mercurial-scm/hg: mercurial/revlog.py comparison

comparison mercurial/revlog.py @ 39872:733db72f0f54

revlog: move revision verification out of verify File revision verification is performing low-level checks of file storage, namely that flags are appropriate and revision data can be resolved. Since these checks are somewhat revlog-specific and may not be appropriate for alternate storage backends, this commit moves those checks from verify.py to revlog.py. Because we're now emitting warnings/errors that apply to specific revisions, we taught the iverifyproblem interface to expose the problematic node and to report this node in verify output. This was necessary to prevent unwanted test changes. After this change, revlog.verifyintegrity() and file verify code in verify.py both iterate over revisions and resolve their fulltext. But they do so in separate loops. (verify.py needs to resolve fulltexts as part of calling renamed() - at least when using revlogs.) This should add overhead. But on the mozilla-unified repo: $ hg verify before: time: real 700.640 secs (user 585.520+0.000 sys 23.480+0.000) after: time: real 682.380 secs (user 570.370+0.000 sys 22.240+0.000) I'm not sure what's going on. Maybe avoiding the filelog attribute proxies shaved off enough time to offset the losses? Maybe fulltext resolution has less overhead than I thought? I've left a comment indicating the potential for optimization. But because it doesn't produce a performance regression on a large repository, I'm not going to worry about it. Differential Revision: https://phab.mercurial-scm.org/D4745

author	Gregory Szorc <gregory.szorc@gmail.com>
date	Mon, 24 Sep 2018 11:27:47 -0700
parents	14e500b58263
children	f8eb71f9e3bd

comparison

equal deleted inserted replaced

-:01c0f01b562b
+:733db72f0f54
 bin,
 hex,
 nullhex,
 nullid,
 nullrev,
+short,
 wdirfilenodeids,
 wdirhex,
 wdirid,
 wdirrev,
 )
 @interfaceutil.implementer(repository.iverifyproblem)
 @attr.s(frozen=True)
 class revlogproblem(object):
 warning = attr.ib(default=None)
 error = attr.ib(default=None)
+node = attr.ib(default=None)
 # index v0:
 #  4 bytes: offset
 #  4 bytes: compressed length
 #  4 bytes: base rev
 if version != state['expectedversion']:
 yield revlogproblem(
 warning=_("warning: '%s' uses revlog format %d; expected %d") %
 (self.indexfile, version, state['expectedversion']))
+state['skipread'] = set()
+for rev in self:
+node = self.node(rev)
+# Verify contents. 4 cases to care about:
+#
+#   common: the most common case
+#   rename: with a rename
+#   meta: file content starts with b'\1\n', the metadata
+#         header defined in filelog.py, but without a rename
+#   ext: content stored externally
+#
+# More formally, their differences are shown below:
+#
+#                       | common | rename | meta  | ext
+#  -------------------------------------------------------
+#   flags()             | 0      | 0      | 0     | not 0
+#   renamed()           | False  | True   | False | ?
+#   rawtext[0:2]=='\1\n'| False  | True   | True  | ?
+#
+# "rawtext" means the raw text stored in revlog data, which
+# could be retrieved by "revision(rev, raw=True)". "text"
+# mentioned below is "revision(rev, raw=False)".
+#
+# There are 3 different lengths stored physically:
+#  1. L1: rawsize, stored in revlog index
+#  2. L2: len(rawtext), stored in revlog data
+#  3. L3: len(text), stored in revlog data if flags==0, or
+#     possibly somewhere else if flags!=0
+#
+# L1 should be equal to L2. L3 could be different from them.
+# "text" may or may not affect commit hash depending on flag
+# processors (see revlog.addflagprocessor).
+#
+#              | common  | rename | meta  | ext
+# -------------------------------------------------
+#    rawsize() | L1      | L1     | L1    | L1
+#       size() | L1      | L2-LM  | L1(*) | L1 (?)
+# len(rawtext) | L2      | L2     | L2    | L2
+#    len(text) | L2      | L2     | L2    | L3
+#  len(read()) | L2      | L2-LM  | L2-LM | L3 (?)
+#
+# LM:  length of metadata, depending on rawtext
+# (*): not ideal, see comment in filelog.size
+# (?): could be "- len(meta)" if the resolved content has
+#      rename metadata
+#
+# Checks needed to be done:
+#  1. length check: L1 == L2, in all cases.
+#  2. hash check: depending on flag processor, we may need to
+#     use either "text" (external), or "rawtext" (in revlog).
+try:
+skipflags = state.get('skipflags', 0)
+if skipflags:
+skipflags &= self.flags(rev)
+if skipflags:
+state['skipread'].add(node)
+else:
+# Side-effect: read content and verify hash.
+self.revision(node)
+l1 = self.rawsize(rev)
+l2 = len(self.revision(node, raw=True))
+if l1 != l2:
+yield revlogproblem(
+error=_('unpacked size is %d, %d expected') % (l2, l1),
+node=node)
+except error.CensoredNodeError:
+if state['erroroncensored']:
+yield revlogproblem(error=_('censored file data'),
+node=node)
+state['skipread'].add(node)
+except Exception as e:
+yield revlogproblem(
+error=_('unpacking %s: %s') % (short(node), e),
+node=node)
+state['skipread'].add(node)
 def storageinfo(self, exclusivefiles=False, sharedfiles=False,
 revisionscount=False, trackedsize=False,
 storedsize=False):
 d = {}

Mercurial > public > mercurial-scm > hg

comparison mercurial/revlog.py @ 39872:733db72f0f54