view tests/test-mq-qrename.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

  $ echo "[extensions]" >> $HGRCPATH
  $ echo "mq=" >> $HGRCPATH

  $ hg init a
  $ cd a

  $ echo 'base' > base
  $ hg ci -Ambase
  adding base

  $ hg qnew -mmqbase mqbase

  $ hg qrename mqbase renamed
  $ mkdir .hg/patches/foo
  $ hg qrename renamed foo

  $ hg qseries
  foo/renamed

  $ ls .hg/patches/foo
  renamed

  $ mkdir .hg/patches/bar
  $ hg qrename foo/renamed bar

  $ hg qseries
  bar/renamed

  $ ls .hg/patches/bar
  renamed

  $ hg qrename bar/renamed baz

  $ hg qseries
  baz

  $ ls .hg/patches/baz
  .hg/patches/baz

  $ hg qrename baz new/dir

  $ hg qseries
  new/dir

  $ ls .hg/patches/new/dir
  .hg/patches/new/dir

  $ cd ..

Test patch being renamed before committed:

  $ hg init b
  $ cd b
  $ hg qinit -c
  $ hg qnew x
  $ hg qrename y
  $ hg qcommit -m rename

  $ cd ..

Test overlapping renames (issue2388)

  $ hg init c
  $ cd c
  $ hg qinit -c
  $ echo a > a
  $ hg add
  adding a
  $ hg qnew patcha
  $ echo b > b
  $ hg add
  adding b
  $ hg qnew patchb
  $ hg ci --mq -m c1
  $ hg qrename patchb patchc
  $ hg qrename patcha patchb
  $ hg st --mq
  M patchb
  M series
  A patchc
  R patcha
  $ cd ..

Test renames with mq repo (issue2097)

  $ hg init issue2097
  $ cd issue2097
  $ hg qnew p0
  $ (cd .hg/patches && hg init)
  $ hg qren p0 p1
  $ hg debugstate --mq
  $ hg ci --mq -mq0
  nothing changed
  [1]
  $ cd ..

Test renaming to a folded patch (issue3058)

  $ hg init issue3058
  $ cd issue3058
  $ hg init --mq
  $ echo a > a
  $ hg add a
  $ hg qnew adda
  $ echo b >> a
  $ hg qnew addb
  $ hg qpop
  popping addb
  now at: adda
  $ hg ci --mq -m "save mq"
  $ hg qfold addb
  $ hg qmv addb
  $ cat .hg/patches/addb
  # HG changeset patch
  # Parent  0000000000000000000000000000000000000000
  
  diff -r 000000000000 a
  --- /dev/null	* (glob)
  +++ b/a	* (glob)
  @@ -0,0 +1,2 @@
  +a
  +b
  $ cd ..