view tests/test-committer.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

  $ unset HGUSER
  $ EMAIL="My Name <[e-mail]>"
  $ export EMAIL

  $ hg init test
  $ cd test
  $ touch asdf
  $ hg add asdf
  $ hg commit -m commit-1
  $ hg tip
  changeset:   0:53f268a58230
  tag:         tip
  user:        My Name <[e-mail]>
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     commit-1
  

  $ unset EMAIL
  $ echo 1234 > asdf
  $ hg commit -u "[e-mail]" -m commit-1
  $ hg tip
  changeset:   1:3871b2a9e9bf
  tag:         tip
  user:        [e-mail]
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     commit-1
  
  $ echo "[ui]" >> .hg/hgrc
  $ echo "username = foobar <[e-mail]>" >> .hg/hgrc
  $ echo 12 > asdf
  $ hg commit -m commit-1
  $ hg tip
  changeset:   2:8eeac6695c1c
  tag:         tip
  user:        foobar <[e-mail]>
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     commit-1
  
  $ echo 1 > asdf
  $ hg commit -u "[e-mail]" -m commit-1
  $ hg tip
  changeset:   3:957606a725e4
  tag:         tip
  user:        [e-mail]
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     commit-1
  
  $ echo 123 > asdf
  $ echo "[ui]" > .hg/hgrc
  $ echo "username = " >> .hg/hgrc
  $ hg commit -m commit-1
  abort: no username supplied
  (use 'hg config --edit' to set your username)
  [255]

# test alternate config var

  $ echo 1234 > asdf
  $ echo "[ui]" > .hg/hgrc
  $ echo "user = Foo Bar II <[e-mail]>" >> .hg/hgrc
  $ hg commit -m commit-1
  $ hg tip
  changeset:   4:6f24bfb4c617
  tag:         tip
  user:        Foo Bar II <[e-mail]>
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     commit-1
  
# test prompt username

  $ cat > .hg/hgrc <<EOF
  > [ui]
  > askusername = True
  > EOF

  $ echo 12345 > asdf
  $ hg commit --config ui.interactive=False -m ask
  enter a commit username: 
  no username found, using '[^']*' instead (re)
  $ hg rollback -q

  $ hg commit --config ui.interactive=True -m ask <<EOF
  > Asked User <[e-mail]>
  > EOF
  enter a commit username: Asked User <[e-mail]>
  $ hg tip
  changeset:   5:84c91d963b70
  tag:         tip
  user:        Asked User <[e-mail]>
  date:        Thu Jan 01 00:00:00 1970 +0000
  summary:     ask
  

# test no .hg/hgrc (uses generated non-interactive username)

  $ echo space > asdf
  $ rm .hg/hgrc
  $ hg commit -m commit-1 2>&1
  no username found, using '[^']*' instead (re)

  $ echo space2 > asdf
  $ hg commit -u ' ' -m commit-1
  transaction abort!
  rollback completed
  abort: empty username!
  [255]

# don't add tests here, previous test is unstable

  $ cd ..