view tests/test-push-cgi.t @ 30288:ceddc3d94d74

revlog: inline start() and end() for perf reasons

When I implemented `hg perfrevlogchunks`, one of the things that
stood out was N * _chunk() calls was ~38x slower than 1
_chunks() call. Specifically, on the mozilla-unified repo:

N*_chunk:  0.528997s
1*_chunks: 0.013735s

This repo has 352,097 changesets. So the average time per changeset
comes out to:

N*_chunk:  1.502us
1*_chunks: 0.039us

If you extrapolate these numbers to a repository with 1M changesets,
that comes out to 1.502s versus 0.039s, which is significant.

At these latencies, Python attribute lookups and function calls
matter. So, this patch inlines some code to cut down on that overhead.

The impact of this patch on N*_chunk() calls is clear:

! wall 0.528997 comb 0.520000 user 0.500000 sys 0.020000 (best of 19)
! wall 0.367723 comb 0.370000 user 0.350000 sys 0.020000 (best of 27)

So, we go from ~38x slower to ~27x. A nice improvement. But there's
still a long way to go.

It's worth noting that functionality like revsets perform changelog
lookups one revision at a time. So this code path is worth optimizing.

#require no-msys # MSYS will translate web paths as if they were file paths

This is a test of the push wire protocol over CGI-based hgweb.

initialize repository

  $ hg init r
  $ cd r
  $ echo a > a
  $ hg ci -A -m "0"
  adding a
  $ echo '[web]' > .hg/hgrc
  $ echo 'allow_push = *' >> .hg/hgrc
  $ echo 'push_ssl = false' >> .hg/hgrc

create hgweb invocation script

  $ cat >hgweb.cgi <<HGWEB
  > import cgitb
  > cgitb.enable()
  > from mercurial import demandimport; demandimport.enable()
  > from mercurial.hgweb import hgweb
  > from mercurial.hgweb import wsgicgi
  > application = hgweb('.', 'test repository')
  > wsgicgi.launch(application)
  > HGWEB
  $ chmod 755 hgweb.cgi

test preparation

  $ . "$TESTDIR/cgienv"
  $ REQUEST_METHOD="POST"; export REQUEST_METHOD
  $ CONTENT_TYPE="application/octet-stream"; export CONTENT_TYPE
  $ hg bundle --type v1 --all bundle.hg
  1 changesets found
  $ CONTENT_LENGTH=279; export CONTENT_LENGTH;

expect failure because heads doesn't match (formerly known as 'unsynced changes')

  $ QUERY_STRING="cmd=unbundle&heads=0000000000000000000000000000000000000000"; export QUERY_STRING
  $ python hgweb.cgi <bundle.hg >page1 2>&1
  $ cat page1
  Status: 200 Script output follows\r (esc)
  Content-Type: application/mercurial-0.1\r (esc)
  Content-Length: 64\r (esc)
  \r (esc)
  0
  repository changed while preparing changes - please try again

successful force push

  $ QUERY_STRING="cmd=unbundle&heads=666f726365"; export QUERY_STRING
  $ python hgweb.cgi <bundle.hg >page2 2>&1
  $ cat page2
  Status: 200 Script output follows\r (esc)
  Content-Type: application/mercurial-0.1\r (esc)
  Content-Length: 102\r (esc)
  \r (esc)
  1
  adding changesets
  adding manifests
  adding file changes
  added 0 changesets with 0 changes to 1 files

successful push, list of heads

  $ QUERY_STRING="cmd=unbundle&heads=f7b1eb17ad24730a1651fccd46c43826d1bbc2ac"; export QUERY_STRING
  $ python hgweb.cgi <bundle.hg >page3 2>&1
  $ cat page3
  Status: 200 Script output follows\r (esc)
  Content-Type: application/mercurial-0.1\r (esc)
  Content-Length: 102\r (esc)
  \r (esc)
  1
  adding changesets
  adding manifests
  adding file changes
  added 0 changesets with 0 changes to 1 files

successful push, SHA1 hash of heads (unbundlehash capability)

  $ QUERY_STRING="cmd=unbundle&heads=686173686564 5a785a5f9e0d433b88ed862b206b011b0c3a9d13"; export QUERY_STRING
  $ python hgweb.cgi <bundle.hg >page4 2>&1
  $ cat page4
  Status: 200 Script output follows\r (esc)
  Content-Type: application/mercurial-0.1\r (esc)
  Content-Length: 102\r (esc)
  \r (esc)
  1
  adding changesets
  adding manifests
  adding file changes
  added 0 changesets with 0 changes to 1 files

  $ cd ..