view tests/test-notify-changegroup.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

  $ cat <<EOF >> $HGRCPATH
  > [extensions]
  > notify=
  > 
  > [hooks]
  > changegroup.notify = python:hgext.notify.hook
  > 
  > [notify]
  > sources = push
  > diffstat = False
  > maxsubject = 10
  > 
  > [usersubs]
  > foo@bar = *
  > 
  > [reposubs]
  > * = baz
  > EOF
  $ hg init a

clone

  $ hg --traceback clone a b
  updating to branch default
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ echo a > b/a

commit

  $ hg --traceback --cwd b commit -Ama
  adding a
  $ echo a >> b/a

commit

  $ hg --traceback --cwd b commit -Amb

push

  $ hg --traceback --cwd b push ../a 2>&1 |
  >     $PYTHON -c 'from __future__ import print_function ; import sys,re; print(re.sub("\n\t", " ", sys.stdin.read()), end="")'
  pushing to ../a
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 2 changesets with 2 changes to 1 files
  MIME-Version: 1.0
  Content-Type: text/plain; charset="us-ascii"
  Content-Transfer-Encoding: 7bit
  Date: * (glob)
  Subject: * (glob)
  From: test
  X-Hg-Notification: changeset cb9a9f314b8b
  Message-Id: <*> (glob)
  To: baz, foo@bar
  
  changeset cb9a9f314b8b in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=cb9a9f314b8b
  summary: a
  
  changeset ba677d0156c1 in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=ba677d0156c1
  summary: b
  
  diffs (6 lines):
  
  diff -r 000000000000 -r ba677d0156c1 a
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/a	Thu Jan 01 00:00:00 1970 +0000
  @@ -0,0 +1,2 @@
  +a
  +a
  $ hg --cwd a rollback
  repository tip rolled back to revision -1 (undo push)

unbundle with unrelated source

  $ hg --cwd b bundle ../test.hg ../a
  searching for changes
  2 changesets found
  $ hg --cwd a unbundle ../test.hg
  adding changesets
  adding manifests
  adding file changes
  added 2 changesets with 2 changes to 1 files
  new changesets cb9a9f314b8b:ba677d0156c1
  (run 'hg update' to get a working copy)
  $ hg --cwd a rollback
  repository tip rolled back to revision -1 (undo unbundle)

unbundle with correct source

  $ hg --config notify.sources=unbundle --cwd a unbundle ../test.hg 2>&1 |
  >     $PYTHON -c 'from __future__ import print_function ; import sys,re; print(re.sub("\n\t", " ", sys.stdin.read()), end="")'
  adding changesets
  adding manifests
  adding file changes
  added 2 changesets with 2 changes to 1 files
  new changesets cb9a9f314b8b:ba677d0156c1
  MIME-Version: 1.0
  Content-Type: text/plain; charset="us-ascii"
  Content-Transfer-Encoding: 7bit
  Date: * (glob)
  Subject: * (glob)
  From: test
  X-Hg-Notification: changeset cb9a9f314b8b
  Message-Id: <*> (glob)
  To: baz, foo@bar
  
  changeset cb9a9f314b8b in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=cb9a9f314b8b
  summary: a
  
  changeset ba677d0156c1 in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=ba677d0156c1
  summary: b
  
  diffs (6 lines):
  
  diff -r 000000000000 -r ba677d0156c1 a
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/a	Thu Jan 01 00:00:00 1970 +0000
  @@ -0,0 +1,2 @@
  +a
  +a
  (run 'hg update' to get a working copy)

Check that using the first committer as the author of a changeset works:
Check that the config option works.
Check that the first committer is indeed used for "From:".
Check that the merge user is NOT used for "From:"

Create new file

  $ echo a > b/b
  $ echo b >> b/b
  $ echo c >> b/b
  $ hg --traceback --cwd b commit -Amnewfile -u committer_1
  adding b

commit as one user

  $ echo x > b/b
  $ echo b >> b/b
  $ echo c >> b/b
  $ hg --traceback --cwd b commit -Amx -u committer_2

commit as other user, change file so we can do an (automatic) merge

  $ hg --cwd b up 2
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ echo a > b/b
  $ echo b >> b/b
  $ echo y >> b/b
  $ hg --traceback --cwd b commit -Amy -u committer_3
  created new head

merge as a different user

  $ hg --cwd b merge --config notify.fromauthor=True
  merging b
  0 files updated, 1 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ hg --traceback --cwd b commit -Am "merged"

push

  $ hg --traceback --cwd b --config notify.fromauthor=True push ../a 2>&1 |
  >     $PYTHON -c 'from __future__ import print_function ; import sys,re; print(re.sub("\n\t", " ", sys.stdin.read()), end="")'
  pushing to ../a
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 4 changesets with 4 changes to 1 files
  MIME-Version: 1.0
  Content-Type: text/plain; charset="us-ascii"
  Content-Transfer-Encoding: 7bit
  Date: * (glob)
  Subject: * (glob)
  From: committer_1
  X-Hg-Notification: changeset 84e487dddc58
  Message-Id: <*> (glob)
  To: baz, foo@bar
  
  changeset 84e487dddc58 in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=84e487dddc58
  summary: newfile
  
  changeset b29c7a2b6b0c in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=b29c7a2b6b0c
  summary: x
  
  changeset 0957c7d64886 in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=0957c7d64886
  summary: y
  
  changeset 485b4e6b0249 in $TESTTMP/a
  details: $TESTTMP/a?cmd=changeset;node=485b4e6b0249
  summary: merged
  
  diffs (7 lines):
  
  diff -r ba677d0156c1 -r 485b4e6b0249 b
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/b	Thu Jan 01 00:00:00 1970 +0000
  @@ -0,0 +1,3 @@
  +x
  +b
  +y
  $ hg --cwd a rollback
  repository tip rolled back to revision 1 (undo push)