view tests/common-pattern.py @ 40021:c537144fdbef

wireprotov2: support response caching

One of the things I've learned from managing VCS servers over the
years is that they are hard to scale. It is well known that some
companies have very beefy (read: very expensive) servers to power
their VCS needs. It is also known that specialized servers for
various VCS exist in order to facilitate scaling servers. (Mercurial
is in this boat.)

One of the aspects that make a VCS server hard to scale is the
high CPU load incurred by constant client clone/pull operations.

To alleviate the scaling pain associated with data retrieval
operations, I want to integrate caching into the Mercurial wire
protocol server as robustly as possible such that servers can
aggressively cache responses and defer as much server load as
possible.

This commit represents the initial implementation of a general
caching layer in wire protocol version 2.

We define a new interface and behavior for a wire protocol cacher
in repository.py. (This is probably where a reviewer should look
first to understand what is going on.)

The bulk of the added code is in wireprotov2server.py, where we
define how a command can opt in to being cached and integrate
caching into command dispatching.

From a very high-level:

* A command can declare itself as cacheable by providing a callable
  that can be used to derive a cache key.
* At dispatch time, if a command is cacheable, we attempt to
  construct a cacher and use it for serving the request and/or
  caching the request.
* The dispatch layer handles the bulk of the business logic for
  caching, making cachers mostly "dumb content stores."
* The mechanism for invalidating cached entries (one of the harder
  parts about caching in general) is by varying the cache key when
  state changes. As such, cachers don't need to be concerned with
  cache invalidation.

Initially, we've hooked up support for caching "manifestdata" and
"filedata" commands. These are the simplest to cache, as they should
be immutable over time. Caching of commands related to changeset
data is a bit harder (because cache validation is impacted by
changes to bookmarks, phases, etc). This will be implemented later.
(Strictly speaking, censoring a file should invalidate caches. I've
added an inline TODO to track this edge case.)

To prove it works, this commit implements a test-only extension
providing in-memory caching backed by an lrucachedict. A new test
showing this extension behaving properly is added. FWIW, the
cacher is ~50 lines of code, demonstrating the relative ease with
which a cache can be added to a server.

While the test cacher is not suitable for production workloads, just
for kicks I performed a clone of just the changeset and manifest data
for the mozilla-unified repository. With a fully warmed cache (of just
the manifest data since changeset data is not cached), server-side
CPU usage dropped from ~73s to ~28s. That's pretty significant and
demonstrates the potential that response caching has on server
scalability!

Differential Revision: https://phab.mercurial-scm.org/D4773

# common patterns in test at can safely be replaced
from __future__ import absolute_import

import os

substitutions = [
    # list of possible compressions
    (br'(zstd,)?zlib,none,bzip2',
     br'$USUAL_COMPRESSIONS$'
    ),
    (br'=(zstd,)?zlib',
     br'=$BUNDLE2_COMPRESSIONS$'
    ),
    # capabilities sent through http
    (br'bundlecaps=HG20%2Cbundle2%3DHG20%250A'
     br'bookmarks%250A'
     br'changegroup%253D01%252C02%250A'
     br'digests%253Dmd5%252Csha1%252Csha512%250A'
     br'error%253Dabort%252Cunsupportedcontent%252Cpushraced%252Cpushkey%250A'
     br'hgtagsfnodes%250A'
     br'listkeys%250A'
     br'phases%253Dheads%250A'
     br'pushkey%250A'
     br'remote-changegroup%253Dhttp%252Chttps%250A'
     br'rev-branch-cache%250A'
     br'stream%253Dv2',
     # (the replacement patterns)
     br'$USUAL_BUNDLE_CAPS$'
    ),
    (br'bundlecaps=HG20%2Cbundle2%3DHG20%250A'
     br'bookmarks%250A'
     br'changegroup%253D01%252C02%250A'
     br'digests%253Dmd5%252Csha1%252Csha512%250A'
     br'error%253Dabort%252Cunsupportedcontent%252Cpushraced%252Cpushkey%250A'
     br'hgtagsfnodes%250A'
     br'listkeys%250A'
     br'phases%253Dheads%250A'
     br'pushkey%250A'
     br'remote-changegroup%253Dhttp%252Chttps',
     # (the replacement patterns)
     br'$USUAL_BUNDLE_CAPS_SERVER$'
     ),
    # bundle2 capabilities sent through ssh
    (br'bundle2=HG20%0A'
     br'bookmarks%0A'
     br'changegroup%3D01%2C02%0A'
     br'digests%3Dmd5%2Csha1%2Csha512%0A'
     br'error%3Dabort%2Cunsupportedcontent%2Cpushraced%2Cpushkey%0A'
     br'hgtagsfnodes%0A'
     br'listkeys%0A'
     br'phases%3Dheads%0A'
     br'pushkey%0A'
     br'remote-changegroup%3Dhttp%2Chttps%0A'
     br'rev-branch-cache%0A'
     br'stream%3Dv2',
     # (replacement patterns)
     br'$USUAL_BUNDLE2_CAPS$'
    ),
    # bundle2 capabilities advertised by the server
    (br'bundle2=HG20%0A'
     br'bookmarks%0A'
     br'changegroup%3D01%2C02%0A'
     br'digests%3Dmd5%2Csha1%2Csha512%0A'
     br'error%3Dabort%2Cunsupportedcontent%2Cpushraced%2Cpushkey%0A'
     br'hgtagsfnodes%0A'
     br'listkeys%0A'
     br'phases%3Dheads%0A'
     br'pushkey%0A'
     br'remote-changegroup%3Dhttp%2Chttps%0A'
     br'rev-branch-cache',
     # (replacement patterns)
     br'$USUAL_BUNDLE2_CAPS_SERVER$'
     ),
    (
     br'bundle2=HG20%0A'
     br'bookmarks%0A'
     br'changegroup%3D01%2C02%0A'
     br'digests%3Dmd5%2Csha1%2Csha512%0A'
     br'error%3Dabort%2Cunsupportedcontent%2Cpushraced%2Cpushkey%0A'
     br'hgtagsfnodes%0A'
     br'listkeys%0A'
     br'pushkey%0A'
     br'remote-changegroup%3Dhttp%2Chttps%0A'
     br'rev-branch-cache%0A'
     br'stream%3Dv2',
     # (replacement patterns)
     br'$USUAL_BUNDLE2_CAPS_NO_PHASES$'
    ),
    # HTTP access log dates
    (br' - - \[\d\d/.../2\d\d\d \d\d:\d\d:\d\d] "(GET|PUT|POST)',
     lambda m: br' - - [$LOGDATE$] "' + m.group(1)
    ),
    # HTTP error log dates
    (br' - - \[\d\d/.../2\d\d\d \d\d:\d\d:\d\d] (HG error:|Exception)',
     lambda m: br' - - [$ERRDATE$] ' + m.group(1)
    ),
    # HTTP header dates- RFC 1123
    (br'([Dd]ate): [A-Za-z]{3}, \d\d [A-Za-z]{3} \d{4} \d\d:\d\d:\d\d GMT',
     lambda m: br'%s: $HTTP_DATE$' % m.group(1)
    ),
    # LFS expiration value
    (br'"expires_at": "\d{4}-\d\d-\d\dT\d\d:\d\d:\d\dZ"',
     br'"expires_at": "$ISO_8601_DATE_TIME$"'
    ),
    # Windows has an extra '/' in the following lines that get globbed away:
    #   pushing to file:/*/$TESTTMP/r2 (glob)
    #   comparing with file:/*/$TESTTMP/r2 (glob)
    #   sub/maybelarge.dat: largefile 34..9c not available from
    #       file:/*/$TESTTMP/largefiles-repo (glob)
    (br'(.*file:/)/?(/\$TESTTMP.*)',
     lambda m: m.group(1) + b'*' + m.group(2) + b' (glob)'
    ),
]

# Various platform error strings, keyed on a common replacement string
_errors = {
    br'$ENOENT$': (
        # strerror()
        br'No such file or directory',

        # FormatMessage(ERROR_FILE_NOT_FOUND)
        br'The system cannot find the file specified',
    ),
    br'$ENOTDIR$': (
        # strerror()
        br'Not a directory',

        # FormatMessage(ERROR_PATH_NOT_FOUND)
        br'The system cannot find the path specified',
    ),
    br'$ECONNRESET$': (
        # strerror()
        br'Connection reset by peer',

        # FormatMessage(WSAECONNRESET)
        br'An existing connection was forcibly closed by the remote host',
    ),
    br'$EADDRINUSE$': (
        # strerror()
        br'Address already in use',

        # FormatMessage(WSAEADDRINUSE)
        br'Only one usage of each socket address'
        br' \(protocol/network address/port\) is normally permitted',
    ),
}

for replace, msgs in _errors.items():
    substitutions.extend((m, replace) for m in msgs)

# Output lines on Windows that can be autocorrected for '\' vs '/' path
# differences.
_winpathfixes = [
    # cloning subrepo s\ss from $TESTTMP/t/s/ss
    # cloning subrepo foo\bar from http://localhost:$HGPORT/foo/bar
    br'(?m)^cloning subrepo \S+\\.*',

    # pulling from $TESTTMP\issue1852a
    br'(?m)^pulling from \$TESTTMP\\.*',

    # pushing to $TESTTMP\a
    br'(?m)^pushing to \$TESTTMP\\.*',

    # pushing subrepo s\ss to $TESTTMP/t/s/ss
    br'(?m)^pushing subrepo \S+\\\S+ to.*',

    # moving d1\d11\a1 to d3/d11/a1
    br'(?m)^moving \S+\\.*',

    # d1\a: not recording move - dummy does not exist
    br'\S+\\\S+: not recording move .+',

    # reverting s\a
    br'(?m)^reverting (?!subrepo ).*\\.*',

    # saved backup bundle to
    #     $TESTTMP\test\.hg\strip-backup/443431ffac4f-2fc5398a-backup.hg
    br'(?m)^saved backup bundle to \$TESTTMP.*\.hg',

    # no changes made to subrepo s\ss since last push to ../tcc/s/ss
    br'(?m)^no changes made to subrepo \S+\\\S+ since.*',

    # changeset 5:9cc5aa7204f0: stuff/maybelarge.dat references missing
    #     $TESTTMP\largefiles-repo-hg\.hg\largefiles\76..38
    br'(?m)^changeset .* references (corrupted|missing) \$TESTTMP\\.*',

    # stuff/maybelarge.dat: largefile 76..38 not available from
    #     file:/*/$TESTTMP\largefiles-repo (glob)
    br'.*: largefile \S+ not available from file:/\*/.+',
]

if os.name == 'nt':
    substitutions.extend([(s, lambda match: match.group().replace(b'\\', b'/'))
                          for s in _winpathfixes])