view tests/test-rebase-rename.t @ 44950:f9734b2d59cc

py3: make stdout line-buffered if connected to a TTY

Status messages that are to be shown on the terminal should be written to the
file descriptor before anything further is done, to keep the user updated.

One common way to achieve this is to make stdout line-buffered if it is
connected to a TTY. This is done on Python 2 (except on Windows, where libc,
which the CPython 2 streams depend on, does not properly support this).

Python 3 rolls it own I/O streams. On Python 3, buffered binary streams can't be
set line-buffered. The previous code (added in 227ba1afcb65) incorrectly
assumed that on Python 3, pycompat.stdout (sys.stdout.buffer) is already
line-buffered. However the interpreter initializes it with a block-buffered
stream or an unbuffered stream (when the -u option or the PYTHONUNBUFFERED
environment variable is set), never with a line-buffered stream.

One example where the current behavior is unacceptable is when running
`hg pull https://www.mercurial-scm.org/repo/hg` on Python 3, where the line
"pulling from https://www.mercurial-scm.org/repo/hg" does not appear on the
terminal before the hg process blocks while waiting for the server.

Various approaches to fix this problem are possible, including:

1. Weaken the contract of procutil.stdout to not give any guarantees about
   buffering behavior. In this case, users of procutil.stdout need to be
   changed to do enough flushes. In particular,
  1. either ui must insert enough flushes for ui.write() and friends, or
  2. ui.write() and friends get split into flushing and fully buffered
     methods, or
  3. users of ui.write() and friends must flush explicitly.
2. Make stdout unbuffered.
3. Make stdout line-buffered. Since Python 3 does not natively support that for
   binary streams, we must implement it ourselves.

(2.) is problematic because using unbuffered I/O changes the performance
characteristics significantly compared to line-buffered (which is used on
Python 2) and this would be a regression.

(1.2.) and (1.3) are a substantial amount of work. It’s unclear whether the
added complexity would be justified, given that raw performance doesn’t matter
that much when writing to a terminal much faster than the user could read it.

(1.1.) pushes complexity into the ui class instead of separating the concern of
how stdout is buffered. Other users of procutil.stdout would still need to take
care of the flushes.

This patch implements (3.). The general performance considerations are very
similar to (1.1.). The extra method invocation and method forwarding add a
little more overhead if the class is used. In exchange, it doesn’t add overhead
if not used.


For the benchmarks, I compared the previous implementation (incorrect on Python
3), (1.1.), (3.) and (2.). The command was chosen so that the streams were
configured as if they were writing to a TTY, but actually write to a pager,
which is also the default:
HGRCPATH=/dev/null python3 ./hg --cwd ~/vcs/mozilla-central --time --pager yes --config pager.pager='cat > /dev/null' status --all

previous:
time: real 7.880 secs (user 7.290+0.050 sys 0.580+0.170)
time: real 7.830 secs (user 7.220+0.070 sys 0.590+0.140)
time: real 7.800 secs (user 7.210+0.050 sys 0.570+0.170)

(1.1.) using Yuya Nishihara’s patch:
time: real 9.860 secs (user 8.670+0.350 sys 1.160+0.830)
time: real 9.540 secs (user 8.430+0.370 sys 1.100+0.770)
time: real 9.830 secs (user 8.630+0.370 sys 1.180+0.840)

(3.) using this patch:
time: real 9.580 secs (user 8.480+0.350 sys 1.090+0.770)
time: real 9.670 secs (user 8.480+0.330 sys 1.170+0.860)
time: real 9.640 secs (user 8.500+0.350 sys 1.130+0.810)

(2.) using a previous patch by me:
time: real 10.480 secs (user 8.850+0.720 sys 1.590+1.500)
time: real 10.490 secs (user 8.750+0.750 sys 1.710+1.470)
time: real 10.240 secs (user 8.600+0.700 sys 1.590+1.510)


As expected, there’s no difference on Python 2, as exactly the same code paths
are used:

previous:
time: real 6.950 secs (user 5.870+0.330 sys 1.070+0.770)
time: real 7.040 secs (user 6.040+0.360 sys 0.980+0.750)
time: real 7.070 secs (user 5.950+0.360 sys 1.100+0.760)

this patch:
time: real 7.010 secs (user 5.900+0.390 sys 1.070+0.730)
time: real 7.000 secs (user 5.850+0.350 sys 1.120+0.760)
time: real 7.000 secs (user 5.790+0.380 sys 1.170+0.710)

  $ cat >> $HGRCPATH <<EOF
  > [extensions]
  > rebase=
  > 
  > [alias]
  > tlog  = log --template "{rev}: {node|short} '{desc}' {branches}\n"
  > tglog = tlog --graph
  > EOF


  $ hg init a
  $ cd a

  $ mkdir d
  $ echo a > a
  $ hg ci -Am A
  adding a

  $ echo b > d/b
  $ hg ci -Am B
  adding d/b

  $ hg mv d d-renamed
  moving d/b to d-renamed/b
  $ hg ci -m 'rename B'

  $ hg up -q -C 1

  $ hg mv a a-renamed
  $ echo x > d/x
  $ hg add d/x

  $ hg ci -m 'rename A'
  created new head

  $ hg tglog
  @  3: 73a3ee40125d 'rename A'
  |
  | o  2: 220d0626d185 'rename B'
  |/
  o  1: 3ab5da9a5c01 'B'
  |
  o  0: 1994f17a630e 'A'
  

Rename is tracked:

  $ hg tlog -p --git -r tip
  3: 73a3ee40125d 'rename A' 
  diff --git a/a b/a-renamed
  rename from a
  rename to a-renamed
  diff --git a/d/x b/d/x
  new file mode 100644
  --- /dev/null
  +++ b/d/x
  @@ -0,0 +1,1 @@
  +x
  
Rebase the revision containing the rename:

  $ hg rebase -s 3 -d 2
  rebasing 3:73a3ee40125d "rename A" (tip)
  saved backup bundle to $TESTTMP/a/.hg/strip-backup/73a3ee40125d-1d78ebcf-rebase.hg

  $ hg tglog
  @  3: 032a9b75e83b 'rename A'
  |
  o  2: 220d0626d185 'rename B'
  |
  o  1: 3ab5da9a5c01 'B'
  |
  o  0: 1994f17a630e 'A'
  

Rename is not lost:

  $ hg tlog -p --git -r tip
  3: 032a9b75e83b 'rename A' 
  diff --git a/a b/a-renamed
  rename from a
  rename to a-renamed
  diff --git a/d-renamed/x b/d-renamed/x
  new file mode 100644
  --- /dev/null
  +++ b/d-renamed/x
  @@ -0,0 +1,1 @@
  +x
  

Rebased revision does not contain information about b (issue3739)

  $ hg log -r 3 --debug
  changeset:   3:032a9b75e83bff1dcfb6cbfa4ef50a704bf1b569
  tag:         tip
  phase:       draft
  parent:      2:220d0626d185f372d9d8f69d9c73b0811d7725f7
  parent:      -1:0000000000000000000000000000000000000000
  manifest:    3:035d66b27a1b06b2d12b46d41a39adb7a200c370
  user:        test
  date:        Thu Jan 01 00:00:00 1970 +0000
  files+:      a-renamed d-renamed/x
  files-:      a
  extra:       branch=default
  extra:       rebase_source=73a3ee40125d6f0f347082e5831ceccb3f005f8a
  description:
  rename A
  
  

  $ repeatchange() {
  >   hg checkout $1
  >   hg cp a z
  >   echo blah >> z
  >   hg commit -Am "$2" --user "$3"
  > }
  $ repeatchange 1 "E" "user1"
  2 files updated, 0 files merged, 3 files removed, 0 files unresolved
  created new head
  $ repeatchange 1 "E" "user2"
  0 files updated, 0 files merged, 1 files removed, 0 files unresolved
  created new head
  $ hg tglog
  @  5: af8ad1f97097 'E'
  |
  | o  4: 60f545c27784 'E'
  |/
  | o  3: 032a9b75e83b 'rename A'
  | |
  | o  2: 220d0626d185 'rename B'
  |/
  o  1: 3ab5da9a5c01 'B'
  |
  o  0: 1994f17a630e 'A'
  
  $ hg rebase -s 5 -d 4
  rebasing 5:af8ad1f97097 "E" (tip)
  note: not rebasing 5:af8ad1f97097 "E" (tip), its destination already has all its changes
  saved backup bundle to $TESTTMP/a/.hg/strip-backup/af8ad1f97097-c3e90708-rebase.hg
  $ hg tglog
  @  4: 60f545c27784 'E'
  |
  | o  3: 032a9b75e83b 'rename A'
  | |
  | o  2: 220d0626d185 'rename B'
  |/
  o  1: 3ab5da9a5c01 'B'
  |
  o  0: 1994f17a630e 'A'
  
  $ hg export tip
  # HG changeset patch
  # User user1
  # Date 0 0
  #      Thu Jan 01 00:00:00 1970 +0000
  # Node ID 60f545c277846e6bad309919bae3ae106f59cb39
  # Parent  3ab5da9a5c01faa02c20f2ec4870a4f689c92da6
  E
  
  diff -r 3ab5da9a5c01 -r 60f545c27784 z
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/z	Thu Jan 01 00:00:00 1970 +0000
  @@ -0,0 +1,2 @@
  +a
  +blah

  $ cd ..


  $ hg init b
  $ cd b

  $ echo a > a
  $ hg ci -Am A
  adding a

  $ echo b > b
  $ hg ci -Am B
  adding b

  $ hg cp b b-copied
  $ hg ci -Am 'copy B'

  $ hg up -q -C 1

  $ hg cp a a-copied
  $ hg ci -m 'copy A'
  created new head

  $ hg tglog
  @  3: 0a8162ff18a8 'copy A'
  |
  | o  2: 39e588434882 'copy B'
  |/
  o  1: 6c81ed0049f8 'B'
  |
  o  0: 1994f17a630e 'A'
  
Copy is tracked:

  $ hg tlog -p --git -r tip
  3: 0a8162ff18a8 'copy A' 
  diff --git a/a b/a-copied
  copy from a
  copy to a-copied
  
Rebase the revision containing the copy:

  $ hg rebase -s 3 -d 2
  rebasing 3:0a8162ff18a8 "copy A" (tip)
  saved backup bundle to $TESTTMP/b/.hg/strip-backup/0a8162ff18a8-dd06302a-rebase.hg

  $ hg tglog
  @  3: 98f6e6dbf45a 'copy A'
  |
  o  2: 39e588434882 'copy B'
  |
  o  1: 6c81ed0049f8 'B'
  |
  o  0: 1994f17a630e 'A'
  

Copy is not lost:

  $ hg tlog -p --git -r tip
  3: 98f6e6dbf45a 'copy A' 
  diff --git a/a b/a-copied
  copy from a
  copy to a-copied
  

Rebased revision does not contain information about b (issue3739)

  $ hg log -r 3 --debug
  changeset:   3:98f6e6dbf45ab54079c2237fbd11066a5c41a11d
  tag:         tip
  phase:       draft
  parent:      2:39e588434882ff77d01229d169cdc77f29e8855e
  parent:      -1:0000000000000000000000000000000000000000
  manifest:    3:2232f329d66fffe3930d43479ae624f66322b04d
  user:        test
  date:        Thu Jan 01 00:00:00 1970 +0000
  files+:      a-copied
  extra:       branch=default
  extra:       rebase_source=0a8162ff18a8900df8df8ef7ac0046955205613e
  description:
  copy A
  
  

  $ cd ..


Test rebase across repeating renames:

  $ hg init repo

  $ cd repo

  $ echo testing > file1.txt
  $ hg add file1.txt
  $ hg ci -m "Adding file1"

  $ hg rename file1.txt file2.txt
  $ hg ci -m "Rename file1 to file2"

  $ echo Unrelated change > unrelated.txt
  $ hg add unrelated.txt
  $ hg ci -m "Unrelated change"

  $ hg rename file2.txt file1.txt
  $ hg ci -m "Rename file2 back to file1"

  $ hg update -r -2
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved

  $ echo Another unrelated change >> unrelated.txt
  $ hg ci -m "Another unrelated change"
  created new head

  $ hg tglog
  @  4: b918d683b091 'Another unrelated change'
  |
  | o  3: 1ac17e43d8aa 'Rename file2 back to file1'
  |/
  o  2: 480101d66d8d 'Unrelated change'
  |
  o  1: be44c61debd2 'Rename file1 to file2'
  |
  o  0: 8ce9a346991d 'Adding file1'
  

  $ hg rebase -s 4 -d 3
  rebasing 4:b918d683b091 "Another unrelated change" (tip)
  saved backup bundle to $TESTTMP/repo/.hg/strip-backup/b918d683b091-3024bc57-rebase.hg

  $ hg diff --stat -c .
   unrelated.txt |  1 +
   1 files changed, 1 insertions(+), 0 deletions(-)

  $ cd ..

Verify that copies get preserved (issue4192).
  $ hg init copy-gets-preserved
  $ cd copy-gets-preserved

  $ echo a > a
  $ hg add a
  $ hg commit --message "File a created"
  $ hg copy a b
  $ echo b > b
  $ hg commit --message "File b created as copy of a and modified"
  $ hg copy b c
  $ echo c > c
  $ hg commit --message "File c created as copy of b and modified"
  $ hg copy c d
  $ echo d > d
  $ hg commit --message "File d created as copy of c and modified"

Note that there are four entries in the log for d
  $ hg tglog --follow d
  @  3: 421b7e82bb85 'File d created as copy of c and modified'
  |
  o  2: 327f772bc074 'File c created as copy of b and modified'
  |
  o  1: 79d255d24ad2 'File b created as copy of a and modified'
  |
  o  0: b220cd6d2326 'File a created'
  
Update back to before we performed copies, and inject an unrelated change.
  $ hg update 0
  0 files updated, 0 files merged, 3 files removed, 0 files unresolved

  $ echo unrelated > unrelated
  $ hg add unrelated
  $ hg commit --message "Unrelated file created"
  created new head
  $ hg update 4
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved

Rebase the copies on top of the unrelated change.
  $ hg rebase --source 1 --dest 4
  rebasing 1:79d255d24ad2 "File b created as copy of a and modified"
  rebasing 2:327f772bc074 "File c created as copy of b and modified"
  rebasing 3:421b7e82bb85 "File d created as copy of c and modified"
  saved backup bundle to $TESTTMP/copy-gets-preserved/.hg/strip-backup/79d255d24ad2-a2265555-rebase.hg
  $ hg update 4
  3 files updated, 0 files merged, 0 files removed, 0 files unresolved

There should still be four entries in the log for d
  $ hg tglog --follow d
  @  4: dbb9ba033561 'File d created as copy of c and modified'
  |
  o  3: af74b229bc02 'File c created as copy of b and modified'
  |
  o  2: 68bf06433839 'File b created as copy of a and modified'
  :
  o  0: b220cd6d2326 'File a created'
  
Same steps as above, but with --collapse on rebase to make sure the
copy records collapse correctly.
  $ hg co 1
  0 files updated, 0 files merged, 3 files removed, 0 files unresolved
  $ echo more >> unrelated
  $ hg ci -m 'unrelated commit is unrelated'
  created new head
  $ hg rebase -s 2 --dest 5 --collapse
  rebasing 2:68bf06433839 "File b created as copy of a and modified"
  rebasing 3:af74b229bc02 "File c created as copy of b and modified"
  merging b and c to c
  rebasing 4:dbb9ba033561 "File d created as copy of c and modified"
  merging c and d to d
  saved backup bundle to $TESTTMP/copy-gets-preserved/.hg/strip-backup/68bf06433839-dde37595-rebase.hg
  $ hg co tip
  3 files updated, 0 files merged, 0 files removed, 0 files unresolved

This should show both revision 3 and 0 since 'd' was transitively a
copy of 'a'.

  $ hg tglog --follow d
  @  3: 5a46b94210e5 'Collapsed revision
  :  * File b created as copy of a and modified
  :  * File c created as copy of b and modified
  :  * File d created as copy of c and modified'
  o  0: b220cd6d2326 'File a created'
  

  $ cd ..