible there are revlog-based repositories that do benefit from
a thread-based worker. I didn't do very comprehensive testing. If there
are, we may want to devise a more proper algorithm for whether to use
the thread-based worker, including possibly config options to limit the
number of threads to use. But until I see evidence that justifies
complexity, simplicity wins.
Differential Revision: https://phab.mercurial-scm.org/D3963
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Wed, 18 Jul 2018 09:49:34 -0700 |
parents | a8a0cafcef79 |
children | 382f4f09f0bd |
line wrap: on
line source
Path conflict checking is currently disabled by default because of issue5716. Turn it on for this test. $ cat >> $HGRCPATH << EOF > [experimental] > merge.checkpathconflicts=True > EOF $ hg init repo $ cd repo $ echo base > base $ hg add base $ hg commit -m "base" $ hg bookmark -i base $ mkdir a $ echo 1 > a/b $ hg add a/b $ hg commit -m "file" $ hg bookmark -i file $ echo 2 > a/b $ hg commit -m "file2" $ hg bookmark -i file2 $ hg up 0 0 files updated, 0 files merged, 1 files removed, 0 files unresolved #if symlink $ mkdir a $ ln -s c a/b $ hg add a/b $ hg commit -m "link" created new head #else $ hg import -q --bypass - <<EOF > # HG changeset patch > link > > diff --git a/a/b b/a/b > new file mode 120000 > --- /dev/null > +++ b/a/b > @@ -0,0 +1,1 @@ > +c > \ No newline at end of file > EOF $ hg up -q #endif $ hg bookmark -i link $ hg up 0 0 files updated, 0 files merged, 1 files removed, 0 files unresolved $ mkdir -p a/b/c $ echo 2 > a/b/c/d $ hg add a/b/c/d $ hg commit -m "dir" created new head $ hg bookmark -i dir Merge - local file conflicts with remote directory $ hg up file 1 files updated, 0 files merged, 1 files removed, 0 files unresolved (activating bookmark file) $ hg bookmark -i $ hg merge --verbose dir resolving manifests a/b: path conflict - a file or link has the same name as a directory the local file has been renamed to a/b~0ed027b96f31 resolve manually then use 'hg resolve --mark a/b' moving a/b to a/b~0ed027b96f31 getting a/b/c/d 1 files updated, 0 files merged, 0 files removed, 1 files unresolved use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon [1] $ hg status M a/b/c/d A a/b~0ed027b96f31 R a/b $ hg resolve --all a/b: path conflict must be resolved manually $ hg forget a/b~0ed027b96f31 && rm a/b~0ed027b96f31 $ hg resolve --mark a/b (no more unresolved files) $ hg commit -m "merge file and dir (deleted file)" Merge - local symlink conflicts with remote directory $ hg up link 1 files updated, 0 files merged, 1 files removed, 0 files unresolved (activating bookmark link) $ hg bookmark -i $ hg merge dir a/b: path conflict - a file or link has the same name as a directory the local file has been renamed to a/b~2ea68033e3be resolve manually then use 'hg resolve --mark a/b' 1 files updated, 0 files merged, 0 files removed, 1 files unresolved use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon [1] $ hg status M a/b/c/d A a/b~2ea68033e3be R a/b $ hg resolve --list P a/b $ hg resolve --all a/b: path conflict must be resolved manually $ hg mv a/b~2ea68033e3be a/b.old $ hg resolve --mark a/b (no more unresolved files) $ hg resolve --list R a/b $ hg commit -m "merge link and dir (renamed link)" Merge - local directory conflicts with remote file or link $ hg up dir 0 files updated, 0 files merged, 1 files removed, 0 files unresolved (activating bookmark dir) $ hg bookmark -i $ hg merge file a/b: path conflict - a file or link has the same name as a directory the remote file has been renamed to a/b~0ed027b96f31 resolve manually then use 'hg resolve --mark a/b' 1 files updated, 0 files merged, 0 files removed, 1 files unresolved use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon [1] $ hg status A a/b~0ed027b96f31 $ hg resolve --all a/b: path conflict must be resolved manually $ hg mv a/b~0ed027b96f31 a/b/old-b $ hg resolve --mark a/b (no more unresolved files) $ hg commit -m "merge dir and file (move file into dir)" created new head $ hg merge file2 merging a/b/old-b and a/b to a/b/old-b 0 files updated, 1 files merged, 0 files removed, 0 files unresolved (branch merge, don't forget to commit) $ cat a/b/old-b 2 $ hg commit -m "merge file2 (copytrace tracked rename)" $ hg merge link a/b: path conflict - a file or link has the same name as a directory the remote file has been renamed to a/b~2ea68033e3be resolve manually then use 'hg resolve --mark a/b' 1 files updated, 0 files merged, 0 files removed, 1 files unresolved use 'hg resolve' to retry unresolved file merges or 'hg merge --abort' to abandon [1] $ hg mv a/b~2ea68033e3be a/b.old #if symlink $ readlink.py a/b.old a/b.old -> c #else $ cat a/b.old c (no-eol) #endif $ hg resolve --mark a/b (no more unresolved files) $ hg commit -m "merge link (rename link)"