view tests/test-import-bypass.t @ 30818:4c0a5a256ae8

localrepo: experimental support for non-zlib revlog compression

The final part of integrating the compression manager APIs into
revlog storage is the plumbing for repositories to advertise they
are using non-zlib storage and for revlogs to instantiate a non-zlib
compression engine.

The main intent of the compression manager work was to zstd all
of the things. Adding zstd to revlogs has proved to be more involved
than other places because revlogs are... special. Very small inputs
and the use of delta chains (which are themselves a form of
compression) are a completely different use case from streaming
compression, which bundles and the wire protocol employ. I've
conducted numerous experiments with zstd in revlogs and have yet
to formalize compression settings and a storage architecture that
I'm confident I won't regret later. In other words, I'm not yet
ready to commit to a new mechanism for using zstd - or any other
compression format - in revlogs.

That being said, having some support for zstd (and other compression
formats) in revlogs in core is beneficial. It can allow others to
conduct experiments.

This patch introduces *highly experimental* support for non-zlib
compression formats in revlogs. Introduced is a config option to
control which compression engine to use. Also introduced is a namespace
of "exp-compression-*" requirements to denote support for non-zlib
compression in revlogs. I've prefixed the namespace with "exp-"
(short for "experimental") because I'm not confident of the
requirements "schema" and in no way want to give the illusion of
supporting these requirements in the future. I fully intend to drop
support for these requirements once we figure out what we're doing
with zstd in revlogs.

A good portion of the patch is teaching the requirements system
about registered compression engines and passing the requested
compression engine as an opener option so revlogs can instantiate
the proper compression engine for new operations.

That's a verbose way of saying "we can now use zstd in revlogs!"

On an `hg pull` conversion of the mozilla-unified repo with no extra
redelta settings (like aggressivemergedeltas), we can see the impact
of zstd vs zlib in revlogs:

$ hg perfrevlogchunks -c
! chunk
! wall 2.032052 comb 2.040000 user 1.990000 sys 0.050000 (best of 5)
! wall 1.866360 comb 1.860000 user 1.820000 sys 0.040000 (best of 6)

! chunk batch
! wall 1.877261 comb 1.870000 user 1.860000 sys 0.010000 (best of 6)
! wall 1.705410 comb 1.710000 user 1.690000 sys 0.020000 (best of 6)

$ hg perfrevlogchunks -m
! chunk
! wall 2.721427 comb 2.720000 user 2.640000 sys 0.080000 (best of 4)
! wall 2.035076 comb 2.030000 user 1.950000 sys 0.080000 (best of 5)

! chunk batch
! wall 2.614561 comb 2.620000 user 2.580000 sys 0.040000 (best of 4)
! wall 1.910252 comb 1.910000 user 1.880000 sys 0.030000 (best of 6)

$ hg perfrevlog -c -d 1
! wall 4.812885 comb 4.820000 user 4.800000 sys 0.020000 (best of 3)
! wall 4.699621 comb 4.710000 user 4.700000 sys 0.010000 (best of 3)

$ hg perfrevlog -m -d 1000
! wall 34.252800 comb 34.250000 user 33.730000 sys 0.520000 (best of 3)
! wall 24.094999 comb 24.090000 user 23.320000 sys 0.770000 (best of 3)

Only modest wins for the changelog. But manifest reading is
significantly faster. What's going on?

One reason might be data volume. zstd decompresses faster. So given
more bytes, it will put more distance between it and zlib.

Another reason is size. In the current design, zstd revlogs are
*larger*:

debugcreatestreamclonebundle (size in bytes)
zlib: 1,638,852,492
zstd: 1,680,601,332

I haven't investigated this fully, but I reckon a significant cause of
larger revlogs is that the zstd frame/header has more bytes than
zlib's. For very small inputs or data that doesn't compress well, we'll
tend to store more uncompressed chunks than with zlib (because the
compressed size isn't smaller than original). This will make revlog
reading faster because it is doing less decompression.

Moving on to bundle performance:

$ hg bundle -a -t none-v2 (total CPU time)
zlib: 102.79s
zstd:  97.75s

So, marginal CPU decrease for reading all chunks in all revlogs
(this is somewhat disappointing).

$ hg bundle -a -t <engine>-v2 (total CPU time)
zlib: 191.59s
zstd: 115.36s

This last test effectively measures the difference between zlib->zlib
and zstd->zstd for revlogs to bundle. This is a rough approximation of
what a server does during `hg clone`.

There are some promising results for zstd. But not enough for me to
feel comfortable advertising it to users. We'll get there...

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

  $ shortlog() {
  >     hg log -G --template '{rev}:{node|short} {author} {date|hgdate} - {branch} - {desc|firstline}\n'
  > }

Test --bypass with other options

  $ hg init repo-options
  $ cd repo-options
  $ echo a > a
  $ hg ci -Am adda
  adding a
  $ echo a >> a
  $ hg branch foo
  marked working directory as branch foo
  (branches are permanent and global, did you want a bookmark?)
  $ hg ci -Am changea
  $ hg export . > ../test.diff
  $ hg up null
  0 files updated, 0 files merged, 1 files removed, 0 files unresolved

Test importing an existing revision
(this also tests that "hg import" disallows combination of '--exact'
and '--edit')

  $ hg import --bypass --exact --edit ../test.diff
  abort: cannot use --exact with --edit
  [255]
  $ hg import --bypass --exact ../test.diff
  applying ../test.diff
  $ shortlog
  o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |
  o  0:07f494440405 test 0 0 - default - adda
  

Test failure without --exact

  $ hg import --bypass ../test.diff
  applying ../test.diff
  unable to find 'a' for patching
  (use '--prefix' to apply patch relative to the current directory)
  abort: patch failed to apply
  [255]
  $ hg st
  $ shortlog
  o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |
  o  0:07f494440405 test 0 0 - default - adda
  

Test --user, --date and --message

  $ hg up 0
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg import --bypass --u test2 -d '1 0' -m patch2 ../test.diff
  applying ../test.diff
  $ cat .hg/last-message.txt
  patch2 (no-eol)
  $ shortlog
  o  2:2e127d1da504 test2 1 0 - default - patch2
  |
  | o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |/
  @  0:07f494440405 test 0 0 - default - adda
  
  $ hg rollback
  repository tip rolled back to revision 1 (undo import)

Test --import-branch
(this also tests that editor is not invoked for '--bypass', if the
patch contains the commit message, regardless of '--edit')

  $ HGEDITOR=cat hg import --bypass --import-branch --edit ../test.diff
  applying ../test.diff
  $ shortlog
  o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |
  @  0:07f494440405 test 0 0 - default - adda
  
  $ hg rollback
  repository tip rolled back to revision 1 (undo import)

Test --strip

  $ hg import --bypass --strip 0 - <<EOF
  > # HG changeset patch
  > # User test
  > # Date 0 0
  > # Branch foo
  > # Node ID 4e322f7ce8e3e4203950eac9ece27bf7e45ffa6c
  > # Parent  07f4944404050f47db2e5c5071e0e84e7a27bba9
  > changea
  > 
  > diff -r 07f494440405 -r 4e322f7ce8e3 a
  > --- a	Thu Jan 01 00:00:00 1970 +0000
  > +++ a	Thu Jan 01 00:00:00 1970 +0000
  > @@ -1,1 +1,2 @@
  >  a
  > +a
  > EOF
  applying patch from stdin
  $ hg rollback
  repository tip rolled back to revision 1 (undo import)

Test --strip with --bypass

  $ mkdir -p dir/dir2
  $ echo bb > dir/dir2/b
  $ echo cc > dir/dir2/c
  $ echo d > dir/d
  $ hg ci -Am 'addabcd'
  adding dir/d
  adding dir/dir2/b
  adding dir/dir2/c
  $ shortlog
  @  2:d805bc8236b6 test 0 0 - default - addabcd
  |
  | o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |/
  o  0:07f494440405 test 0 0 - default - adda
  
  $ hg import --bypass --strip 2 --prefix dir/ - <<EOF
  > # HG changeset patch
  > # User test
  > # Date 0 0
  > # Branch foo
  > changeabcd
  > 
  > diff --git a/foo/a b/foo/a
  > new file mode 100644
  > --- /dev/null
  > +++ b/foo/a
  > @@ -0,0 +1 @@
  > +a
  > diff --git a/foo/dir2/b b/foo/dir2/b2
  > rename from foo/dir2/b
  > rename to foo/dir2/b2
  > diff --git a/foo/dir2/c b/foo/dir2/c
  > --- a/foo/dir2/c
  > +++ b/foo/dir2/c
  > @@ -0,0 +1 @@
  > +cc
  > diff --git a/foo/d b/foo/d
  > deleted file mode 100644
  > --- a/foo/d
  > +++ /dev/null
  > @@ -1,1 +0,0 @@
  > -d
  > EOF
  applying patch from stdin

  $ shortlog
  o  3:5bd46886ca3e test 0 0 - default - changeabcd
  |
  @  2:d805bc8236b6 test 0 0 - default - addabcd
  |
  | o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |/
  o  0:07f494440405 test 0 0 - default - adda
  
  $ hg diff --change 3 --git
  diff --git a/dir/a b/dir/a
  new file mode 100644
  --- /dev/null
  +++ b/dir/a
  @@ -0,0 +1,1 @@
  +a
  diff --git a/dir/d b/dir/d
  deleted file mode 100644
  --- a/dir/d
  +++ /dev/null
  @@ -1,1 +0,0 @@
  -d
  diff --git a/dir/dir2/b b/dir/dir2/b2
  rename from dir/dir2/b
  rename to dir/dir2/b2
  diff --git a/dir/dir2/c b/dir/dir2/c
  --- a/dir/dir2/c
  +++ b/dir/dir2/c
  @@ -1,1 +1,2 @@
   cc
  +cc
  $ hg -q --config extensions.strip= strip .

Test unsupported combinations

  $ hg import --bypass --no-commit ../test.diff
  abort: cannot use --no-commit with --bypass
  [255]
  $ hg import --bypass --similarity 50 ../test.diff
  abort: cannot use --similarity with --bypass
  [255]
  $ hg import --exact --prefix dir/ ../test.diff
  abort: cannot use --exact with --prefix
  [255]

Test commit editor
(this also tests that editor is invoked, if the patch doesn't contain
the commit message, regardless of '--edit')

  $ cat > ../test.diff <<EOF
  > diff -r 07f494440405 -r 4e322f7ce8e3 a
  > --- a/a	Thu Jan 01 00:00:00 1970 +0000
  > +++ b/a	Thu Jan 01 00:00:00 1970 +0000
  > @@ -1,1 +1,2 @@
  > -a
  > +b
  > +c
  > EOF
  $ HGEDITOR=cat hg import --bypass ../test.diff
  applying ../test.diff
  
  
  HG: Enter commit message.  Lines beginning with 'HG:' are removed.
  HG: Leave message empty to abort commit.
  HG: --
  HG: user: test
  HG: branch 'default'
  HG: changed a
  abort: empty commit message
  [255]

Test patch.eol is handled
(this also tests that editor is not invoked for '--bypass', if the
commit message is explicitly specified, regardless of '--edit')

  $ $PYTHON -c 'file("a", "wb").write("a\r\n")'
  $ hg ci -m makeacrlf
  $ HGEDITOR=cat hg import -m 'should fail because of eol' --edit --bypass ../test.diff
  applying ../test.diff
  patching file a
  Hunk #1 FAILED at 0
  abort: patch failed to apply
  [255]
  $ hg --config patch.eol=auto import -d '0 0' -m 'test patch.eol' --bypass ../test.diff
  applying ../test.diff
  $ shortlog
  o  3:c606edafba99 test 0 0 - default - test patch.eol
  |
  @  2:872023de769d test 0 0 - default - makeacrlf
  |
  | o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |/
  o  0:07f494440405 test 0 0 - default - adda
  

Test applying multiple patches

  $ hg up -qC 0
  $ echo e > e
  $ hg ci -Am adde
  adding e
  created new head
  $ hg export . > ../patch1.diff
  $ hg up -qC 1
  $ echo f > f
  $ hg ci -Am addf
  adding f
  $ hg export . > ../patch2.diff
  $ cd ..
  $ hg clone -r1 repo-options repo-multi1
  adding changesets
  adding manifests
  adding file changes
  added 2 changesets with 2 changes to 1 files
  updating to branch foo
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ cd repo-multi1
  $ hg up 0
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ hg import --bypass ../patch1.diff ../patch2.diff
  applying ../patch1.diff
  applying ../patch2.diff
  $ shortlog
  o  3:bc8ca3f8a7c4 test 0 0 - default - addf
  |
  o  2:16581080145e test 0 0 - default - adde
  |
  | o  1:4e322f7ce8e3 test 0 0 - foo - changea
  |/
  @  0:07f494440405 test 0 0 - default - adda
  

Test applying multiple patches with --exact

  $ cd ..
  $ hg clone -r1 repo-options repo-multi2
  adding changesets
  adding manifests
  adding file changes
  added 2 changesets with 2 changes to 1 files
  updating to branch foo
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ cd repo-multi2
  $ hg import --bypass --exact ../patch1.diff ../patch2.diff
  applying ../patch1.diff
  applying ../patch2.diff
  $ shortlog
  o  3:d60cb8989666 test 0 0 - foo - addf
  |
  | o  2:16581080145e test 0 0 - default - adde
  | |
  @ |  1:4e322f7ce8e3 test 0 0 - foo - changea
  |/
  o  0:07f494440405 test 0 0 - default - adda
  

  $ cd ..

Test avoiding editor invocation at applying the patch with --exact
even if commit message is empty

  $ cd repo-options

  $ echo a >> a
  $ hg commit -m ' '
  $ hg tip -T "{node}\n"
  1b77bc7d1db9f0e7f1716d515b630516ab386c89
  $ hg export -o ../empty-log.diff .
  $ hg update -q -C ".^1"
  $ hg --config extensions.strip= strip -q tip
  $ HGEDITOR=cat hg import --exact --bypass ../empty-log.diff
  applying ../empty-log.diff
  $ hg tip -T "{node}\n"
  1b77bc7d1db9f0e7f1716d515b630516ab386c89

  $ cd ..

#if symlink execbit

Test complicated patch with --exact

  $ hg init repo-exact
  $ cd repo-exact
  $ echo a > a
  $ echo c > c
  $ echo d > d
  $ echo e > e
  $ echo f > f
  $ chmod +x f
  $ ln -s c linkc
  $ hg ci -Am t
  adding a
  adding c
  adding d
  adding e
  adding f
  adding linkc
  $ hg cp a aa1
  $ echo b >> a
  $ echo b > b
  $ hg add b
  $ hg cp a aa2
  $ echo aa >> aa2
  $ chmod +x e
  $ chmod -x f
  $ ln -s a linka
  $ hg rm d
  $ hg rm linkc
  $ hg mv c cc
  $ hg ci -m patch
  $ hg export --git . > ../test.diff
  $ hg up -C null
  0 files updated, 0 files merged, 7 files removed, 0 files unresolved
  $ hg purge
  $ hg st
  $ hg import --bypass --exact ../test.diff
  applying ../test.diff

The patch should have matched the exported revision and generated no additional
data. If not, diff both heads to debug it.

  $ shortlog
  o  1:2978fd5c8aa4 test 0 0 - default - patch
  |
  o  0:a0e19e636a43 test 0 0 - default - t
  
#endif

  $ cd ..