##// END OF EJS Templates
ssh: avoid reading beyond the end of stream when using compression...
ssh: avoid reading beyond the end of stream when using compression Compressed streams can be used as part of getbundle. The normal read() operation of bufferedinputpipe will try to fulfill the request exactly and can deadlock if the server sends less as it is done. At the same time, the bundle2 logic will stop reading when it believes it has gotten all parts of the bundle, which can leave behind end of stream markers as used by bzip2 and zstd. To solve this, introduce a new optional unbufferedread interface and provided it in bufferedinputpipe and doublepipe. If there is buffered data left, it will be returned, otherwise it will issue a single read request and return whatever it obtains. Reorganize the decompression handlers to try harder to read until the end of stream, especially if the requested read can already be fulfilled. Check for end of stream is messy with Python 2, none of the standard compression modules properly exposes it. At least with zstd and bzip2, decompressing will remember EOS and fail for empty input after the EOS has been seen. For zlib, the only way to detect it with Python 2 is to duplicate the decompressobj and force some additional data into it. The common handler can be further optimized, but works as PoC. Differential Revision: https://phab.mercurial-scm.org/D3937

File last commit:

r37392:ac42e39b default
r38735:27391d74 default
Show More
narrowrevlog.py
80 lines | 3.0 KiB | text/x-python | PythonLexer
# narrowrevlog.py - revlog storing irrelevant nodes as "ellipsis" nodes
#
# Copyright 2017 Google, Inc.
#
# 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
from mercurial import (
revlog,
util,
)
def readtransform(self, text):
return text, False
def writetransform(self, text):
return text, False
def rawtransform(self, text):
return False
revlog.addflagprocessor(revlog.REVIDX_ELLIPSIS,
(readtransform, writetransform, rawtransform))
def setup():
# We just wanted to add the flag processor, which is done at module
# load time.
pass
def makenarrowfilelog(fl, narrowmatch):
class narrowfilelog(fl.__class__):
def renamed(self, node):
# Renames that come from outside the narrowspec are
# problematic at least for git-diffs, because we lack the
# base text for the rename. This logic was introduced in
# 3cd72b1 of narrowhg (authored by martinvonz, reviewed by
# adgar), but that revision doesn't have any additional
# commentary on what problems we can encounter.
m = super(narrowfilelog, self).renamed(node)
if m and not narrowmatch(m[0]):
return None
return m
def size(self, rev):
# We take advantage of the fact that remotefilelog
# lacks a node() method to just skip the
# rename-checking logic when on remotefilelog. This
# might be incorrect on other non-revlog-based storage
# engines, but for now this seems to be fine.
#
# TODO: when remotefilelog is in core, improve this to
# explicitly look for remotefilelog instead of cheating
# with a hasattr check.
if util.safehasattr(self, 'node'):
node = self.node(rev)
# Because renamed() is overridden above to
# sometimes return None even if there is metadata
# in the revlog, size can be incorrect for
# copies/renames, so we need to make sure we call
# the super class's implementation of renamed()
# for the purpose of size calculation.
if super(narrowfilelog, self).renamed(node):
return len(self.read(node))
return super(narrowfilelog, self).size(rev)
def cmp(self, node, text):
different = super(narrowfilelog, self).cmp(node, text)
if different:
# Similar to size() above, if the file was copied from
# a file outside the narrowspec, the super class's
# would have returned True because we tricked it into
# thinking that the file was not renamed.
if super(narrowfilelog, self).renamed(node):
t2 = self.read(node)
return t2 != text
return different
fl.__class__ = narrowfilelog