##// END OF EJS Templates
hgweb: pre-init mimetypes module (fixes ugly bug in python-2.6.2 mimetypes)...
hgweb: pre-init mimetypes module (fixes ugly bug in python-2.6.2 mimetypes) The ugly initialization hack in mimetypes' initialization procedure can make hg serve hit the maximum recursion depth when in threaded mode.

File last commit:

r8210:344751cd default
r8224:1075f5c1 default
Show More
store.py
295 lines | 10.1 KiB | text/x-python | PythonLexer
Adrian Buehlmann
move filename encoding functions from util.py to new store.py
r6839 # store.py - repository store handling for Mercurial
#
# Copyright 2008 Matt Mackall <mpm@selenic.com>
#
# This software may be used and distributed according to the terms
# of the GNU General Public License, incorporated herein by reference.
Adrian Buehlmann
introduce fncache repository layout...
r7229 from i18n import _
Adrian Buehlmann
introduce store classes...
r6840 import os, stat, osutil, util
Adrian Buehlmann
introduce fncache repository layout...
r7229 _sha = util.sha1
Adrian Buehlmann
move filename encoding functions from util.py to new store.py
r6839 def _buildencodefun():
e = '_'
win_reserved = [ord(x) for x in '\\:*?"<>|']
cmap = dict([ (chr(x), chr(x)) for x in xrange(127) ])
for x in (range(32) + range(126, 256) + win_reserved):
cmap[chr(x)] = "~%02x" % x
for x in range(ord("A"), ord("Z")+1) + [ord(e)]:
cmap[chr(x)] = e + chr(x).lower()
dmap = {}
for k, v in cmap.iteritems():
dmap[v] = k
def decode(s):
i = 0
while i < len(s):
for l in xrange(1, 4):
try:
yield dmap[s[i:i+l]]
i += l
break
except KeyError:
pass
else:
raise KeyError
return (lambda s: "".join([cmap[c] for c in s]),
lambda s: "".join(list(decode(s))))
encodefilename, decodefilename = _buildencodefun()
Adrian Buehlmann
introduce fncache repository layout...
r7229 def _build_lower_encodefun():
win_reserved = [ord(x) for x in '\\:*?"<>|']
cmap = dict([ (chr(x), chr(x)) for x in xrange(127) ])
for x in (range(32) + range(126, 256) + win_reserved):
cmap[chr(x)] = "~%02x" % x
for x in range(ord("A"), ord("Z")+1):
cmap[chr(x)] = chr(x).lower()
return lambda s: "".join([cmap[c] for c in s])
lowerencode = _build_lower_encodefun()
_windows_reserved_filenames = '''con prn aux nul
com1 com2 com3 com4 com5 com6 com7 com8 com9
lpt1 lpt2 lpt3 lpt4 lpt5 lpt6 lpt7 lpt8 lpt9'''.split()
def auxencode(path):
res = []
for n in path.split('/'):
if n:
base = n.split('.')[0]
if base and (base in _windows_reserved_filenames):
# encode third letter ('aux' -> 'au~78')
ec = "~%02x" % ord(n[2])
n = n[0:2] + ec + n[3:]
Adrian Buehlmann
store: encode trailing period and space on directory names (issue1417)...
r7515 if n[-1] in '. ':
# encode last period or space ('foo...' -> 'foo..~2e')
n = n[:-1] + "~%02x" % ord(n[-1])
Adrian Buehlmann
introduce fncache repository layout...
r7229 res.append(n)
return '/'.join(res)
MAX_PATH_LEN_IN_HGSTORE = 120
DIR_PREFIX_LEN = 8
_MAX_SHORTENED_DIRS_LEN = 8 * (DIR_PREFIX_LEN + 1) - 4
def hybridencode(path):
'''encodes path with a length limit
Encodes all paths that begin with 'data/', according to the following.
Default encoding (reversible):
Encodes all uppercase letters 'X' as '_x'. All reserved or illegal
characters are encoded as '~xx', where xx is the two digit hex code
of the character (see encodefilename).
Relevant path components consisting of Windows reserved filenames are
masked by encoding the third character ('aux' -> 'au~78', see auxencode).
Hashed encoding (not reversible):
If the default-encoded path is longer than MAX_PATH_LEN_IN_HGSTORE, a
non-reversible hybrid hashing of the path is done instead.
This encoding uses up to DIR_PREFIX_LEN characters of all directory
levels of the lowerencoded path, but not more levels than can fit into
_MAX_SHORTENED_DIRS_LEN.
Then follows the filler followed by the sha digest of the full path.
The filler is the beginning of the basename of the lowerencoded path
(the basename is everything after the last path separator). The filler
is as long as possible, filling in characters from the basename until
the encoded path has MAX_PATH_LEN_IN_HGSTORE characters (or all chars
of the basename have been taken).
The extension (e.g. '.i' or '.d') is preserved.
The string 'data/' at the beginning is replaced with 'dh/', if the hashed
encoding was used.
'''
if not path.startswith('data/'):
return path
ndpath = path[len('data/'):]
res = 'data/' + auxencode(encodefilename(ndpath))
if len(res) > MAX_PATH_LEN_IN_HGSTORE:
digest = _sha(path).hexdigest()
aep = auxencode(lowerencode(ndpath))
_root, ext = os.path.splitext(aep)
parts = aep.split('/')
basename = parts[-1]
sdirs = []
for p in parts[:-1]:
d = p[:DIR_PREFIX_LEN]
Adrian Buehlmann
store: don't create dirs ending in period or space for hashed paths (issue1417)...
r7514 if d[-1] in '. ':
# Windows can't access dirs ending in period or space
d = d[:-1] + '_'
Adrian Buehlmann
introduce fncache repository layout...
r7229 t = '/'.join(sdirs) + '/' + d
if len(t) > _MAX_SHORTENED_DIRS_LEN:
break
sdirs.append(d)
dirs = '/'.join(sdirs)
if len(dirs) > 0:
dirs += '/'
res = 'dh/' + dirs + digest + ext
space_left = MAX_PATH_LEN_IN_HGSTORE - len(res)
if space_left > 0:
filler = basename[:space_left]
res = 'dh/' + dirs + filler + digest + ext
return res
Matt Mackall
store: simplify class hierarchy
r6898 def _calcmode(path):
try:
# files in .hg/ will be created using this mode
mode = os.stat(path).st_mode
# avoid some useless chmods
Matt Mackall
util: split out posix, windows, and win32 modules
r7890 if (0777 & ~util.umask) == (0777 & mode):
Matt Mackall
store: simplify class hierarchy
r6898 mode = None
except OSError:
mode = None
return mode
Matt Mackall
clone: get a list of files to clone from store
r6903 _data = 'data 00manifest.d 00manifest.i 00changelog.d 00changelog.i'
Matt Mackall
store: simplify class hierarchy
r6898 class basicstore:
Adrian Buehlmann
introduce store classes...
r6840 '''base class for local repository stores'''
Adrian Buehlmann
Fix for Issue1260
r6988 def __init__(self, path, opener, pathjoiner):
self.pathjoiner = pathjoiner
Adrian Buehlmann
introduce store classes...
r6840 self.path = path
Matt Mackall
store: simplify class hierarchy
r6898 self.createmode = _calcmode(path)
self.opener = opener(self.path)
self.opener.createmode = self.createmode
Adrian Buehlmann
introduce store classes...
r6840
def join(self, f):
Adrian Buehlmann
Fix for Issue1260
r6988 return self.pathjoiner(self.path, f)
Adrian Buehlmann
introduce store classes...
r6840
Matt Mackall
store: simplify walking...
r6899 def _walk(self, relpath, recurse):
Matt Mackall
store: change handling of decoding errors
r6900 '''yields (unencoded, encoded, size)'''
Adrian Buehlmann
Fix for Issue1260
r6988 path = self.pathjoiner(self.path, relpath)
Adrian Buehlmann
introduce store classes...
r6840 striplen = len(self.path) + len(os.sep)
Matt Mackall
store: simplify walking...
r6899 l = []
if os.path.isdir(path):
visit = [path]
while visit:
p = visit.pop()
for f, kind, st in osutil.listdir(p, stat=True):
Adrian Buehlmann
Fix for Issue1260
r6988 fp = self.pathjoiner(p, f)
Matt Mackall
store: simplify walking...
r6899 if kind == stat.S_IFREG and f[-2:] in ('.d', '.i'):
Matt Mackall
store: change handling of decoding errors
r6900 n = util.pconvert(fp[striplen:])
l.append((n, n, st.st_size))
Matt Mackall
store: simplify walking...
r6899 elif kind == stat.S_IFDIR and recurse:
visit.append(fp)
Matt Mackall
replace util.sort with sorted built-in...
r8209 return sorted(l)
Adrian Buehlmann
introduce store classes...
r6840
Matt Mackall
store: change handling of decoding errors
r6900 def datafiles(self):
Matt Mackall
store: simplify walking...
r6899 return self._walk('data', True)
Adrian Buehlmann
introduce store classes...
r6840
def walk(self):
Matt Mackall
store: change handling of decoding errors
r6900 '''yields (unencoded, encoded, size)'''
Adrian Buehlmann
introduce store classes...
r6840 # yield data files first
Adrian Buehlmann
verify: check repo.store
r6892 for x in self.datafiles():
Adrian Buehlmann
introduce store classes...
r6840 yield x
# yield manifest before changelog
Matt Mackall
replace various uses of list.reverse()
r8210 for x in reversed(self._walk('', False)):
Adrian Buehlmann
introduce store classes...
r6840 yield x
Matt Mackall
clone: get a list of files to clone from store
r6903 def copylist(self):
return ['requires'] + _data.split()
Matt Mackall
store: simplify class hierarchy
r6898 class encodedstore(basicstore):
Adrian Buehlmann
Fix for Issue1260
r6988 def __init__(self, path, opener, pathjoiner):
self.pathjoiner = pathjoiner
self.path = self.pathjoiner(path, 'store')
Matt Mackall
store: simplify class hierarchy
r6898 self.createmode = _calcmode(self.path)
Matt Mackall
store: take opener as an argument
r6896 op = opener(self.path)
Adrian Buehlmann
introduce store classes...
r6840 op.createmode = self.createmode
Matt Mackall
store: drop self.encodefn
r6902 self.opener = lambda f, *args, **kw: op(encodefilename(f), *args, **kw)
Adrian Buehlmann
introduce store classes...
r6840
Matt Mackall
store: change handling of decoding errors
r6900 def datafiles(self):
for a, b, size in self._walk('data', True):
Adrian Buehlmann
verify: check repo.store
r6892 try:
Matt Mackall
store: change handling of decoding errors
r6900 a = decodefilename(a)
Adrian Buehlmann
verify: check repo.store
r6892 except KeyError:
Matt Mackall
store: change handling of decoding errors
r6900 a = None
yield a, b, size
Adrian Buehlmann
introduce store classes...
r6840
def join(self, f):
Adrian Buehlmann
Fix for Issue1260
r6988 return self.pathjoiner(self.path, encodefilename(f))
Adrian Buehlmann
introduce store classes...
r6840
Matt Mackall
clone: get a list of files to clone from store
r6903 def copylist(self):
return (['requires', '00changelog.i'] +
Adrian Buehlmann
Fix for Issue1260
r6988 [self.pathjoiner('store', f) for f in _data.split()])
Matt Mackall
clone: get a list of files to clone from store
r6903
Adrian Buehlmann
introduce fncache repository layout...
r7229 def fncache(opener):
'''yields the entries in the fncache file'''
try:
fp = opener('fncache', mode='rb')
except IOError:
# skip nonexistent file
return
for n, line in enumerate(fp):
if (len(line) < 2) or (line[-1] != '\n'):
t = _('invalid entry in fncache, line %s') % (n + 1)
raise util.Abort(t)
yield line[:-1]
fp.close()
class fncacheopener(object):
def __init__(self, opener):
self.opener = opener
self.entries = None
def loadfncache(self):
self.entries = {}
for f in fncache(self.opener):
self.entries[f] = True
def __call__(self, path, mode='r', *args, **kw):
if mode not in ('r', 'rb') and path.startswith('data/'):
if self.entries is None:
self.loadfncache()
if path not in self.entries:
self.opener('fncache', 'ab').write(path + '\n')
# fncache may contain non-existent files after rollback / strip
self.entries[path] = True
return self.opener(hybridencode(path), mode, *args, **kw)
class fncachestore(basicstore):
def __init__(self, path, opener, pathjoiner):
self.pathjoiner = pathjoiner
self.path = self.pathjoiner(path, 'store')
self.createmode = _calcmode(self.path)
self._op = opener(self.path)
self._op.createmode = self.createmode
self.opener = fncacheopener(self._op)
def join(self, f):
return self.pathjoiner(self.path, hybridencode(f))
def datafiles(self):
rewrite = False
existing = []
pjoin = self.pathjoiner
spath = self.path
for f in fncache(self._op):
ef = hybridencode(f)
try:
st = os.stat(pjoin(spath, ef))
yield f, ef, st.st_size
existing.append(f)
except OSError:
# nonexistent entry
rewrite = True
if rewrite:
# rewrite fncache to remove nonexistent entries
# (may be caused by rollback / strip)
fp = self._op('fncache', mode='wb')
for p in existing:
fp.write(p + '\n')
fp.close()
def copylist(self):
Thomas Arendsen Hein
Remove trailing spaces.
r7236 d = _data + ' dh fncache'
Adrian Buehlmann
introduce fncache repository layout...
r7229 return (['requires', '00changelog.i'] +
[self.pathjoiner('store', f) for f in d.split()])
Patrick Mezard
store: pathjoiner default value is os.path.join
r6989 def store(requirements, path, opener, pathjoiner=None):
pathjoiner = pathjoiner or os.path.join
Matt Mackall
store: simplify class hierarchy
r6898 if 'store' in requirements:
Adrian Buehlmann
introduce fncache repository layout...
r7229 if 'fncache' in requirements:
return fncachestore(path, opener, pathjoiner)
Adrian Buehlmann
Fix for Issue1260
r6988 return encodedstore(path, opener, pathjoiner)
return basicstore(path, opener, pathjoiner)