view tests/test-mq-merge.t @ 47010:76ae43d5b1db stable

repoview: fix memory leak of filtered repo classes

The leak occurs in long-running server processes with
extensions, and is measured at 110kB per request.

Before this change, the contents of the `_filteredrepotypes`
cache are not properly garbage collected, despite it begin
a `WeakKeyDictionary`.

Extensions have a tendency to generate a new repository class
for each `localrepo` instantiation. Server processes based
on `hgwebdir_mod` will instantiate a new `localrepo` for each
HTTP request that involves a repository.

As a result, with a testing process that repeatedly opens a
repository with several extensions activated
(`topic` notably among them), we see a steady increase in
resident memory of 110kB per repository instantiation before this
change. This is also true, if we call `gc.collect()` at each
instantiation, like `hgwebdir_mod` does, or not.

The cause of the leak is that the *values* aren't weak references.
This change uses `weakref.ref` for the values, and this makes
in our measurements the resident size increase drop to 5kB per
repository instantiation, with no explicit call of `gc.collect()`
at all.
There is currently no reason to believe that this remaining leak
of 5kB is related to or even due to Mercurial core.

We've also seen evidence that `ui.ui` instances weren't properly
garbage collected before the change (with the change, they are).
This could explain why the figures are relatively high.

In theory, the collection of weak references could lead to
much more misses in the cache, so we measured the impact on
the original case that was motivation for introducing that cache
in 7e89bd0cfb86 (see also issue5043): `hg convert` of the
mozilla-central repository. The bad news here is that there is a
major memory leak there, both with and without the present changeset.
There were no more cache misses, and we could see no
more memory leak with this change: the resident size after importing
roughly 100000 changesets was at 12.4GB before, and 12.5GB after.
The small increase is mentioned for completeness only, and we
believe that it should be ignored, at least as long as the main
leak isn't fixed. At less than 1% of the main leak, even finding out
whether it is merely noise would be wasteful.

Original context where this was spotted and first mitigated:
  https://foss.heptapod.net/heptapod/heptapod/-/issues/466

The leak reduction was also obtained in Heptapod inner HTTP server,
which amounts to the same as `hgwebdir_mod` for these questions.

The measurements done with Python 3.9, similar figures seen with 3.8.
More work on our side would be needed to give measurements with 2.7,
because of testing server process does not support it.

Setup extension:

  $ cat <<EOF >> $HGRCPATH
  > [extensions]
  > mq =
  > [mq]
  > git = keep
  > EOF

Test merge with mq changeset as the second parent:

  $ hg init m
  $ cd m
  $ touch a b c
  $ hg add a
  $ hg commit -m a
  $ hg add b
  $ hg qnew -d "0 0" b
  $ hg update 0
  0 files updated, 0 files merged, 1 files removed, 0 files unresolved
  $ hg add c
  $ hg commit -m c
  created new head
  $ hg merge
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ hg commit -m merge
  abort: cannot commit over an applied mq patch
  [255]
  $ cd ..

Issue529: mq aborts when merging patch deleting files

  $ checkundo()
  > {
  >     if [ -f .hg/store/undo ]; then
  >         echo ".hg/store/undo still exists"
  >     fi
  > }

Commit two dummy files in "init" changeset:

  $ hg init t
  $ cd t
  $ echo a > a
  $ echo b > b
  $ hg ci -Am init
  adding a
  adding b
  $ hg tag -l init

Create a patch removing a:

  $ hg qnew rm_a
  $ hg rm a
  $ hg qrefresh -m "rm a"

Save the patch queue so we can merge it later:

  $ hg qsave -c -e
  copy $TESTTMP/t/.hg/patches to $TESTTMP/t/.hg/patches.1
  $ checkundo

Update b and commit in an "update" changeset:

  $ hg up -C init
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ echo b >> b
  $ hg st
  M b
  $ hg ci -m update
  created new head

# Here, qpush used to abort with :
# The system cannot find the file specified => a
  $ hg manifest
  a
  b

  $ hg qpush -a -m
  merging with queue at: $TESTTMP/t/.hg/patches.1
  applying rm_a
  now at: rm_a

  $ checkundo
  $ hg manifest
  b

Ensure status is correct after merge:

  $ hg qpop -a
  popping rm_a
  popping .hg.patches.merge.marker
  patch queue now empty

  $ cd ..

Classic MQ merge sequence *with an explicit named queue*:

  $ hg init t2
  $ cd t2
  $ echo '[diff]' > .hg/hgrc
  $ echo 'nodates = 1' >> .hg/hgrc
  $ echo a > a
  $ hg ci -Am init
  adding a
  $ echo b > a
  $ hg ci -m changea
  $ hg up -C 0
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg cp a aa
  $ echo c >> a
  $ hg qnew --git -f -e patcha
  $ echo d >> a
  $ hg qnew -d '0 0' -f -e patcha2

Create the reference queue:

  $ hg qsave -c -e -n refqueue
  copy $TESTTMP/t2/.hg/patches to $TESTTMP/t2/.hg/refqueue
  $ hg up -C 1
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved

Merge:

  $ HGMERGE=internal:other hg qpush -a -m -n refqueue
  merging with queue at: $TESTTMP/t2/.hg/refqueue
  applying patcha
  patching file a
  Hunk #1 succeeded at 2 with fuzz 1 (offset 0 lines).
  fuzz found when applying patch, stopping
  patch didn't work out, merging patcha
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  0 files updated, 2 files merged, 0 files removed, 0 files unresolved
  applying patcha2
  now at: patcha2

Check patcha is still a git patch:

  $ cat .hg/patches/patcha
  # HG changeset patch
  # Parent  d3873e73d99ef67873dac33fbcc66268d5d2b6f4
  
  diff --git a/a b/a
  --- a/a
  +++ b/a
  @@ -1,1 +1,2 @@
  -b
  +a
  +c
  diff --git a/a b/aa
  copy from a
  copy to aa
  --- a/a
  +++ b/aa
  @@ -1,1 +1,1 @@
  -b
  +a

Check patcha2 is still a regular patch:

  $ cat .hg/patches/patcha2
  # HG changeset patch
  # Date 0 0
  # Parent  ???????????????????????????????????????? (glob)
  
  diff -r ???????????? -r ???????????? a (glob)
  --- a/a
  +++ b/a
  @@ -1,2 +1,3 @@
   a
   c
  +d

  $ cd ..