tests/test-hgignore.t

revlog: use an LRU cache for delta chain bases

Profiling using statprof revealed a hotspot during changegroup
application calculating delta chain bases on generaldelta repos.
Essentially, revlog._addrevision() was performing a lot of redundant
work tracing the delta chain as part of determining when the chain
distance was acceptable. This was most pronounced when adding
revisions to manifests, which can have delta chains thousands of
revisions long.

There was a delta chain base cache on revlogs before, but it only
captured a single revision. This was acceptable before generaldelta,
when _addrevision would build deltas from the previous revision and
thus we'd pretty much guarantee a cache hit when resolving the delta
chain base on a subsequent _addrevision call. However, it isn't
suitable for generaldelta because parent revisions aren't necessarily
the last processed revision.

This patch converts the delta chain base cache to an LRU dict cache.
The cache can hold multiple entries, so generaldelta repos have a
higher chance of getting a cache hit.

The impact of this change when processing changegroup additions is
significant. On a generaldelta conversion of the "mozilla-unified"
repo (which contains heads of the main Firefox repositories in
chronological order - this means there are lots of transitions between
heads in revlog order), this change has the following impact when
performing an `hg unbundle` of an uncompressed bundle of the repo:

before: 5:42 CPU time
after:  4:34 CPU time

Most of this time is saved when applying the changelog and manifest
revlogs:

before: 2:30 CPU time
after:  1:17 CPU time

That nearly a 50% reduction in CPU time applying changesets and
manifests!

Applying a gzipped bundle of the same repo (effectively simulating a
`hg clone` over HTTP) showed a similar speedup:

before: 5:53 CPU time
after:  4:46 CPU time

Wall time improvements were basically the same as CPU time.

I didn't measure explicitly, but it feels like most of the time
is saved when processing manifests. This makes sense, as large
manifests tend to have very long delta chains and thus benefit the
most from this cache.

So, this change effectively makes changegroup application (which is
used by `hg unbundle`, `hg clone`, `hg pull`, `hg unshelve`, and
various other commands) significantly faster when delta chains are
long (which can happen on repos with large numbers of files and thus
large manifests).

In theory, this change can result in more memory utilization. However,
we're caching a dict of ints. At most we have 200 ints + Python object
overhead per revlog. And, the cache is really only populated when
performing read-heavy operations, such as adding changegroups or
scanning an individual revlog. For memory bloat to be an issue, we'd
need to scan/read several revisions from several revlogs all while
having active references to several revlogs. I don't think there are
many operations that do this, so I don't think memory bloat from the
cache will be an issue.

  $ hg init ignorerepo
  $ cd ignorerepo

Issue562: .hgignore requires newline at end:

  $ touch foo
  $ touch bar
  $ touch baz
  $ cat > makeignore.py <<EOF
  > f = open(".hgignore", "w")
  > f.write("ignore\n")
  > f.write("foo\n")
  > # No EOL here
  > f.write("bar")
  > f.close()
  > EOF

  $ python makeignore.py

Should display baz only:

  $ hg status
  ? baz

  $ rm foo bar baz .hgignore makeignore.py

  $ touch a.o
  $ touch a.c
  $ touch syntax
  $ mkdir dir
  $ touch dir/a.o
  $ touch dir/b.o
  $ touch dir/c.o

  $ hg add dir/a.o
  $ hg commit -m 0
  $ hg add dir/b.o

  $ hg status
  A dir/b.o
  ? a.c
  ? a.o
  ? dir/c.o
  ? syntax

  $ echo "*.o" > .hgignore
  $ hg status
  abort: $TESTTMP/ignorerepo/.hgignore: invalid pattern (relre): *.o (glob)
  [255]

  $ echo ".*\.o" > .hgignore
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? syntax

Ensure that comments work:

  $ touch 'foo#bar' 'quux#'
#if no-windows
  $ touch 'baz\#wat'
#endif
  $ cat <<'EOF' >> .hgignore
  > # full-line comment
  >   # whitespace-only comment line
  > syntax# pattern, no whitespace, then comment
  > a.c  # pattern, then whitespace, then comment
  > baz\\# # escaped comment character
  > foo\#b # escaped comment character
  > quux\## escaped comment character at end of name
  > EOF
  $ hg status
  A dir/b.o
  ? .hgignore
  $ rm 'foo#bar' 'quux#'
#if no-windows
  $ rm 'baz\#wat'
#endif

Check it does not ignore the current directory '.':

  $ echo "^\." > .hgignore
  $ hg status
  A dir/b.o
  ? a.c
  ? a.o
  ? dir/c.o
  ? syntax

Test that patterns from ui.ignore options are read:

  $ echo > .hgignore
  $ cat >> $HGRCPATH << EOF
  > [ui]
  > ignore.other = $TESTTMP/ignorerepo/.hg/testhgignore
  > EOF
  $ echo "glob:**.o" > .hg/testhgignore
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? syntax

empty out testhgignore
  $ echo > .hg/testhgignore

Test relative ignore path (issue4473):

  $ cat >> $HGRCPATH << EOF
  > [ui]
  > ignore.relative = .hg/testhgignorerel
  > EOF
  $ echo "glob:*.o" > .hg/testhgignorerel
  $ cd dir
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? syntax

  $ cd ..
  $ echo > .hg/testhgignorerel
  $ echo "syntax: glob" > .hgignore
  $ echo "re:.*\.o" >> .hgignore
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? syntax

  $ echo "syntax: invalid" > .hgignore
  $ hg status
  $TESTTMP/ignorerepo/.hgignore: ignoring invalid syntax 'invalid' (glob)
  A dir/b.o
  ? .hgignore
  ? a.c
  ? a.o
  ? dir/c.o
  ? syntax

  $ echo "syntax: glob" > .hgignore
  $ echo "*.o" >> .hgignore
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? syntax

  $ echo "relglob:syntax*" > .hgignore
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? a.o
  ? dir/c.o

  $ echo "relglob:*" > .hgignore
  $ hg status
  A dir/b.o

  $ cd dir
  $ hg status .
  A b.o

  $ hg debugignore
  (?:(?:|.*/)[^/]*(?:/|$))

  $ hg debugignore b.o
  b.o is ignored
  (ignore rule in $TESTTMP/ignorerepo/.hgignore, line 1: '*') (glob)

  $ cd ..

Check patterns that match only the directory

  $ echo "^dir\$" > .hgignore
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? a.o
  ? syntax

Check recursive glob pattern matches no directories (dir/**/c.o matches dir/c.o)

  $ echo "syntax: glob" > .hgignore
  $ echo "dir/**/c.o" >> .hgignore
  $ touch dir/c.o
  $ mkdir dir/subdir
  $ touch dir/subdir/c.o
  $ hg status
  A dir/b.o
  ? .hgignore
  ? a.c
  ? a.o
  ? syntax
  $ hg debugignore a.c
  a.c is not ignored
  $ hg debugignore dir/c.o
  dir/c.o is ignored
  (ignore rule in $TESTTMP/ignorerepo/.hgignore, line 2: 'dir/**/c.o') (glob)

Check using 'include:' in ignore file

  $ hg purge --all --config extensions.purge=
  $ touch foo.included

  $ echo ".*.included" > otherignore
  $ hg status -I "include:otherignore"
  ? foo.included

  $ echo "include:otherignore" >> .hgignore
  $ hg status
  A dir/b.o
  ? .hgignore
  ? otherignore

Check recursive uses of 'include:'

  $ echo "include:nested/ignore" >> otherignore
  $ mkdir nested
  $ echo "glob:*ignore" > nested/ignore
  $ hg status
  A dir/b.o

  $ cp otherignore goodignore
  $ echo "include:badignore" >> otherignore
  $ hg status
  skipping unreadable pattern file 'badignore': No such file or directory
  A dir/b.o

  $ mv goodignore otherignore

Check using 'include:' while in a non-root directory

  $ cd ..
  $ hg -R ignorerepo status
  A dir/b.o
  $ cd ignorerepo

Check including subincludes

  $ hg revert -q --all
  $ hg purge --all --config extensions.purge=
  $ echo ".hgignore" > .hgignore
  $ mkdir dir1 dir2
  $ touch dir1/file1 dir1/file2 dir2/file1 dir2/file2
  $ echo "subinclude:dir2/.hgignore" >> .hgignore
  $ echo "glob:file*2" > dir2/.hgignore
  $ hg status
  ? dir1/file1
  ? dir1/file2
  ? dir2/file1

Check including subincludes with regexs

  $ echo "subinclude:dir1/.hgignore" >> .hgignore
  $ echo "regexp:f.le1" > dir1/.hgignore

  $ hg status
  ? dir1/file2
  ? dir2/file1

Check multiple levels of sub-ignores

  $ mkdir dir1/subdir
  $ touch dir1/subdir/subfile1 dir1/subdir/subfile3 dir1/subdir/subfile4
  $ echo "subinclude:subdir/.hgignore" >> dir1/.hgignore
  $ echo "glob:subfil*3" >> dir1/subdir/.hgignore

  $ hg status
  ? dir1/file2
  ? dir1/subdir/subfile4
  ? dir2/file1

Check include subignore at the same level

  $ mv dir1/subdir/.hgignore dir1/.hgignoretwo
  $ echo "regexp:f.le1" > dir1/.hgignore
  $ echo "subinclude:.hgignoretwo" >> dir1/.hgignore
  $ echo "glob:file*2" > dir1/.hgignoretwo

  $ hg status | grep file2
  [1]
  $ hg debugignore dir1/file2
  dir1/file2 is ignored
  (ignore rule in dir2/.hgignore, line 1: 'file*2')

#if windows

Windows paths are accepted on input

  $ rm dir1/.hgignore
  $ echo "dir1/file*" >> .hgignore
  $ hg debugignore "dir1\file2"
  dir1\file2 is ignored
  (ignore rule in $TESTTMP\ignorerepo\.hgignore, line 4: 'dir1/file*')
  $ hg up -qC .

#endif