tests/test-inotify.t

branchmap: update cache of 'unserved' filter on new changesets

The `commitctx` and `addchangegroup` methods of repo upgrade branchcache after
completion. This behavior aims to keep the branchcache in sync for read only
process as hgweb. See ee317dbfb9d0 for details.

Since changelog filtering is used, those calls only update the cache for unfiltered repo.
One of no interest for typical read only process like hgweb.

Note: By chance in basic case, `repo.unfiltered() == repo.filtered('unserved')`

This changesets have the "unserved" cache updated instead. I think this is the
only cache that matter for hgweb.

We could imagine updating all possible branchcaches instead but:
- I'm not sure it would have any benefit impact. It may even increase the odd of
  all cache being invalidated.
- This is more complicated change.

So I'm going for updating a single cache only which is already better that
updating a cache nobody cares about.

This changeset have a few expected impact on the testsuite are different cache
are updated.

  $ "$TESTDIR/hghave" inotify || exit 80
  $ hg init repo1
  $ cd repo1
  $ touch a b c d e
  $ mkdir dir
  $ mkdir dir/bar
  $ touch dir/x dir/y dir/bar/foo
  $ hg ci -Am m
  adding a
  adding b
  adding c
  adding d
  adding dir/bar/foo
  adding dir/x
  adding dir/y
  adding e
  $ cd ..
  $ hg clone repo1 repo2
  updating to branch default
  8 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ echo "[extensions]" >> $HGRCPATH
  $ echo "inotify=" >> $HGRCPATH
  $ cd repo2
  $ echo b >> a

check that daemon started automatically works correctly
and make sure that inotify.pidfile works

  $ hg --config "inotify.pidfile=../hg2.pid" status
  M a

make sure that pidfile worked. Output should be silent.

  $ kill `cat ../hg2.pid`
  $ cd ../repo1

inserve

  $ hg inserve -d --pid-file=hg.pid
  $ cat hg.pid >> "$DAEMON_PIDS"

let the daemon finish its stuff

  $ sleep 1

cannot start, already bound

  $ hg inserve
  abort: inotify-server: cannot start: socket is already bound
  [255]

issue907

  $ hg status
  ? hg.pid

clean

  $ hg status -c
  C a
  C b
  C c
  C d
  C dir/bar/foo
  C dir/x
  C dir/y
  C e

all

  $ hg status -A
  ? hg.pid
  C a
  C b
  C c
  C d
  C dir/bar/foo
  C dir/x
  C dir/y
  C e

path patterns

  $ echo x > dir/x
  $ hg status .
  M dir/x
  ? hg.pid
  $ hg status dir
  M dir/x
  $ cd dir
  $ hg status .
  M x
  $ cd ..

issue 1375
testing that we can remove a folder and then add a file with the same name
issue 1375

  $ mkdir h
  $ echo h > h/h
  $ hg ci -Am t
  adding h/h
  adding hg.pid
  $ hg rm h
  removing h/h
  $ echo h >h
  $ hg add h
  $ hg status
  A h
  R h/h
  $ hg ci -m0

Test for issue1735: inotify watches files in .hg/merge

  $ hg st
  $ echo a > a
  $ hg ci -Am a
  $ hg st
  $ echo b >> a
  $ hg ci -m ab
  $ hg st
  $ echo c >> a
  $ hg st
  M a
  $ HGMERGE=internal:local hg up 0
  1 files updated, 1 files merged, 2 files removed, 0 files unresolved
  $ hg st
  M a
  $ HGMERGE=internal:local hg up
  3 files updated, 1 files merged, 0 files removed, 0 files unresolved
  $ hg st
  M a

Test for 1844: "hg ci folder" will not commit all changes beneath "folder"

  $ mkdir 1844
  $ echo a > 1844/foo
  $ hg add 1844
  adding 1844/foo
  $ hg ci -m 'working'
  $ echo b >> 1844/foo
  $ hg ci 1844 -m 'broken'

Test for issue884: "Build products not ignored until .hgignore is touched"

  $ echo '^build$' > .hgignore
  $ hg add .hgignore
  $ hg ci .hgignore -m 'ignorelist'

Now, lets add some build products...

  $ mkdir build
  $ touch build/x
  $ touch build/y

build/x & build/y shouldn't appear in "hg st"

  $ hg st
  $ kill `cat hg.pid`

  $ cd ..

Ensure that if the repo is in a directory whose name is too long, the
unix domain socket is reached through a symlink (issue1208).

  $ mkdir 0_3456789_10_456789_20_456789_30_456789_40_456789_50_45678_
  $ cd 0_3456789_10_456789_20_456789_30_456789_40_456789_50_45678_
  $ mkdir 60_456789_70_456789_80_456789_90_456789_100_56789_
  $ cd 60_456789_70_456789_80_456789_90_456789_100_56789_

  $ hg --config inotify.pidfile=hg3.pid clone -q ../../repo1
  $ readlink repo1/.hg/inotify.sock
  */inotify.sock (glob)

Trying to start the server a second time should fail as usual.

  $ hg --cwd repo1 inserve
  abort: inotify-server: cannot start: socket is already bound
  [255]

  $ kill `cat hg3.pid`