##// 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:

r37102:f0b6fbea default
r37304:9bfcbe4f default
Show More
color.py
532 lines | 17.6 KiB | text/x-python | PythonLexer
# utility for color output for Mercurial commands
#
# Copyright (C) 2007 Kevin Christen <kevin.christen@gmail.com> and other
#
# 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 re
from .i18n import _
from . import (
encoding,
pycompat,
)
from .utils import (
stringutil,
)
try:
import curses
# Mapping from effect name to terminfo attribute name (or raw code) or
# color number. This will also force-load the curses module.
_baseterminfoparams = {
'none': (True, 'sgr0', ''),
'standout': (True, 'smso', ''),
'underline': (True, 'smul', ''),
'reverse': (True, 'rev', ''),
'inverse': (True, 'rev', ''),
'blink': (True, 'blink', ''),
'dim': (True, 'dim', ''),
'bold': (True, 'bold', ''),
'invisible': (True, 'invis', ''),
'italic': (True, 'sitm', ''),
'black': (False, curses.COLOR_BLACK, ''),
'red': (False, curses.COLOR_RED, ''),
'green': (False, curses.COLOR_GREEN, ''),
'yellow': (False, curses.COLOR_YELLOW, ''),
'blue': (False, curses.COLOR_BLUE, ''),
'magenta': (False, curses.COLOR_MAGENTA, ''),
'cyan': (False, curses.COLOR_CYAN, ''),
'white': (False, curses.COLOR_WHITE, ''),
}
except ImportError:
curses = None
_baseterminfoparams = {}
# start and stop parameters for effects
_effects = {
'none': 0,
'black': 30,
'red': 31,
'green': 32,
'yellow': 33,
'blue': 34,
'magenta': 35,
'cyan': 36,
'white': 37,
'bold': 1,
'italic': 3,
'underline': 4,
'inverse': 7,
'dim': 2,
'black_background': 40,
'red_background': 41,
'green_background': 42,
'yellow_background': 43,
'blue_background': 44,
'purple_background': 45,
'cyan_background': 46,
'white_background': 47,
}
_defaultstyles = {
'grep.match': 'red bold',
'grep.linenumber': 'green',
'grep.rev': 'green',
'grep.change': 'green',
'grep.sep': 'cyan',
'grep.filename': 'magenta',
'grep.user': 'magenta',
'grep.date': 'magenta',
'bookmarks.active': 'green',
'branches.active': 'none',
'branches.closed': 'black bold',
'branches.current': 'green',
'branches.inactive': 'none',
'diff.changed': 'white',
'diff.deleted': 'red',
'diff.deleted.highlight': 'red bold underline',
'diff.diffline': 'bold',
'diff.extended': 'cyan bold',
'diff.file_a': 'red bold',
'diff.file_b': 'green bold',
'diff.hunk': 'magenta',
'diff.inserted': 'green',
'diff.inserted.highlight': 'green bold underline',
'diff.tab': '',
'diff.trailingwhitespace': 'bold red_background',
'changeset.public': '',
'changeset.draft': '',
'changeset.secret': '',
'diffstat.deleted': 'red',
'diffstat.inserted': 'green',
'formatvariant.name.mismatchconfig': 'red',
'formatvariant.name.mismatchdefault': 'yellow',
'formatvariant.name.uptodate': 'green',
'formatvariant.repo.mismatchconfig': 'red',
'formatvariant.repo.mismatchdefault': 'yellow',
'formatvariant.repo.uptodate': 'green',
'formatvariant.config.special': 'yellow',
'formatvariant.config.default': 'green',
'formatvariant.default': '',
'histedit.remaining': 'red bold',
'ui.prompt': 'yellow',
'log.changeset': 'yellow',
'patchbomb.finalsummary': '',
'patchbomb.from': 'magenta',
'patchbomb.to': 'cyan',
'patchbomb.subject': 'green',
'patchbomb.diffstats': '',
'rebase.rebased': 'blue',
'rebase.remaining': 'red bold',
'resolve.resolved': 'green bold',
'resolve.unresolved': 'red bold',
'shelve.age': 'cyan',
'shelve.newest': 'green bold',
'shelve.name': 'blue bold',
'status.added': 'green bold',
'status.clean': 'none',
'status.copied': 'none',
'status.deleted': 'cyan bold underline',
'status.ignored': 'black bold',
'status.modified': 'blue bold',
'status.removed': 'red bold',
'status.unknown': 'magenta bold underline',
'tags.normal': 'green',
'tags.local': 'black bold',
}
def loadcolortable(ui, extname, colortable):
_defaultstyles.update(colortable)
def _terminfosetup(ui, mode, formatted):
'''Initialize terminfo data and the terminal if we're in terminfo mode.'''
# If we failed to load curses, we go ahead and return.
if curses is None:
return
# Otherwise, see what the config file says.
if mode not in ('auto', 'terminfo'):
return
ui._terminfoparams.update(_baseterminfoparams)
for key, val in ui.configitems('color'):
if key.startswith('color.'):
newval = (False, int(val), '')
ui._terminfoparams[key[6:]] = newval
elif key.startswith('terminfo.'):
newval = (True, '', val.replace('\\E', '\x1b'))
ui._terminfoparams[key[9:]] = newval
try:
curses.setupterm()
except curses.error as e:
ui._terminfoparams.clear()
return
for key, (b, e, c) in ui._terminfoparams.items():
if not b:
continue
if not c and not curses.tigetstr(e):
# Most terminals don't support dim, invis, etc, so don't be
# noisy and use ui.debug().
ui.debug("no terminfo entry for %s\n" % e)
del ui._terminfoparams[key]
if not curses.tigetstr('setaf') or not curses.tigetstr('setab'):
# Only warn about missing terminfo entries if we explicitly asked for
# terminfo mode and we're in a formatted terminal.
if mode == "terminfo" and formatted:
ui.warn(_("no terminfo entry for setab/setaf: reverting to "
"ECMA-48 color\n"))
ui._terminfoparams.clear()
def setup(ui):
"""configure color on a ui
That function both set the colormode for the ui object and read
the configuration looking for custom colors and effect definitions."""
mode = _modesetup(ui)
ui._colormode = mode
if mode and mode != 'debug':
configstyles(ui)
def _modesetup(ui):
if ui.plain('color'):
return None
config = ui.config('ui', 'color')
if config == 'debug':
return 'debug'
auto = (config == 'auto')
always = False
if not auto and stringutil.parsebool(config):
# We want the config to behave like a boolean, "on" is actually auto,
# but "always" value is treated as a special case to reduce confusion.
if ui.configsource('ui', 'color') == '--color' or config == 'always':
always = True
else:
auto = True
if not always and not auto:
return None
formatted = (always or (encoding.environ.get('TERM') != 'dumb'
and ui.formatted()))
mode = ui.config('color', 'mode')
# If pager is active, color.pagermode overrides color.mode.
if getattr(ui, 'pageractive', False):
mode = ui.config('color', 'pagermode', mode)
realmode = mode
if pycompat.iswindows:
from . import win32
term = encoding.environ.get('TERM')
# TERM won't be defined in a vanilla cmd.exe environment.
# UNIX-like environments on Windows such as Cygwin and MSYS will
# set TERM. They appear to make a best effort attempt at setting it
# to something appropriate. However, not all environments with TERM
# defined support ANSI.
ansienviron = term and 'xterm' in term
if mode == 'auto':
# Since "ansi" could result in terminal gibberish, we error on the
# side of selecting "win32". However, if w32effects is not defined,
# we almost certainly don't support "win32", so don't even try.
# w32ffects is not populated when stdout is redirected, so checking
# it first avoids win32 calls in a state known to error out.
if ansienviron or not w32effects or win32.enablevtmode():
realmode = 'ansi'
else:
realmode = 'win32'
# An empty w32effects is a clue that stdout is redirected, and thus
# cannot enable VT mode.
elif mode == 'ansi' and w32effects and not ansienviron:
win32.enablevtmode()
elif mode == 'auto':
realmode = 'ansi'
def modewarn():
# only warn if color.mode was explicitly set and we're in
# a formatted terminal
if mode == realmode and formatted:
ui.warn(_('warning: failed to set color mode to %s\n') % mode)
if realmode == 'win32':
ui._terminfoparams.clear()
if not w32effects:
modewarn()
return None
elif realmode == 'ansi':
ui._terminfoparams.clear()
elif realmode == 'terminfo':
_terminfosetup(ui, mode, formatted)
if not ui._terminfoparams:
## FIXME Shouldn't we return None in this case too?
modewarn()
realmode = 'ansi'
else:
return None
if always or (auto and formatted):
return realmode
return None
def configstyles(ui):
ui._styles.update(_defaultstyles)
for status, cfgeffects in ui.configitems('color'):
if '.' not in status or status.startswith(('color.', 'terminfo.')):
continue
cfgeffects = ui.configlist('color', status)
if cfgeffects:
good = []
for e in cfgeffects:
if valideffect(ui, e):
good.append(e)
else:
ui.warn(_("ignoring unknown color/effect %r "
"(configured in color.%s)\n")
% (e, status))
ui._styles[status] = ' '.join(good)
def _activeeffects(ui):
'''Return the effects map for the color mode set on the ui.'''
if ui._colormode == 'win32':
return w32effects
elif ui._colormode is not None:
return _effects
return {}
def valideffect(ui, effect):
'Determine if the effect is valid or not.'
return ((not ui._terminfoparams and effect in _activeeffects(ui))
or (effect in ui._terminfoparams
or effect[:-11] in ui._terminfoparams))
def _effect_str(ui, effect):
'''Helper function for render_effects().'''
bg = False
if effect.endswith('_background'):
bg = True
effect = effect[:-11]
try:
attr, val, termcode = ui._terminfoparams[effect]
except KeyError:
return ''
if attr:
if termcode:
return termcode
else:
return curses.tigetstr(val)
elif bg:
return curses.tparm(curses.tigetstr('setab'), val)
else:
return curses.tparm(curses.tigetstr('setaf'), val)
def _mergeeffects(text, start, stop):
"""Insert start sequence at every occurrence of stop sequence
>>> s = _mergeeffects(b'cyan', b'[C]', b'|')
>>> s = _mergeeffects(s + b'yellow', b'[Y]', b'|')
>>> s = _mergeeffects(b'ma' + s + b'genta', b'[M]', b'|')
>>> s = _mergeeffects(b'red' + s, b'[R]', b'|')
>>> s
'[R]red[M]ma[Y][C]cyan|[R][M][Y]yellow|[R][M]genta|'
"""
parts = []
for t in text.split(stop):
if not t:
continue
parts.extend([start, t, stop])
return ''.join(parts)
def _render_effects(ui, text, effects):
'Wrap text in commands to turn on each effect.'
if not text:
return text
if ui._terminfoparams:
start = ''.join(_effect_str(ui, effect)
for effect in ['none'] + effects.split())
stop = _effect_str(ui, 'none')
else:
activeeffects = _activeeffects(ui)
start = [pycompat.bytestr(activeeffects[e])
for e in ['none'] + effects.split()]
start = '\033[' + ';'.join(start) + 'm'
stop = '\033[' + pycompat.bytestr(activeeffects['none']) + 'm'
return _mergeeffects(text, start, stop)
_ansieffectre = re.compile(br'\x1b\[[0-9;]*m')
def stripeffects(text):
"""Strip ANSI control codes which could be inserted by colorlabel()"""
return _ansieffectre.sub('', text)
def colorlabel(ui, msg, label):
"""add color control code according to the mode"""
if ui._colormode == 'debug':
if label and msg:
if msg.endswith('\n'):
msg = "[%s|%s]\n" % (label, msg[:-1])
else:
msg = "[%s|%s]" % (label, msg)
elif ui._colormode is not None:
effects = []
for l in label.split():
s = ui._styles.get(l, '')
if s:
effects.append(s)
elif valideffect(ui, l):
effects.append(l)
effects = ' '.join(effects)
if effects:
msg = '\n'.join([_render_effects(ui, line, effects)
for line in msg.split('\n')])
return msg
w32effects = None
if pycompat.iswindows:
import ctypes
_kernel32 = ctypes.windll.kernel32
_WORD = ctypes.c_ushort
_INVALID_HANDLE_VALUE = -1
class _COORD(ctypes.Structure):
_fields_ = [('X', ctypes.c_short),
('Y', ctypes.c_short)]
class _SMALL_RECT(ctypes.Structure):
_fields_ = [('Left', ctypes.c_short),
('Top', ctypes.c_short),
('Right', ctypes.c_short),
('Bottom', ctypes.c_short)]
class _CONSOLE_SCREEN_BUFFER_INFO(ctypes.Structure):
_fields_ = [('dwSize', _COORD),
('dwCursorPosition', _COORD),
('wAttributes', _WORD),
('srWindow', _SMALL_RECT),
('dwMaximumWindowSize', _COORD)]
_STD_OUTPUT_HANDLE = 0xfffffff5 # (DWORD)-11
_STD_ERROR_HANDLE = 0xfffffff4 # (DWORD)-12
_FOREGROUND_BLUE = 0x0001
_FOREGROUND_GREEN = 0x0002
_FOREGROUND_RED = 0x0004
_FOREGROUND_INTENSITY = 0x0008
_BACKGROUND_BLUE = 0x0010
_BACKGROUND_GREEN = 0x0020
_BACKGROUND_RED = 0x0040
_BACKGROUND_INTENSITY = 0x0080
_COMMON_LVB_REVERSE_VIDEO = 0x4000
_COMMON_LVB_UNDERSCORE = 0x8000
# http://msdn.microsoft.com/en-us/library/ms682088%28VS.85%29.aspx
w32effects = {
'none': -1,
'black': 0,
'red': _FOREGROUND_RED,
'green': _FOREGROUND_GREEN,
'yellow': _FOREGROUND_RED | _FOREGROUND_GREEN,
'blue': _FOREGROUND_BLUE,
'magenta': _FOREGROUND_BLUE | _FOREGROUND_RED,
'cyan': _FOREGROUND_BLUE | _FOREGROUND_GREEN,
'white': _FOREGROUND_RED | _FOREGROUND_GREEN | _FOREGROUND_BLUE,
'bold': _FOREGROUND_INTENSITY,
'black_background': 0x100, # unused value > 0x0f
'red_background': _BACKGROUND_RED,
'green_background': _BACKGROUND_GREEN,
'yellow_background': _BACKGROUND_RED | _BACKGROUND_GREEN,
'blue_background': _BACKGROUND_BLUE,
'purple_background': _BACKGROUND_BLUE | _BACKGROUND_RED,
'cyan_background': _BACKGROUND_BLUE | _BACKGROUND_GREEN,
'white_background': (_BACKGROUND_RED | _BACKGROUND_GREEN |
_BACKGROUND_BLUE),
'bold_background': _BACKGROUND_INTENSITY,
'underline': _COMMON_LVB_UNDERSCORE, # double-byte charsets only
'inverse': _COMMON_LVB_REVERSE_VIDEO, # double-byte charsets only
}
passthrough = {_FOREGROUND_INTENSITY,
_BACKGROUND_INTENSITY,
_COMMON_LVB_UNDERSCORE,
_COMMON_LVB_REVERSE_VIDEO}
stdout = _kernel32.GetStdHandle(
_STD_OUTPUT_HANDLE) # don't close the handle returned
if stdout is None or stdout == _INVALID_HANDLE_VALUE:
w32effects = None
else:
csbi = _CONSOLE_SCREEN_BUFFER_INFO()
if not _kernel32.GetConsoleScreenBufferInfo(
stdout, ctypes.byref(csbi)):
# stdout may not support GetConsoleScreenBufferInfo()
# when called from subprocess or redirected
w32effects = None
else:
origattr = csbi.wAttributes
ansire = re.compile('\033\[([^m]*)m([^\033]*)(.*)',
re.MULTILINE | re.DOTALL)
def win32print(ui, writefunc, *msgs, **opts):
for text in msgs:
_win32print(ui, text, writefunc, **opts)
def _win32print(ui, text, writefunc, **opts):
label = opts.get(r'label', '')
attr = origattr
def mapcolor(val, attr):
if val == -1:
return origattr
elif val in passthrough:
return attr | val
elif val > 0x0f:
return (val & 0x70) | (attr & 0x8f)
else:
return (val & 0x07) | (attr & 0xf8)
# determine console attributes based on labels
for l in label.split():
style = ui._styles.get(l, '')
for effect in style.split():
try:
attr = mapcolor(w32effects[effect], attr)
except KeyError:
# w32effects could not have certain attributes so we skip
# them if not found
pass
# hack to ensure regexp finds data
if not text.startswith('\033['):
text = '\033[m' + text
# Look for ANSI-like codes embedded in text
m = re.match(ansire, text)
try:
while m:
for sattr in m.group(1).split(';'):
if sattr:
attr = mapcolor(int(sattr), attr)
ui.flush()
_kernel32.SetConsoleTextAttribute(stdout, attr)
writefunc(m.group(2), **opts)
m = re.match(ansire, m.group(3))
finally:
# Explicitly reset original attributes
ui.flush()
_kernel32.SetConsoleTextAttribute(stdout, origattr)