tests/test-default-push.t

wireproto: compress data from a generator

Currently, the "getbundle" wire protocol command obtains a generator of
data, converts it to a util.chunkbuffer, then converts it back to a
generator via the protocol's groupchunks() implementation. For the SSH
protocol, groupchunks() simply reads 4kb chunks then write()s the
data to a file descriptor. For the HTTP protocol, groupchunks() reads
32kb chunks, feeds those into a zlib compressor, emits compressed data
as it is available, and that is sent to the WSGI layer, where it is
likely turned into HTTP chunked transfer chunks as is or further
buffered and turned into a larger chunk.

For both the SSH and HTTP protocols, there is inefficiency from using
util.chunkbuffer.

For SSH, emitting consistent 4kb chunks sounds nice. However, the file
descriptor it is writing to is almost certainly buffered. That means
that a Python .write() probably doesn't translate into exactly what is
written to the I/O layer.

For HTTP, we're going through an intermediate layer to zlib compress
data. So all util.chunkbuffer is doing is ensuring that the chunks we
feed into the zlib compressor are of uniform size. This means more CPU
time in Python buffering and emitting chunks in util.chunkbuffer but
fewer function calls to zlib.

This patch introduces and implements a new wire protocol abstract
method: compresschunks(). It is like groupchunks() except it operates
on a generator instead of something with a .read(). The SSH
implementation simply proxies chunks. The HTTP implementation uses
zlib compression.

To avoid duplicate code, the HTTP groupchunks() has been reimplemented
in terms of compresschunks().

To prove this all works, the "getbundle" wire protocol command has been
switched to compresschunks(). This removes the util.chunkbuffer from
that command. Now, data essentially streams straight from the
changegroup emitter to the wire, possibly through a zlib compressor.
Generators all the way, baby.

There were slim to no performance changes on the server as measured
with the mozilla-central repository. This is likely because CPU
time is dominated by reading revlogs, producing the changegroup, and
zlib compressing the output stream. Still, this brings us a little
closer to our ideal of using generators everywhere.

  $ hg init a

  $ echo a > a/a
  $ hg --cwd a ci -Ama
  adding a

  $ hg clone a c
  updating to branch default
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved

  $ hg clone a b
  updating to branch default
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved

  $ echo b >> b/a
  $ hg --cwd b ci -mb

Push should provide a hint when both 'default' and 'default-push' not set:
  $ cd c
  $ hg push --config paths.default=
  abort: default repository not configured!
  (see 'hg help config.paths')
  [255]

  $ cd ..

Push should push to 'default' when 'default-push' not set:

  $ hg --cwd b push
  pushing to $TESTTMP/a (glob)
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files

Push should push to 'default-push' when set:

  $ echo '[paths]' >> b/.hg/hgrc
  $ echo 'default-push = ../c' >> b/.hg/hgrc
  $ hg --cwd b push
  pushing to $TESTTMP/c (glob)
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files

But push should push to 'default' if explicitly specified (issue5000):

  $ hg --cwd b push default
  pushing to $TESTTMP/a (glob)
  searching for changes
  no changes found
  [1]

Push should push to 'default-push' when 'default' is not set

  $ hg -q clone a push-default-only
  $ cd push-default-only
  $ rm .hg/hgrc

  $ touch foo
  $ hg -q commit -A -m 'add foo'
  $ hg --config paths.default-push=../a push
  pushing to $TESTTMP/a (glob)
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files

  $ cd ..

Pushing to a path that isn't defined should not fall back to default

  $ hg --cwd b push doesnotexist
  abort: repository doesnotexist does not exist!
  [255]

:pushurl is used when defined

  $ hg -q clone a pushurlsource
  $ hg -q clone a pushurldest
  $ cd pushurlsource

Windows needs a leading slash to make a URL that passes all of the checks
  $ WD=`pwd`
#if windows
  $ WD="/$WD"
#endif
  $ cat > .hg/hgrc << EOF
  > [paths]
  > default = https://example.com/not/relevant
  > default:pushurl = file://$WD/../pushurldest
  > EOF

  $ touch pushurl
  $ hg -q commit -A -m 'add pushurl'
  $ hg push
  pushing to file:/*/$TESTTMP/pushurlsource/../pushurldest (glob)
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files

:pushrev is used when no -r is passed

  $ cat >> .hg/hgrc << EOF
  > default:pushrev = .
  > EOF
  $ hg -q up -r 0
  $ echo head1 > foo
  $ hg -q commit -A -m head1
  $ hg -q up -r 0
  $ echo head2 > foo
  $ hg -q commit -A -m head2
  $ hg push -f
  pushing to file:/*/$TESTTMP/pushurlsource/../pushurldest (glob)
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files (+1 heads)

  $ hg --config 'paths.default:pushrev=draft()' push -f
  pushing to file:/*/$TESTTMP/pushurlsource/../pushurldest (glob)
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files (+1 heads)

Invalid :pushrev raises appropriately

  $ hg --config 'paths.default:pushrev=notdefined()' push
  pushing to file:/*/$TESTTMP/pushurlsource/../pushurldest (glob)
  hg: parse error: unknown identifier: notdefined
  [255]

  $ hg --config 'paths.default:pushrev=(' push
  pushing to file:/*/$TESTTMP/pushurlsource/../pushurldest (glob)
  hg: parse error at 1: not a prefix: end
  [255]

  $ cd ..