##// END OF EJS Templates
localrepo: iteratively derive local repository type...
localrepo: iteratively derive local repository type This commit implements the dynamic local repository type derivation that was explained in the recent commit bfeab472e3c0 "localrepo: create new function for instantiating a local repo object." Instead of a static localrepository class/type which must be customized after construction, we now dynamically construct a type by building up base classes/types to represent specific repository interfaces. Conceptually, the end state is similar to what was happening when various extensions would monkeypatch the __class__ of newly-constructed repo instances. However, the approach is inverted. Instead of making the instance then customizing it, we do the customization up front by influencing the behavior of the type then we instantiate that custom type. This approach gives us much more flexibility. For example, we can use completely separate classes for implementing different aspects of the repository. For example, we could have one class representing revlog-based file storage and another representing non-revlog based file storage. When then choose which implementation to use based on the presence of repo requirements. A concern with this approach is that it creates a lot more types and complexity and that complexity adds overhead. Yes, it is true that this approach will result in more types being created. Yes, this is more complicated than traditional "instantiate a static type." However, I believe the alternatives to supporting alternate storage backends are just as complicated. (Before I arrived at this solution, I had patches storing factory functions on local repo instances for e.g. constructing a file storage instance. We ended up having a handful of these. And this was logically identical to assigning custom methods. Since we were logically changing the type of the instance, I figured it would be better to just use specialized types instead of introducing levels of abstraction at run-time.) On the performance front, I don't believe that having N base classes has any significant performance overhead compared to just a single base class. Intuition says that Python will need to iterate the base classes to find an attribute. However, CPython caches method lookups: as long as the __class__ or MRO isn't changing, method attribute lookup should be constant time after first access. And non-method attributes are stored in __dict__, of which there is only 1 per object, so the number of base classes for __dict__ is irrelevant. Anyway, this commit splits up the monolithic completelocalrepository interface into sub-interfaces: 1 for file storage and 1 representing everything else. We've taught ``makelocalrepository()`` to call a series of factory functions which will produce types implementing specific interfaces. It then calls type() to create a new type from the built-up list of base types. This commit should be considered a start and not the end state. I suspect we'll hit a number of problems as we start to implement alternate storage backends: * Passing custom arguments to __init__ and setting custom attributes on __dict__. * Customizing the set of interfaces that are needed. e.g. the "readonly" intent could translate to not requesting an interface providing methods related to writing. * More ergonomic way for extensions to insert themselves so their callbacks aren't unconditionally called. * Wanting to modify vfs instances, other arguments passed to __init__. That being said, this code is usable in its current state and I'm convinced future commits will demonstrate the value in this approach. Differential Revision: https://phab.mercurial-scm.org/D4642

File last commit:

r39280:73cf21b2 default
r39800:e4e88157 default
Show More
hg.py
652 lines | 23.7 KiB | text/x-python | PythonLexer
# hg.py - hg backend for convert extension
#
# Copyright 2005-2009 Matt Mackall <mpm@selenic.com> and others
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
# Notes for hg->hg conversion:
#
# * Old versions of Mercurial didn't trim the whitespace from the ends
# of commit messages, but new versions do. Changesets created by
# those older versions, then converted, may thus have different
# hashes for changesets that are otherwise identical.
#
# * Using "--config convert.hg.saverev=true" will make the source
# identifier to be stored in the converted revision. This will cause
# the converted revision to have a different identity than the
# source.
from __future__ import absolute_import
import os
import re
import time
from mercurial.i18n import _
from mercurial import (
bookmarks,
context,
error,
exchange,
hg,
lock as lockmod,
merge as mergemod,
node as nodemod,
phases,
scmutil,
util,
)
from mercurial.utils import dateutil
stringio = util.stringio
from . import common
mapfile = common.mapfile
NoRepo = common.NoRepo
sha1re = re.compile(br'\b[0-9a-f]{12,40}\b')
class mercurial_sink(common.converter_sink):
def __init__(self, ui, repotype, path):
common.converter_sink.__init__(self, ui, repotype, path)
self.branchnames = ui.configbool('convert', 'hg.usebranchnames')
self.clonebranches = ui.configbool('convert', 'hg.clonebranches')
self.tagsbranch = ui.config('convert', 'hg.tagsbranch')
self.lastbranch = None
if os.path.isdir(path) and len(os.listdir(path)) > 0:
try:
self.repo = hg.repository(self.ui, path)
if not self.repo.local():
raise NoRepo(_('%s is not a local Mercurial repository')
% path)
except error.RepoError as err:
ui.traceback()
raise NoRepo(err.args[0])
else:
try:
ui.status(_('initializing destination %s repository\n') % path)
self.repo = hg.repository(self.ui, path, create=True)
if not self.repo.local():
raise NoRepo(_('%s is not a local Mercurial repository')
% path)
self.created.append(path)
except error.RepoError:
ui.traceback()
raise NoRepo(_("could not create hg repository %s as sink")
% path)
self.lock = None
self.wlock = None
self.filemapmode = False
self.subrevmaps = {}
def before(self):
self.ui.debug('run hg sink pre-conversion action\n')
self.wlock = self.repo.wlock()
self.lock = self.repo.lock()
def after(self):
self.ui.debug('run hg sink post-conversion action\n')
if self.lock:
self.lock.release()
if self.wlock:
self.wlock.release()
def revmapfile(self):
return self.repo.vfs.join("shamap")
def authorfile(self):
return self.repo.vfs.join("authormap")
def setbranch(self, branch, pbranches):
if not self.clonebranches:
return
setbranch = (branch != self.lastbranch)
self.lastbranch = branch
if not branch:
branch = 'default'
pbranches = [(b[0], b[1] and b[1] or 'default') for b in pbranches]
if pbranches:
pbranch = pbranches[0][1]
else:
pbranch = 'default'
branchpath = os.path.join(self.path, branch)
if setbranch:
self.after()
try:
self.repo = hg.repository(self.ui, branchpath)
except Exception:
self.repo = hg.repository(self.ui, branchpath, create=True)
self.before()
# pbranches may bring revisions from other branches (merge parents)
# Make sure we have them, or pull them.
missings = {}
for b in pbranches:
try:
self.repo.lookup(b[0])
except Exception:
missings.setdefault(b[1], []).append(b[0])
if missings:
self.after()
for pbranch, heads in sorted(missings.iteritems()):
pbranchpath = os.path.join(self.path, pbranch)
prepo = hg.peer(self.ui, {}, pbranchpath)
self.ui.note(_('pulling from %s into %s\n') % (pbranch, branch))
exchange.pull(self.repo, prepo,
[prepo.lookup(h) for h in heads])
self.before()
def _rewritetags(self, source, revmap, data):
fp = stringio()
for line in data.splitlines():
s = line.split(' ', 1)
if len(s) != 2:
self.ui.warn(_('invalid tag entry: "%s"\n') % line)
fp.write('%s\n' % line) # Bogus, but keep for hash stability
continue
revid = revmap.get(source.lookuprev(s[0]))
if not revid:
if s[0] == nodemod.nullhex:
revid = s[0]
else:
# missing, but keep for hash stability
self.ui.warn(_('missing tag entry: "%s"\n') % line)
fp.write('%s\n' % line)
continue
fp.write('%s %s\n' % (revid, s[1]))
return fp.getvalue()
def _rewritesubstate(self, source, data):
fp = stringio()
for line in data.splitlines():
s = line.split(' ', 1)
if len(s) != 2:
continue
revid = s[0]
subpath = s[1]
if revid != nodemod.nullhex:
revmap = self.subrevmaps.get(subpath)
if revmap is None:
revmap = mapfile(self.ui,
self.repo.wjoin(subpath, '.hg/shamap'))
self.subrevmaps[subpath] = revmap
# It is reasonable that one or more of the subrepos don't
# need to be converted, in which case they can be cloned
# into place instead of converted. Therefore, only warn
# once.
msg = _('no ".hgsubstate" updates will be made for "%s"\n')
if len(revmap) == 0:
sub = self.repo.wvfs.reljoin(subpath, '.hg')
if self.repo.wvfs.exists(sub):
self.ui.warn(msg % subpath)
newid = revmap.get(revid)
if not newid:
if len(revmap) > 0:
self.ui.warn(_("%s is missing from %s/.hg/shamap\n") %
(revid, subpath))
else:
revid = newid
fp.write('%s %s\n' % (revid, subpath))
return fp.getvalue()
def _calculatemergedfiles(self, source, p1ctx, p2ctx):
"""Calculates the files from p2 that we need to pull in when merging p1
and p2, given that the merge is coming from the given source.
This prevents us from losing files that only exist in the target p2 and
that don't come from the source repo (like if you're merging multiple
repositories together).
"""
anc = [p1ctx.ancestor(p2ctx)]
# Calculate what files are coming from p2
actions, diverge, rename = mergemod.calculateupdates(
self.repo, p1ctx, p2ctx, anc,
True, # branchmerge
True, # force
False, # acceptremote
False, # followcopies
)
for file, (action, info, msg) in actions.iteritems():
if source.targetfilebelongstosource(file):
# If the file belongs to the source repo, ignore the p2
# since it will be covered by the existing fileset.
continue
# If the file requires actual merging, abort. We don't have enough
# context to resolve merges correctly.
if action in ['m', 'dm', 'cd', 'dc']:
raise error.Abort(_("unable to convert merge commit "
"since target parents do not merge cleanly (file "
"%s, parents %s and %s)") % (file, p1ctx,
p2ctx))
elif action == 'k':
# 'keep' means nothing changed from p1
continue
else:
# Any other change means we want to take the p2 version
yield file
def putcommit(self, files, copies, parents, commit, source, revmap, full,
cleanp2):
files = dict(files)
def getfilectx(repo, memctx, f):
if p2ctx and f in p2files and f not in copies:
self.ui.debug('reusing %s from p2\n' % f)
try:
return p2ctx[f]
except error.ManifestLookupError:
# If the file doesn't exist in p2, then we're syncing a
# delete, so just return None.
return None
try:
v = files[f]
except KeyError:
return None
data, mode = source.getfile(f, v)
if data is None:
return None
if f == '.hgtags':
data = self._rewritetags(source, revmap, data)
if f == '.hgsubstate':
data = self._rewritesubstate(source, data)
return context.memfilectx(self.repo, memctx, f, data, 'l' in mode,
'x' in mode, copies.get(f))
pl = []
for p in parents:
if p not in pl:
pl.append(p)
parents = pl
nparents = len(parents)
if self.filemapmode and nparents == 1:
m1node = self.repo.changelog.read(nodemod.bin(parents[0]))[0]
parent = parents[0]
if len(parents) < 2:
parents.append(nodemod.nullid)
if len(parents) < 2:
parents.append(nodemod.nullid)
p2 = parents.pop(0)
text = commit.desc
sha1s = re.findall(sha1re, text)
for sha1 in sha1s:
oldrev = source.lookuprev(sha1)
newrev = revmap.get(oldrev)
if newrev is not None:
text = text.replace(sha1, newrev[:len(sha1)])
extra = commit.extra.copy()
sourcename = self.repo.ui.config('convert', 'hg.sourcename')
if sourcename:
extra['convert_source'] = sourcename
for label in ('source', 'transplant_source', 'rebase_source',
'intermediate-source'):
node = extra.get(label)
if node is None:
continue
# Only transplant stores its reference in binary
if label == 'transplant_source':
node = nodemod.hex(node)
newrev = revmap.get(node)
if newrev is not None:
if label == 'transplant_source':
newrev = nodemod.bin(newrev)
extra[label] = newrev
if self.branchnames and commit.branch:
extra['branch'] = commit.branch
if commit.rev and commit.saverev:
extra['convert_revision'] = commit.rev
while parents:
p1 = p2
p2 = parents.pop(0)
p1ctx = self.repo[p1]
p2ctx = None
if p2 != nodemod.nullid:
p2ctx = self.repo[p2]
fileset = set(files)
if full:
fileset.update(self.repo[p1])
fileset.update(self.repo[p2])
if p2ctx:
p2files = set(cleanp2)
for file in self._calculatemergedfiles(source, p1ctx, p2ctx):
p2files.add(file)
fileset.add(file)
ctx = context.memctx(self.repo, (p1, p2), text, fileset,
getfilectx, commit.author, commit.date, extra)
# We won't know if the conversion changes the node until after the
# commit, so copy the source's phase for now.
self.repo.ui.setconfig('phases', 'new-commit',
phases.phasenames[commit.phase], 'convert')
with self.repo.transaction("convert") as tr:
node = nodemod.hex(self.repo.commitctx(ctx))
# If the node value has changed, but the phase is lower than
# draft, set it back to draft since it hasn't been exposed
# anywhere.
if commit.rev != node:
ctx = self.repo[node]
if ctx.phase() < phases.draft:
phases.registernew(self.repo, tr, phases.draft,
[ctx.node()])
text = "(octopus merge fixup)\n"
p2 = node
if self.filemapmode and nparents == 1:
man = self.repo.manifestlog.getstorage(b'')
mnode = self.repo.changelog.read(nodemod.bin(p2))[0]
closed = 'close' in commit.extra
if not closed and not man.cmp(m1node, man.revision(mnode)):
self.ui.status(_("filtering out empty revision\n"))
self.repo.rollback(force=True)
return parent
return p2
def puttags(self, tags):
tagparent = self.repo.branchtip(self.tagsbranch, ignoremissing=True)
tagparent = tagparent or nodemod.nullid
oldlines = set()
for branch, heads in self.repo.branchmap().iteritems():
for h in heads:
if '.hgtags' in self.repo[h]:
oldlines.update(
set(self.repo[h]['.hgtags'].data().splitlines(True)))
oldlines = sorted(list(oldlines))
newlines = sorted([("%s %s\n" % (tags[tag], tag)) for tag in tags])
if newlines == oldlines:
return None, None
# if the old and new tags match, then there is nothing to update
oldtags = set()
newtags = set()
for line in oldlines:
s = line.strip().split(' ', 1)
if len(s) != 2:
continue
oldtags.add(s[1])
for line in newlines:
s = line.strip().split(' ', 1)
if len(s) != 2:
continue
if s[1] not in oldtags:
newtags.add(s[1].strip())
if not newtags:
return None, None
data = "".join(newlines)
def getfilectx(repo, memctx, f):
return context.memfilectx(repo, memctx, f, data, False, False, None)
self.ui.status(_("updating tags\n"))
date = "%d 0" % int(time.mktime(time.gmtime()))
extra = {'branch': self.tagsbranch}
ctx = context.memctx(self.repo, (tagparent, None), "update tags",
[".hgtags"], getfilectx, "convert-repo", date,
extra)
node = self.repo.commitctx(ctx)
return nodemod.hex(node), nodemod.hex(tagparent)
def setfilemapmode(self, active):
self.filemapmode = active
def putbookmarks(self, updatedbookmark):
if not len(updatedbookmark):
return
wlock = lock = tr = None
try:
wlock = self.repo.wlock()
lock = self.repo.lock()
tr = self.repo.transaction('bookmark')
self.ui.status(_("updating bookmarks\n"))
destmarks = self.repo._bookmarks
changes = [(bookmark, nodemod.bin(updatedbookmark[bookmark]))
for bookmark in updatedbookmark]
destmarks.applychanges(self.repo, tr, changes)
tr.close()
finally:
lockmod.release(lock, wlock, tr)
def hascommitfrommap(self, rev):
# the exact semantics of clonebranches is unclear so we can't say no
return rev in self.repo or self.clonebranches
def hascommitforsplicemap(self, rev):
if rev not in self.repo and self.clonebranches:
raise error.Abort(_('revision %s not found in destination '
'repository (lookups with clonebranches=true '
'are not implemented)') % rev)
return rev in self.repo
class mercurial_source(common.converter_source):
def __init__(self, ui, repotype, path, revs=None):
common.converter_source.__init__(self, ui, repotype, path, revs)
self.ignoreerrors = ui.configbool('convert', 'hg.ignoreerrors')
self.ignored = set()
self.saverev = ui.configbool('convert', 'hg.saverev')
try:
self.repo = hg.repository(self.ui, path)
# try to provoke an exception if this isn't really a hg
# repo, but some other bogus compatible-looking url
if not self.repo.local():
raise error.RepoError
except error.RepoError:
ui.traceback()
raise NoRepo(_("%s is not a local Mercurial repository") % path)
self.lastrev = None
self.lastctx = None
self._changescache = None, None
self.convertfp = None
# Restrict converted revisions to startrev descendants
startnode = ui.config('convert', 'hg.startrev')
hgrevs = ui.config('convert', 'hg.revs')
if hgrevs is None:
if startnode is not None:
try:
startnode = self.repo.lookup(startnode)
except error.RepoError:
raise error.Abort(_('%s is not a valid start revision')
% startnode)
startrev = self.repo.changelog.rev(startnode)
children = {startnode: 1}
for r in self.repo.changelog.descendants([startrev]):
children[self.repo.changelog.node(r)] = 1
self.keep = children.__contains__
else:
self.keep = util.always
if revs:
self._heads = [self.repo.lookup(r) for r in revs]
else:
self._heads = self.repo.heads()
else:
if revs or startnode is not None:
raise error.Abort(_('hg.revs cannot be combined with '
'hg.startrev or --rev'))
nodes = set()
parents = set()
for r in scmutil.revrange(self.repo, [hgrevs]):
ctx = self.repo[r]
nodes.add(ctx.node())
parents.update(p.node() for p in ctx.parents())
self.keep = nodes.__contains__
self._heads = nodes - parents
def _changectx(self, rev):
if self.lastrev != rev:
self.lastctx = self.repo[rev]
self.lastrev = rev
return self.lastctx
def _parents(self, ctx):
return [p for p in ctx.parents() if p and self.keep(p.node())]
def getheads(self):
return [nodemod.hex(h) for h in self._heads if self.keep(h)]
def getfile(self, name, rev):
try:
fctx = self._changectx(rev)[name]
return fctx.data(), fctx.flags()
except error.LookupError:
return None, None
def _changedfiles(self, ctx1, ctx2):
ma, r = [], []
maappend = ma.append
rappend = r.append
d = ctx1.manifest().diff(ctx2.manifest())
for f, ((node1, flag1), (node2, flag2)) in d.iteritems():
if node2 is None:
rappend(f)
else:
maappend(f)
return ma, r
def getchanges(self, rev, full):
ctx = self._changectx(rev)
parents = self._parents(ctx)
if full or not parents:
files = copyfiles = ctx.manifest()
if parents:
if self._changescache[0] == rev:
ma, r = self._changescache[1]
else:
ma, r = self._changedfiles(parents[0], ctx)
if not full:
files = ma + r
copyfiles = ma
# _getcopies() is also run for roots and before filtering so missing
# revlogs are detected early
copies = self._getcopies(ctx, parents, copyfiles)
cleanp2 = set()
if len(parents) == 2:
d = parents[1].manifest().diff(ctx.manifest(), clean=True)
for f, value in d.iteritems():
if value is None:
cleanp2.add(f)
changes = [(f, rev) for f in files if f not in self.ignored]
changes.sort()
return changes, copies, cleanp2
def _getcopies(self, ctx, parents, files):
copies = {}
for name in files:
if name in self.ignored:
continue
try:
copysource, _copynode = ctx.filectx(name).renamed()
if copysource in self.ignored:
continue
# Ignore copy sources not in parent revisions
if not any(copysource in p for p in parents):
continue
copies[name] = copysource
except TypeError:
pass
except error.LookupError as e:
if not self.ignoreerrors:
raise
self.ignored.add(name)
self.ui.warn(_('ignoring: %s\n') % e)
return copies
def getcommit(self, rev):
ctx = self._changectx(rev)
_parents = self._parents(ctx)
parents = [p.hex() for p in _parents]
optparents = [p.hex() for p in ctx.parents() if p and p not in _parents]
crev = rev
return common.commit(author=ctx.user(),
date=dateutil.datestr(ctx.date(),
'%Y-%m-%d %H:%M:%S %1%2'),
desc=ctx.description(),
rev=crev,
parents=parents,
optparents=optparents,
branch=ctx.branch(),
extra=ctx.extra(),
sortkey=ctx.rev(),
saverev=self.saverev,
phase=ctx.phase())
def gettags(self):
# This will get written to .hgtags, filter non global tags out.
tags = [t for t in self.repo.tagslist()
if self.repo.tagtype(t[0]) == 'global']
return dict([(name, nodemod.hex(node)) for name, node in tags
if self.keep(node)])
def getchangedfiles(self, rev, i):
ctx = self._changectx(rev)
parents = self._parents(ctx)
if not parents and i is None:
i = 0
ma, r = ctx.manifest().keys(), []
else:
i = i or 0
ma, r = self._changedfiles(parents[i], ctx)
ma, r = [[f for f in l if f not in self.ignored] for l in (ma, r)]
if i == 0:
self._changescache = (rev, (ma, r))
return ma + r
def converted(self, rev, destrev):
if self.convertfp is None:
self.convertfp = open(self.repo.vfs.join('shamap'), 'ab')
self.convertfp.write(util.tonativeeol('%s %s\n' % (destrev, rev)))
self.convertfp.flush()
def before(self):
self.ui.debug('run hg source pre-conversion action\n')
def after(self):
self.ui.debug('run hg source post-conversion action\n')
def hasnativeorder(self):
return True
def hasnativeclose(self):
return True
def lookuprev(self, rev):
try:
return nodemod.hex(self.repo.lookup(rev))
except (error.RepoError, error.LookupError):
return None
def getbookmarks(self):
return bookmarks.listbookmarks(self.repo)
def checkrevformat(self, revstr, mapname='splicemap'):
""" Mercurial, revision string is a 40 byte hex """
self.checkhexformat(revstr, mapname)