view mercurial/hgweb/wsgicgi.py @ 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.

# hgweb/wsgicgi.py - CGI->WSGI translator
#
# Copyright 2006 Eric Hopper <[e-mail]>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
#
# This was originally copied from the public domain code at
# http://www.python.org/dev/peps/pep-0333/#the-server-gateway-side

from __future__ import absolute_import

import os
import sys

from .. import (
    util,
)

from . import (
    common,
)

def launch(application):
    util.setbinary(sys.stdin)
    util.setbinary(sys.stdout)

    environ = dict(os.environ.iteritems())
    environ.setdefault('PATH_INFO', '')
    if environ.get('SERVER_SOFTWARE', '').startswith('Microsoft-IIS'):
        # IIS includes script_name in PATH_INFO
        scriptname = environ['SCRIPT_NAME']
        if environ['PATH_INFO'].startswith(scriptname):
            environ['PATH_INFO'] = environ['PATH_INFO'][len(scriptname):]

    stdin = sys.stdin
    if environ.get('HTTP_EXPECT', '').lower() == '100-continue':
        stdin = common.continuereader(stdin, sys.stdout.write)

    environ['wsgi.input'] = stdin
    environ['wsgi.errors'] = sys.stderr
    environ['wsgi.version'] = (1, 0)
    environ['wsgi.multithread'] = False
    environ['wsgi.multiprocess'] = True
    environ['wsgi.run_once'] = True

    if environ.get('HTTPS', 'off').lower() in ('on', '1', 'yes'):
        environ['wsgi.url_scheme'] = 'https'
    else:
        environ['wsgi.url_scheme'] = 'http'

    headers_set = []
    headers_sent = []
    out = sys.stdout

    def write(data):
        if not headers_set:
            raise AssertionError("write() before start_response()")

        elif not headers_sent:
            # Before the first output, send the stored headers
            status, response_headers = headers_sent[:] = headers_set
            out.write('Status: %s\r\n' % status)
            for header in response_headers:
                out.write('%s: %s\r\n' % header)
            out.write('\r\n')

        out.write(data)
        out.flush()

    def start_response(status, response_headers, exc_info=None):
        if exc_info:
            try:
                if headers_sent:
                    # Re-raise original exception if headers sent
                    raise exc_info[0](exc_info[1], exc_info[2])
            finally:
                exc_info = None     # avoid dangling circular ref
        elif headers_set:
            raise AssertionError("Headers already set!")

        headers_set[:] = [status, response_headers]
        return write

    content = application(environ, start_response)
    try:
        for chunk in content:
            write(chunk)
        if not headers_sent:
            write('')   # send headers now if body was empty
    finally:
        getattr(content, 'close', lambda : None)()