##// END OF EJS Templates
wireproto: add streams to frame-based protocol...
wireproto: add streams to frame-based protocol Previously, the frame-based protocol was just a series of frames, with each frame associated with a request ID. In order to scale the protocol, we'll want to enable the use of compression. While it is possible to enable compression at the socket/pipe level, this has its disadvantages. The big one is it undermines the point of frames being standalone, atomic units that can be read and written: if you add compression above the framing protocol, you are back to having a stream-based protocol as opposed to something frame-based. So in order to preserve frames, compression needs to occur at the frame payload level. Compressing each frame's payload individually will limit compression ratios because the window size of the compressor will be limited by the max frame size, which is 32-64kb as currently defined. It will also add CPU overhead, as it is more efficient for compressors to operate on fewer, larger blocks of data than more, smaller blocks. So compressing each frame independently is out. This means we need to compress each frame's payload as if it is part of a larger stream. The simplest approach is to have 1 stream per connection. This could certainly work. However, it has disadvantages (documented below). We could also have 1 stream per RPC/command invocation. (This is the model HTTP/2 goes with.) This also has disadvantages. The main disadvantage to one global stream is that it has the very real potential to create CPU bottlenecks doing compression. Networks are only getting faster and the performance of single CPU cores has been relatively flat. Newer compression formats like zstandard offer better CPU cycle efficiency than predecessors like zlib. But it still all too common to saturate your CPU with compression overhead long before you saturate the network pipe. The main disadvantage with streams per request is that you can't reap the benefits of the compression context for multiple requests. For example, if you send 1000 RPC requests (or HTTP/2 requests for that matter), the response to each would have its own compression context. The overall size of the raw responses would be larger because compression contexts wouldn't be able to reference data from another request or response. The approach for streams as implemented in this commit is to support N streams per connection and for streams to potentially span requests and responses. As explained by the added internals docs, this facilitates servers and clients delegating independent streams and compression to independent threads / CPU cores. This helps alleviate the CPU bottleneck of compression. This design also allows compression contexts to be reused across requests/responses. This can result in improved compression ratios and less overhead for compressors and decompressors having to build new contexts. Another feature that was defined was the ability for individual frames within a stream to declare whether that individual frame's payload uses the content encoding (read: compression) defined by the stream. The idea here is that some servers may serve data from a combination of caches and dynamic resolution. Data coming from caches may be pre-compressed. We want to facilitate servers being able to essentially stream bytes from caches to the wire with minimal overhead. Being able to mix and match with frames are compressed within a stream enables these types of advanced server functionality. This commit defines the new streams mechanism. Basic code for supporting streams in frames has been added. But that code is seriously lacking and doesn't fully conform to the defined protocol. For example, we don't close any streams. And support for content encoding within streams is not yet implemented. The change was rather invasive and I didn't think it would be reasonable to implement the entire feature in a single commit. For the record, I would have loved to reuse an existing multiplexing protocol to build the new wire protocol on top of. However, I couldn't find a protocol that offers the performance and scaling characteristics that I desired. Namely, it should support multiple compression contexts to facilitate scaling out to multiple CPU cores and compression contexts should be able to live longer than single RPC requests. HTTP/2 *almost* fits the bill. But the semantics of HTTP message exchange state that streams can only live for a single request-response. We /could/ tunnel on top of HTTP/2 streams and frames with HEADER and DATA frames. But there's no guarantee that HTTP/2 libraries and proxies would allow us to use HTTP/2 streams and frames without the HTTP message exchange semantics defined in RFC 7540 Section 8. Other RPC protocols like gRPC tunnel are built on top of HTTP/2 and thus preserve its semantics of stream per RPC invocation. Even QUIC does this. We could attempt to invent a higher-level stream that spans HTTP/2 streams. But this would be violating HTTP/2 because there is no guarantee that HTTP/2 streams are routed to the same server. The best we can do - which is what this protocol does - is shoehorn all request and response data into a single HTTP message and create streams within. At that point, we've defined a Content-Type in HTTP parlance. It just so happens our media type can also work as a standalone, stream-based protocol, without leaning on HTTP or similar protocol. Differential Revision: https://phab.mercurial-scm.org/D2907

File last commit:

r37247:63144f33 default
r37304:9bfcbe4f default
Show More
templatefilters.py
439 lines | 13.3 KiB | text/x-python | PythonLexer
# templatefilters.py - common template expansion filters
#
# Copyright 2005-2008 Matt Mackall <mpm@selenic.com>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
from __future__ import absolute_import
import os
import re
import time
from .i18n import _
from . import (
encoding,
error,
node,
pycompat,
registrar,
templateutil,
url,
util,
)
from .utils import (
dateutil,
stringutil,
)
urlerr = util.urlerr
urlreq = util.urlreq
if pycompat.ispy3:
long = int
# filters are callables like:
# fn(obj)
# with:
# obj - object to be filtered (text, date, list and so on)
filters = {}
templatefilter = registrar.templatefilter(filters)
@templatefilter('addbreaks', intype=bytes)
def addbreaks(text):
"""Any text. Add an XHTML "<br />" tag before the end of
every line except the last.
"""
return text.replace('\n', '<br/>\n')
agescales = [("year", 3600 * 24 * 365, 'Y'),
("month", 3600 * 24 * 30, 'M'),
("week", 3600 * 24 * 7, 'W'),
("day", 3600 * 24, 'd'),
("hour", 3600, 'h'),
("minute", 60, 'm'),
("second", 1, 's')]
@templatefilter('age', intype=templateutil.date)
def age(date, abbrev=False):
"""Date. Returns a human-readable date/time difference between the
given date/time and the current date/time.
"""
def plural(t, c):
if c == 1:
return t
return t + "s"
def fmt(t, c, a):
if abbrev:
return "%d%s" % (c, a)
return "%d %s" % (c, plural(t, c))
now = time.time()
then = date[0]
future = False
if then > now:
future = True
delta = max(1, int(then - now))
if delta > agescales[0][1] * 30:
return 'in the distant future'
else:
delta = max(1, int(now - then))
if delta > agescales[0][1] * 2:
return dateutil.shortdate(date)
for t, s, a in agescales:
n = delta // s
if n >= 2 or s == 1:
if future:
return '%s from now' % fmt(t, n, a)
return '%s ago' % fmt(t, n, a)
@templatefilter('basename', intype=bytes)
def basename(path):
"""Any text. Treats the text as a path, and returns the last
component of the path after splitting by the path separator.
For example, "foo/bar/baz" becomes "baz" and "foo/bar//" becomes "".
"""
return os.path.basename(path)
@templatefilter('count')
def count(i):
"""List or text. Returns the length as an integer."""
try:
return len(i)
except TypeError:
raise error.ParseError(_('not countable'))
@templatefilter('dirname', intype=bytes)
def dirname(path):
"""Any text. Treats the text as a path, and strips the last
component of the path after splitting by the path separator.
"""
return os.path.dirname(path)
@templatefilter('domain', intype=bytes)
def domain(author):
"""Any text. Finds the first string that looks like an email
address, and extracts just the domain component. Example: ``User
<user@example.com>`` becomes ``example.com``.
"""
f = author.find('@')
if f == -1:
return ''
author = author[f + 1:]
f = author.find('>')
if f >= 0:
author = author[:f]
return author
@templatefilter('email', intype=bytes)
def email(text):
"""Any text. Extracts the first string that looks like an email
address. Example: ``User <user@example.com>`` becomes
``user@example.com``.
"""
return stringutil.email(text)
@templatefilter('escape', intype=bytes)
def escape(text):
"""Any text. Replaces the special XML/XHTML characters "&", "<"
and ">" with XML entities, and filters out NUL characters.
"""
return url.escape(text.replace('\0', ''), True)
para_re = None
space_re = None
def fill(text, width, initindent='', hangindent=''):
'''fill many paragraphs with optional indentation.'''
global para_re, space_re
if para_re is None:
para_re = re.compile('(\n\n|\n\\s*[-*]\\s*)', re.M)
space_re = re.compile(br' +')
def findparas():
start = 0
while True:
m = para_re.search(text, start)
if not m:
uctext = encoding.unifromlocal(text[start:])
w = len(uctext)
while 0 < w and uctext[w - 1].isspace():
w -= 1
yield (encoding.unitolocal(uctext[:w]),
encoding.unitolocal(uctext[w:]))
break
yield text[start:m.start(0)], m.group(1)
start = m.end(1)
return "".join([stringutil.wrap(space_re.sub(' ',
stringutil.wrap(para, width)),
width, initindent, hangindent) + rest
for para, rest in findparas()])
@templatefilter('fill68', intype=bytes)
def fill68(text):
"""Any text. Wraps the text to fit in 68 columns."""
return fill(text, 68)
@templatefilter('fill76', intype=bytes)
def fill76(text):
"""Any text. Wraps the text to fit in 76 columns."""
return fill(text, 76)
@templatefilter('firstline', intype=bytes)
def firstline(text):
"""Any text. Returns the first line of text."""
try:
return text.splitlines(True)[0].rstrip('\r\n')
except IndexError:
return ''
@templatefilter('hex', intype=bytes)
def hexfilter(text):
"""Any text. Convert a binary Mercurial node identifier into
its long hexadecimal representation.
"""
return node.hex(text)
@templatefilter('hgdate', intype=templateutil.date)
def hgdate(text):
"""Date. Returns the date as a pair of numbers: "1157407993
25200" (Unix timestamp, timezone offset).
"""
return "%d %d" % text
@templatefilter('isodate', intype=templateutil.date)
def isodate(text):
"""Date. Returns the date in ISO 8601 format: "2009-08-18 13:00
+0200".
"""
return dateutil.datestr(text, '%Y-%m-%d %H:%M %1%2')
@templatefilter('isodatesec', intype=templateutil.date)
def isodatesec(text):
"""Date. Returns the date in ISO 8601 format, including
seconds: "2009-08-18 13:00:13 +0200". See also the rfc3339date
filter.
"""
return dateutil.datestr(text, '%Y-%m-%d %H:%M:%S %1%2')
def indent(text, prefix):
'''indent each non-empty line of text after first with prefix.'''
lines = text.splitlines()
num_lines = len(lines)
endswithnewline = text[-1:] == '\n'
def indenter():
for i in xrange(num_lines):
l = lines[i]
if i and l.strip():
yield prefix
yield l
if i < num_lines - 1 or endswithnewline:
yield '\n'
return "".join(indenter())
@templatefilter('json')
def json(obj, paranoid=True):
if obj is None:
return 'null'
elif obj is False:
return 'false'
elif obj is True:
return 'true'
elif isinstance(obj, (int, long, float)):
return pycompat.bytestr(obj)
elif isinstance(obj, bytes):
return '"%s"' % encoding.jsonescape(obj, paranoid=paranoid)
elif isinstance(obj, str):
# This branch is unreachable on Python 2, because bytes == str
# and we'll return in the next-earlier block in the elif
# ladder. On Python 3, this helps us catch bugs before they
# hurt someone.
raise error.ProgrammingError(
'Mercurial only does output with bytes on Python 3: %r' % obj)
elif util.safehasattr(obj, 'keys'):
out = ['"%s": %s' % (encoding.jsonescape(k, paranoid=paranoid),
json(v, paranoid))
for k, v in sorted(obj.iteritems())]
return '{' + ', '.join(out) + '}'
elif util.safehasattr(obj, '__iter__'):
out = [json(i, paranoid) for i in obj]
return '[' + ', '.join(out) + ']'
raise error.ProgrammingError('cannot encode %r' % obj)
@templatefilter('lower', intype=bytes)
def lower(text):
"""Any text. Converts the text to lowercase."""
return encoding.lower(text)
@templatefilter('nonempty', intype=bytes)
def nonempty(text):
"""Any text. Returns '(none)' if the string is empty."""
return text or "(none)"
@templatefilter('obfuscate', intype=bytes)
def obfuscate(text):
"""Any text. Returns the input text rendered as a sequence of
XML entities.
"""
text = unicode(text, pycompat.sysstr(encoding.encoding), r'replace')
return ''.join(['&#%d;' % ord(c) for c in text])
@templatefilter('permissions', intype=bytes)
def permissions(flags):
if "l" in flags:
return "lrwxrwxrwx"
if "x" in flags:
return "-rwxr-xr-x"
return "-rw-r--r--"
@templatefilter('person', intype=bytes)
def person(author):
"""Any text. Returns the name before an email address,
interpreting it as per RFC 5322.
"""
return stringutil.person(author)
@templatefilter('revescape', intype=bytes)
def revescape(text):
"""Any text. Escapes all "special" characters, except @.
Forward slashes are escaped twice to prevent web servers from prematurely
unescaping them. For example, "@foo bar/baz" becomes "@foo%20bar%252Fbaz".
"""
return urlreq.quote(text, safe='/@').replace('/', '%252F')
@templatefilter('rfc3339date', intype=templateutil.date)
def rfc3339date(text):
"""Date. Returns a date using the Internet date format
specified in RFC 3339: "2009-08-18T13:00:13+02:00".
"""
return dateutil.datestr(text, "%Y-%m-%dT%H:%M:%S%1:%2")
@templatefilter('rfc822date', intype=templateutil.date)
def rfc822date(text):
"""Date. Returns a date using the same format used in email
headers: "Tue, 18 Aug 2009 13:00:13 +0200".
"""
return dateutil.datestr(text, "%a, %d %b %Y %H:%M:%S %1%2")
@templatefilter('short', intype=bytes)
def short(text):
"""Changeset hash. Returns the short form of a changeset hash,
i.e. a 12 hexadecimal digit string.
"""
return text[:12]
@templatefilter('shortbisect', intype=bytes)
def shortbisect(label):
"""Any text. Treats `label` as a bisection status, and
returns a single-character representing the status (G: good, B: bad,
S: skipped, U: untested, I: ignored). Returns single space if `text`
is not a valid bisection status.
"""
if label:
return label[0:1].upper()
return ' '
@templatefilter('shortdate', intype=templateutil.date)
def shortdate(text):
"""Date. Returns a date like "2006-09-18"."""
return dateutil.shortdate(text)
@templatefilter('slashpath', intype=bytes)
def slashpath(path):
"""Any text. Replaces the native path separator with slash."""
return util.pconvert(path)
@templatefilter('splitlines', intype=bytes)
def splitlines(text):
"""Any text. Split text into a list of lines."""
return templateutil.hybridlist(text.splitlines(), name='line')
@templatefilter('stringescape', intype=bytes)
def stringescape(text):
return stringutil.escapestr(text)
@templatefilter('stringify', intype=bytes)
def stringify(thing):
"""Any type. Turns the value into text by converting values into
text and concatenating them.
"""
return thing # coerced by the intype
@templatefilter('stripdir', intype=bytes)
def stripdir(text):
"""Treat the text as path and strip a directory level, if
possible. For example, "foo" and "foo/bar" becomes "foo".
"""
dir = os.path.dirname(text)
if dir == "":
return os.path.basename(text)
else:
return dir
@templatefilter('tabindent', intype=bytes)
def tabindent(text):
"""Any text. Returns the text, with every non-empty line
except the first starting with a tab character.
"""
return indent(text, '\t')
@templatefilter('upper', intype=bytes)
def upper(text):
"""Any text. Converts the text to uppercase."""
return encoding.upper(text)
@templatefilter('urlescape', intype=bytes)
def urlescape(text):
"""Any text. Escapes all "special" characters. For example,
"foo bar" becomes "foo%20bar".
"""
return urlreq.quote(text)
@templatefilter('user', intype=bytes)
def userfilter(text):
"""Any text. Returns a short representation of a user name or email
address."""
return stringutil.shortuser(text)
@templatefilter('emailuser', intype=bytes)
def emailuser(text):
"""Any text. Returns the user portion of an email address."""
return stringutil.emailuser(text)
@templatefilter('utf8', intype=bytes)
def utf8(text):
"""Any text. Converts from the local character encoding to UTF-8."""
return encoding.fromlocal(text)
@templatefilter('xmlescape', intype=bytes)
def xmlescape(text):
text = (text
.replace('&', '&amp;')
.replace('<', '&lt;')
.replace('>', '&gt;')
.replace('"', '&quot;')
.replace("'", '&#39;')) # &apos; invalid in HTML
return re.sub('[\x00-\x08\x0B\x0C\x0E-\x1F]', ' ', text)
def websub(text, websubtable):
""":websub: Any text. Only applies to hgweb. Applies the regular
expression replacements defined in the websub section.
"""
if websubtable:
for regexp, format in websubtable:
text = regexp.sub(format, text)
return text
def loadfilter(ui, extname, registrarobj):
"""Load template filter from specified registrarobj
"""
for name, func in registrarobj._table.iteritems():
filters[name] = func
# tell hggettext to extract docstrings from these functions:
i18nfunctions = filters.values()