2005-08-28 01:21:25 +04:00
|
|
|
# manifest.py - manifest revision class for mercurial
|
|
|
|
#
|
2007-06-19 10:51:34 +04:00
|
|
|
# Copyright 2005-2007 Matt Mackall <mpm@selenic.com>
|
2005-08-28 01:21:25 +04:00
|
|
|
#
|
2009-04-26 03:08:54 +04:00
|
|
|
# This software may be used and distributed according to the terms of the
|
2010-01-20 07:20:08 +03:00
|
|
|
# GNU General Public License version 2 or any later version.
|
2005-08-28 01:21:25 +04:00
|
|
|
|
2006-12-15 05:25:19 +03:00
|
|
|
from i18n import _
|
manifest: for diff(), only iterate over files, not flags
From manifest.diff(), we return a dict from filename to pairs of pairs
of file nodeids and flags (values of the form ((n1,n2),(fl1,fl2))). To
create this dict, we currently generate one dict for files (with
(n1,n2) values) and one for flags (with (fl1,fl2) values) and then
join these dicts. Missing files are represented by None and missing
flags by '', but due to the dict joining, the inner pairs themselves
can also be None. The only caller, merge.manifestmerge(), then unpacks
these values while checking for None values.
By inlining the calls to dicthelpers and simplifying it to only
iterate over files (ignoring flags-only differences), we can simplify
life for our caller.
2014-10-15 09:48:44 +04:00
|
|
|
import mdiff, parsers, error, revlog, util
|
2009-04-28 19:40:46 +04:00
|
|
|
import array, struct
|
2005-08-28 01:21:25 +04:00
|
|
|
|
2015-03-07 19:42:05 +03:00
|
|
|
|
2015-03-07 20:04:39 +03:00
|
|
|
class _lazymanifest(dict):
|
|
|
|
"""This is the pure implementation of lazymanifest.
|
|
|
|
|
|
|
|
It has not been optimized *at all* and is not lazy.
|
|
|
|
"""
|
|
|
|
|
|
|
|
def __init__(self, data):
|
|
|
|
# This init method does a little bit of excessive-looking
|
|
|
|
# precondition checking. This is so that the behavior of this
|
|
|
|
# class exactly matches its C counterpart to try and help
|
|
|
|
# prevent surprise breakage for anyone that develops against
|
|
|
|
# the pure version.
|
|
|
|
if data and data[-1] != '\n':
|
|
|
|
raise ValueError('Manifest did not end in a newline.')
|
|
|
|
dict.__init__(self)
|
|
|
|
prev = None
|
|
|
|
for l in data.splitlines():
|
|
|
|
if prev is not None and prev > l:
|
|
|
|
raise ValueError('Manifest lines not in sorted order.')
|
|
|
|
prev = l
|
|
|
|
f, n = l.split('\0')
|
|
|
|
if len(n) > 40:
|
|
|
|
self[f] = revlog.bin(n[:40]), n[40:]
|
|
|
|
else:
|
|
|
|
self[f] = revlog.bin(n), ''
|
2015-03-07 19:43:12 +03:00
|
|
|
|
2014-12-12 21:40:44 +03:00
|
|
|
def __setitem__(self, k, v):
|
2015-03-07 20:04:39 +03:00
|
|
|
node, flag = v
|
|
|
|
assert node is not None
|
|
|
|
if len(node) > 21:
|
|
|
|
node = node[:21] # match c implementation behavior
|
|
|
|
dict.__setitem__(self, k, (node, flag))
|
|
|
|
|
|
|
|
def __iter__(self):
|
|
|
|
return ((f, e[0], e[1]) for f, e in sorted(self.iteritems()))
|
|
|
|
|
2006-07-16 11:59:20 +04:00
|
|
|
def copy(self):
|
2015-03-07 20:04:39 +03:00
|
|
|
c = _lazymanifest('')
|
|
|
|
c.update(self)
|
|
|
|
return c
|
|
|
|
|
|
|
|
def diff(self, m2, clean=False):
|
|
|
|
'''Finds changes between the current manifest and m2.'''
|
|
|
|
diff = {}
|
|
|
|
|
|
|
|
for fn, e1 in self.iteritems():
|
|
|
|
if fn not in m2:
|
|
|
|
diff[fn] = e1, (None, '')
|
|
|
|
else:
|
|
|
|
e2 = m2[fn]
|
|
|
|
if e1 != e2:
|
|
|
|
diff[fn] = e1, e2
|
|
|
|
elif clean:
|
|
|
|
diff[fn] = None
|
|
|
|
|
|
|
|
for fn, e2 in m2.iteritems():
|
|
|
|
if fn not in self:
|
|
|
|
diff[fn] = (None, ''), e2
|
|
|
|
|
|
|
|
return diff
|
|
|
|
|
|
|
|
def filtercopy(self, filterfn):
|
|
|
|
c = _lazymanifest('')
|
|
|
|
for f, n, fl in self:
|
|
|
|
if filterfn(f):
|
|
|
|
c[f] = n, fl
|
|
|
|
return c
|
|
|
|
|
|
|
|
def text(self):
|
|
|
|
"""Get the full data of this manifest as a bytestring."""
|
|
|
|
fl = sorted(self)
|
|
|
|
|
|
|
|
_hex = revlog.hex
|
|
|
|
# if this is changed to support newlines in filenames,
|
|
|
|
# be sure to check the templates/ dir again (especially *-raw.tmpl)
|
|
|
|
return ''.join("%s\0%s%s\n" % (
|
|
|
|
f, _hex(n[:20]), flag) for f, n, flag in fl)
|
|
|
|
|
2015-03-07 05:29:47 +03:00
|
|
|
try:
|
|
|
|
_lazymanifest = parsers.lazymanifest
|
|
|
|
except AttributeError:
|
|
|
|
pass
|
|
|
|
|
2015-03-07 20:04:39 +03:00
|
|
|
class manifestdict(object):
|
|
|
|
def __init__(self, data=''):
|
|
|
|
self._lm = _lazymanifest(data)
|
|
|
|
|
|
|
|
def __getitem__(self, key):
|
|
|
|
return self._lm[key][0]
|
|
|
|
|
|
|
|
def __len__(self):
|
|
|
|
return len(self._lm)
|
|
|
|
|
|
|
|
def __setitem__(self, key, node):
|
|
|
|
self._lm[key] = node, self.flags(key, '')
|
|
|
|
|
|
|
|
def __contains__(self, key):
|
|
|
|
return key in self._lm
|
|
|
|
|
|
|
|
def __delitem__(self, key):
|
|
|
|
del self._lm[key]
|
|
|
|
|
|
|
|
def keys(self):
|
|
|
|
return [x[0] for x in self._lm]
|
|
|
|
|
|
|
|
def iterkeys(self):
|
|
|
|
return iter(self.keys())
|
|
|
|
|
2014-07-13 04:57:25 +04:00
|
|
|
def intersectfiles(self, files):
|
2015-03-07 20:04:39 +03:00
|
|
|
'''make a new lazymanifest with the intersection of self with files
|
2014-07-13 04:57:25 +04:00
|
|
|
|
|
|
|
The algorithm assumes that files is much smaller than self.'''
|
|
|
|
ret = manifestdict()
|
2015-03-07 20:04:39 +03:00
|
|
|
lm = self._lm
|
2014-07-13 04:57:25 +04:00
|
|
|
for fn in files:
|
2015-03-07 20:04:39 +03:00
|
|
|
if fn in lm:
|
|
|
|
ret._lm[fn] = self._lm[fn]
|
2014-07-13 04:57:25 +04:00
|
|
|
return ret
|
2014-10-15 04:09:16 +04:00
|
|
|
|
2015-02-28 00:57:37 +03:00
|
|
|
def filesnotin(self, m2):
|
|
|
|
'''Set of files in this manifest that are not in the other'''
|
|
|
|
files = set(self.iterkeys())
|
|
|
|
files.difference_update(m2.iterkeys())
|
|
|
|
return files
|
|
|
|
|
2014-10-23 08:38:30 +04:00
|
|
|
def matches(self, match):
|
|
|
|
'''generate a new manifest filtered by the match argument'''
|
|
|
|
if match.always():
|
|
|
|
return self.copy()
|
|
|
|
|
|
|
|
files = match.files()
|
|
|
|
if (match.matchfn == match.exact or
|
|
|
|
(not match.anypats() and util.all(fn in self for fn in files))):
|
|
|
|
return self.intersectfiles(files)
|
|
|
|
|
2015-03-07 20:04:39 +03:00
|
|
|
lm = manifestdict('')
|
|
|
|
lm._lm = self._lm.filtercopy(match)
|
|
|
|
return lm
|
2014-10-23 08:38:30 +04:00
|
|
|
|
2014-12-16 00:04:28 +03:00
|
|
|
def diff(self, m2, clean=False):
|
|
|
|
'''Finds changes between the current manifest and m2.
|
|
|
|
|
|
|
|
Args:
|
|
|
|
m2: the manifest to which this manifest should be compared.
|
|
|
|
clean: if true, include files unchanged between these manifests
|
|
|
|
with a None value in the returned dictionary.
|
|
|
|
|
|
|
|
The result is returned as a dict with filename as key and
|
|
|
|
values of the form ((n1,fl1),(n2,fl2)), where n1/n2 is the
|
|
|
|
nodeid in the current/other manifest and fl1/fl2 is the flag
|
|
|
|
in the current/other manifest. Where the file does not exist,
|
|
|
|
the nodeid will be None and the flags will be the empty
|
|
|
|
string.
|
|
|
|
'''
|
2015-03-07 20:04:39 +03:00
|
|
|
return self._lm.diff(m2._lm, clean)
|
manifest: for diff(), only iterate over files, not flags
From manifest.diff(), we return a dict from filename to pairs of pairs
of file nodeids and flags (values of the form ((n1,n2),(fl1,fl2))). To
create this dict, we currently generate one dict for files (with
(n1,n2) values) and one for flags (with (fl1,fl2) values) and then
join these dicts. Missing files are represented by None and missing
flags by '', but due to the dict joining, the inner pairs themselves
can also be None. The only caller, merge.manifestmerge(), then unpacks
these values while checking for None values.
By inlining the calls to dicthelpers and simplifying it to only
iterate over files (ignoring flags-only differences), we can simplify
life for our caller.
2014-10-15 09:48:44 +04:00
|
|
|
|
2015-03-07 20:04:39 +03:00
|
|
|
def setflag(self, key, flag):
|
|
|
|
self._lm[key] = self[key], flag
|
manifest: for diff(), only iterate over files, not flags
From manifest.diff(), we return a dict from filename to pairs of pairs
of file nodeids and flags (values of the form ((n1,n2),(fl1,fl2))). To
create this dict, we currently generate one dict for files (with
(n1,n2) values) and one for flags (with (fl1,fl2) values) and then
join these dicts. Missing files are represented by None and missing
flags by '', but due to the dict joining, the inner pairs themselves
can also be None. The only caller, merge.manifestmerge(), then unpacks
these values while checking for None values.
By inlining the calls to dicthelpers and simplifying it to only
iterate over files (ignoring flags-only differences), we can simplify
life for our caller.
2014-10-15 09:48:44 +04:00
|
|
|
|
2015-03-07 20:04:39 +03:00
|
|
|
def get(self, key, default=None):
|
|
|
|
try:
|
|
|
|
return self._lm[key][0]
|
|
|
|
except KeyError:
|
|
|
|
return default
|
2006-07-16 11:59:20 +04:00
|
|
|
|
2015-03-07 20:04:39 +03:00
|
|
|
def flags(self, key, default=''):
|
|
|
|
try:
|
|
|
|
return self._lm[key][1]
|
|
|
|
except KeyError:
|
|
|
|
return default
|
2014-10-08 22:47:30 +04:00
|
|
|
|
2015-03-07 20:04:39 +03:00
|
|
|
def __iter__(self):
|
|
|
|
return (x[0] for x in self._lm)
|
|
|
|
|
|
|
|
def copy(self):
|
|
|
|
c = manifestdict('')
|
|
|
|
c._lm = self._lm.copy()
|
|
|
|
return c
|
|
|
|
|
|
|
|
def iteritems(self):
|
|
|
|
return (x[:2] for x in self._lm)
|
|
|
|
|
|
|
|
def text(self):
|
|
|
|
return self._lm.text()
|
2014-08-07 17:46:05 +04:00
|
|
|
|
2014-10-08 23:20:14 +04:00
|
|
|
def fastdelta(self, base, changes):
|
|
|
|
"""Given a base manifest text as an array.array and a list of changes
|
|
|
|
relative to that text, compute a delta that can be used by revlog.
|
|
|
|
"""
|
|
|
|
delta = []
|
|
|
|
dstart = None
|
|
|
|
dend = None
|
|
|
|
dline = [""]
|
|
|
|
start = 0
|
|
|
|
# zero copy representation of base as a buffer
|
|
|
|
addbuf = util.buffer(base)
|
|
|
|
|
|
|
|
# start with a readonly loop that finds the offset of
|
|
|
|
# each line and creates the deltas
|
|
|
|
for f, todelete in changes:
|
|
|
|
# bs will either be the index of the item or the insert point
|
|
|
|
start, end = _msearch(addbuf, f, start)
|
|
|
|
if not todelete:
|
2015-03-07 20:04:39 +03:00
|
|
|
h, fl = self._lm[f]
|
|
|
|
l = "%s\0%s%s\n" % (f, revlog.hex(h), fl)
|
2014-10-08 23:20:14 +04:00
|
|
|
else:
|
|
|
|
if start == end:
|
|
|
|
# item we want to delete was not found, error out
|
|
|
|
raise AssertionError(
|
|
|
|
_("failed to remove %s from manifest") % f)
|
|
|
|
l = ""
|
|
|
|
if dstart is not None and dstart <= start and dend >= start:
|
|
|
|
if dend < end:
|
|
|
|
dend = end
|
|
|
|
if l:
|
|
|
|
dline.append(l)
|
|
|
|
else:
|
|
|
|
if dstart is not None:
|
|
|
|
delta.append([dstart, dend, "".join(dline)])
|
|
|
|
dstart = start
|
|
|
|
dend = end
|
|
|
|
dline = [l]
|
|
|
|
|
|
|
|
if dstart is not None:
|
|
|
|
delta.append([dstart, dend, "".join(dline)])
|
|
|
|
# apply the delta to the base, and get a delta for addrevision
|
|
|
|
deltatext, arraytext = _addlistdelta(base, delta)
|
|
|
|
return arraytext, deltatext
|
|
|
|
|
2014-10-08 23:21:59 +04:00
|
|
|
def _msearch(m, s, lo=0, hi=None):
|
|
|
|
'''return a tuple (start, end) that says where to find s within m.
|
|
|
|
|
|
|
|
If the string is found m[start:end] are the line containing
|
|
|
|
that string. If start == end the string was not found and
|
|
|
|
they indicate the proper sorted insertion point.
|
|
|
|
|
|
|
|
m should be a buffer or a string
|
|
|
|
s is a string'''
|
|
|
|
def advance(i, c):
|
|
|
|
while i < lenm and m[i] != c:
|
|
|
|
i += 1
|
|
|
|
return i
|
|
|
|
if not s:
|
|
|
|
return (lo, lo)
|
|
|
|
lenm = len(m)
|
|
|
|
if not hi:
|
|
|
|
hi = lenm
|
|
|
|
while lo < hi:
|
|
|
|
mid = (lo + hi) // 2
|
|
|
|
start = mid
|
|
|
|
while start > 0 and m[start - 1] != '\n':
|
|
|
|
start -= 1
|
|
|
|
end = advance(start, '\0')
|
|
|
|
if m[start:end] < s:
|
|
|
|
# we know that after the null there are 40 bytes of sha1
|
|
|
|
# this translates to the bisect lo = mid + 1
|
|
|
|
lo = advance(end + 40, '\n') + 1
|
|
|
|
else:
|
|
|
|
# this translates to the bisect hi = mid
|
|
|
|
hi = start
|
|
|
|
end = advance(lo, '\0')
|
|
|
|
found = m[lo:end]
|
|
|
|
if s == found:
|
|
|
|
# we know that after the null there are 40 bytes of sha1
|
|
|
|
end = advance(end + 40, '\n')
|
|
|
|
return (lo, end + 1)
|
|
|
|
else:
|
|
|
|
return (lo, lo)
|
|
|
|
|
2014-09-11 18:14:34 +04:00
|
|
|
def _checkforbidden(l):
|
2014-08-07 17:46:05 +04:00
|
|
|
"""Check filenames for illegal characters."""
|
|
|
|
for f in l:
|
|
|
|
if '\n' in f or '\r' in f:
|
|
|
|
raise error.RevlogError(
|
|
|
|
_("'\\n' and '\\r' disallowed in filenames: %r") % f)
|
|
|
|
|
|
|
|
|
2014-08-07 20:47:20 +04:00
|
|
|
# apply the changes collected during the bisect loop to our addlist
|
|
|
|
# return a delta suitable for addrevision
|
2014-09-11 18:14:34 +04:00
|
|
|
def _addlistdelta(addlist, x):
|
2014-08-07 20:47:20 +04:00
|
|
|
# for large addlist arrays, building a new array is cheaper
|
|
|
|
# than repeatedly modifying the existing one
|
|
|
|
currentposition = 0
|
|
|
|
newaddlist = array.array('c')
|
|
|
|
|
|
|
|
for start, end, content in x:
|
|
|
|
newaddlist += addlist[currentposition:start]
|
|
|
|
if content:
|
|
|
|
newaddlist += array.array('c', content)
|
|
|
|
|
|
|
|
currentposition = end
|
|
|
|
|
|
|
|
newaddlist += addlist[currentposition:]
|
|
|
|
|
|
|
|
deltatext = "".join(struct.pack(">lll", start, end, len(content))
|
|
|
|
+ content for start, end, content in x)
|
|
|
|
return deltatext, newaddlist
|
|
|
|
|
2009-01-12 07:55:36 +03:00
|
|
|
class manifest(revlog.revlog):
|
2007-03-23 03:52:38 +03:00
|
|
|
def __init__(self, opener):
|
2015-01-24 04:06:03 +03:00
|
|
|
# During normal operations, we expect to deal with not more than four
|
|
|
|
# revs at a time (such as during commit --amend). When rebasing large
|
|
|
|
# stacks of commits, the number can go up, hence the config knob below.
|
|
|
|
cachesize = 4
|
|
|
|
opts = getattr(opener, 'options', None)
|
|
|
|
if opts is not None:
|
|
|
|
cachesize = opts.get('manifestcachesize', cachesize)
|
|
|
|
self._mancache = util.lrucachedict(cachesize)
|
2009-01-12 07:55:36 +03:00
|
|
|
revlog.revlog.__init__(self, opener, "00manifest.i")
|
2005-08-28 01:21:25 +04:00
|
|
|
|
2006-09-30 00:00:54 +04:00
|
|
|
def readdelta(self, node):
|
2008-11-13 00:32:16 +03:00
|
|
|
r = self.rev(node)
|
2015-03-07 19:43:12 +03:00
|
|
|
d = mdiff.patchtext(self.revdiff(self.deltaparent(r), r))
|
|
|
|
return manifestdict(d)
|
2006-10-01 21:26:33 +04:00
|
|
|
|
2011-03-21 03:43:28 +03:00
|
|
|
def readfast(self, node):
|
|
|
|
'''use the faster of readdelta or read'''
|
|
|
|
r = self.rev(node)
|
2011-05-05 20:05:24 +04:00
|
|
|
deltaparent = self.deltaparent(r)
|
|
|
|
if deltaparent != revlog.nullrev and deltaparent in self.parentrevs(r):
|
2011-03-21 03:43:28 +03:00
|
|
|
return self.readdelta(node)
|
|
|
|
return self.read(node)
|
|
|
|
|
2005-08-28 01:21:25 +04:00
|
|
|
def read(self, node):
|
2009-01-12 07:55:36 +03:00
|
|
|
if node == revlog.nullid:
|
|
|
|
return manifestdict() # don't upset local cache
|
2013-02-09 19:43:02 +04:00
|
|
|
if node in self._mancache:
|
|
|
|
return self._mancache[node][0]
|
2005-08-28 01:21:25 +04:00
|
|
|
text = self.revision(node)
|
2009-09-02 23:05:01 +04:00
|
|
|
arraytext = array.array('c', text)
|
2015-03-07 19:43:12 +03:00
|
|
|
m = manifestdict(text)
|
2015-02-24 20:08:54 +03:00
|
|
|
self._mancache[node] = (m, arraytext)
|
|
|
|
return m
|
2005-08-28 01:21:25 +04:00
|
|
|
|
2006-05-19 10:31:12 +04:00
|
|
|
def find(self, node, f):
|
|
|
|
'''look up entry for a single file efficiently.
|
2007-03-07 21:25:58 +03:00
|
|
|
return (node, flags) pair if found, (None, None) if not.'''
|
2013-02-09 19:43:02 +04:00
|
|
|
if node in self._mancache:
|
2015-02-24 20:08:54 +03:00
|
|
|
m = self._mancache[node][0]
|
|
|
|
return m.get(f), m.flags(f)
|
2006-05-19 10:31:12 +04:00
|
|
|
text = self.revision(node)
|
2015-03-07 20:04:39 +03:00
|
|
|
try:
|
|
|
|
return manifestdict(text)._lm[f]
|
|
|
|
except KeyError:
|
2006-05-19 10:31:12 +04:00
|
|
|
return None, None
|
|
|
|
|
2015-02-24 20:08:54 +03:00
|
|
|
def add(self, m, transaction, link, p1, p2, added, removed):
|
2014-10-08 20:59:11 +04:00
|
|
|
if p1 in self._mancache:
|
|
|
|
# If our first parent is in the manifest cache, we can
|
|
|
|
# compute a delta here using properties we know about the
|
|
|
|
# manifest up-front, which may save time later for the
|
|
|
|
# revlog layer.
|
2005-08-28 01:21:25 +04:00
|
|
|
|
2014-09-11 18:14:34 +04:00
|
|
|
_checkforbidden(added)
|
2005-08-28 01:21:25 +04:00
|
|
|
# combine the changed lists into one list for sorting
|
2009-09-02 23:05:43 +04:00
|
|
|
work = [(x, False) for x in added]
|
|
|
|
work.extend((x, True) for x in removed)
|
2012-08-28 01:14:27 +04:00
|
|
|
# this could use heapq.merge() (from Python 2.6+) or equivalent
|
2009-09-02 23:05:43 +04:00
|
|
|
# since the lists are already sorted
|
2005-08-28 01:21:25 +04:00
|
|
|
work.sort()
|
|
|
|
|
2015-02-24 20:08:54 +03:00
|
|
|
arraytext, deltatext = m.fastdelta(self._mancache[p1][1], work)
|
2014-10-08 23:20:14 +04:00
|
|
|
cachedelta = self.rev(p1), deltatext
|
2011-12-16 01:27:11 +04:00
|
|
|
text = util.buffer(arraytext)
|
2014-10-08 20:59:11 +04:00
|
|
|
else:
|
|
|
|
# The first parent manifest isn't already loaded, so we'll
|
|
|
|
# just encode a fulltext of the manifest and pass that
|
|
|
|
# through to the revlog layer, and let it handle the delta
|
|
|
|
# process.
|
2015-02-24 20:08:54 +03:00
|
|
|
text = m.text()
|
2014-10-08 20:59:11 +04:00
|
|
|
arraytext = array.array('c', text)
|
|
|
|
cachedelta = None
|
2005-08-28 01:21:25 +04:00
|
|
|
|
2009-09-04 12:47:55 +04:00
|
|
|
n = self.addrevision(text, transaction, link, p1, p2, cachedelta)
|
2015-02-24 20:08:54 +03:00
|
|
|
self._mancache[n] = (m, arraytext)
|
2005-08-28 01:21:25 +04:00
|
|
|
|
|
|
|
return n
|