"""
dirstate.py - working directory tracking for mercurial
Copyright 2005, 2006 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.
"""
from node import *
from i18n import _
import struct, os, time, bisect, stat, strutil, util, re, errno, ignore
import cStringIO
class dirstate(object):
format = ">cllll"
def __init__(self, opener, ui, root):
self.opener = opener
self.root = root
self.dirty = 0
self.ui = ui
self._slash = None
def __getattr__(self, name):
if name == 'map':
self.read()
return self.map
elif name == 'copymap':
self.read()
return self.copymap
elif name == '_branch':
try:
self._branch = self.opener("branch").read().strip()\
or "default"
except IOError:
self._branch = "default"
return self._branch
elif name == 'pl':
self.pl = [nullid, nullid]
try:
st = self.opener("dirstate").read(40)
if len(st) == 40:
self.pl = st[:20], st[20:40]
except IOError, err:
if err.errno != errno.ENOENT: raise
return self.pl
elif name == 'dirs':
self.dirs = {}
for f in self.map:
self.updatedirs(f, 1)
return self.dirs
elif name == '_ignore':
files = [self.wjoin('.hgignore')] + self.ui.hgignorefiles()
self._ignore = ignore.ignore(self.root, files, self.ui.warn)
return self._ignore
else:
raise AttributeError, name
def wjoin(self, f):
return os.path.join(self.root, f)
def getcwd(self):
cwd = os.getcwd()
if cwd == self.root: return ''
# self.root ends with a path separator if self.root is '/' or 'C:\'
rootsep = self.root
if not rootsep.endswith(os.sep):
rootsep += os.sep
if cwd.startswith(rootsep):
return cwd[len(rootsep):]
else:
# we're outside the repo. return an absolute path.
return cwd
def pathto(self, f, cwd=None):
if cwd is None:
cwd = self.getcwd()
path = util.pathto(self.root, cwd, f)
if self._slash is None:
self._slash = self.ui.configbool('ui', 'slash') and os.sep != '/'
if self._slash:
path = path.replace(os.sep, '/')
return path
def __del__(self):
if self.dirty:
self.write()
def __getitem__(self, key):
return self.map[key]
_unknown = ('?', 0, 0, 0)
def get(self, key):
try:
return self[key]
except KeyError:
return self._unknown
def __contains__(self, key):
return key in self.map
def parents(self):
return self.pl
def branch(self):
return self._branch
def markdirty(self):
if not self.dirty:
self.dirty = 1
def setparents(self, p1, p2=nullid):
self.markdirty()
self.pl = p1, p2
def setbranch(self, branch):
self._branch = branch
self.opener("branch", "w").write(branch + '\n')
def state(self, key):
try:
return self[key][0]
except KeyError:
return "?"
def read(self):
self.map = {}
self.copymap = {}
self.pl = [nullid, nullid]
try:
st = self.opener("dirstate").read()
except IOError, err:
if err.errno != errno.ENOENT: raise
return
if not st:
return
self.pl = [st[:20], st[20: 40]]
# deref fields so they will be local in loop
dmap = self.map
copymap = self.copymap
format = self.format
unpack = struct.unpack
pos = 40
e_size = struct.calcsize(format)
while pos < len(st):
newpos = pos + e_size
e = unpack(format, st[pos:newpos])
l = e[4]
pos = newpos
newpos = pos + l
f = st[pos:newpos]
if '\0' in f:
f, c = f.split('\0')
copymap[f] = c
dmap[f] = e[:4]
pos = newpos
def reload(self):
for a in "map copymap _branch pl dirs _ignore".split():
if hasattr(self, a):
self.__delattr__(a)
def copy(self, source, dest):
self.markdirty()
self.copymap[dest] = source
def copied(self, file):
return self.copymap.get(file, None)
def copies(self):
return self.copymap
def updatedirs(self, path, delta):
for c in strutil.findall(path, '/'):
pc = path[:c]
self.dirs.setdefault(pc, 0)
self.dirs[pc] += delta
def checkinterfering(self, files):
def prefixes(f):
for c in strutil.rfindall(f, '/'):
yield f[:c]
seendirs = {}
for f in files:
# shadows
if self.dirs.get(f):
raise util.Abort(_('directory named %r already in dirstate') %
f)
for d in prefixes(f):
if d in seendirs:
break
if d in self.map:
raise util.Abort(_('file named %r already in dirstate') %
d)
seendirs[d] = True
# disallowed
if '\r' in f or '\n' in f:
raise util.Abort(_("'\\n' and '\\r' disallowed in filenames"))
def update(self, files, state, **kw):
''' current states:
n normal
m needs merging
r marked for removal
a marked for addition'''
if not files: return
self.markdirty()
if state == "a":
self.checkinterfering(files)
for f in files:
if state == "r":
self.map[f] = ('r', 0, 0, 0)
self.updatedirs(f, -1)
else:
if state == "a":
self.updatedirs(f, 1)
s = os.lstat(self.wjoin(f))
st_size = kw.get('st_size', s.st_size)
st_mtime = kw.get('st_mtime', s.st_mtime)
self.map[f] = (state, s.st_mode, st_size, st_mtime)
if self.copymap.has_key(f):
del self.copymap[f]
def forget(self, files):
if not files: return
self.markdirty()
for f in files:
try:
del self.map[f]
self.updatedirs(f, -1)
except KeyError:
self.ui.warn(_("not in dirstate: %s!\n") % f)
pass
def rebuild(self, parent, files):
self.reload()
for f in files:
if files.execf(f):
self.map[f] = ('n', 0777, -1, 0)
else:
self.map[f] = ('n', 0666, -1, 0)
self.pl = (parent, nullid)
self.markdirty()
def write(self):
if not self.dirty:
return
cs = cStringIO.StringIO()
cs.write("".join(self.pl))
for f, e in self.map.iteritems():
c = self.copied(f)
if c:
f = f + "\0" + c
e = struct.pack(self.format, e[0], e[1], e[2], e[3], len(f))
cs.write(e)
cs.write(f)
st = self.opener("dirstate", "w", atomictemp=True)
st.write(cs.getvalue())
st.rename()
self.dirty = 0
def filterfiles(self, files):
ret = {}
unknown = []
for x in files:
if x == '.':
return self.map.copy()
if x not in self.map:
unknown.append(x)
else:
ret[x] = self.map[x]
if not unknown:
return ret
b = self.map.keys()
b.sort()
blen = len(b)
for x in unknown:
bs = bisect.bisect(b, "%s%s" % (x, '/'))
while bs < blen:
s = b[bs]
if len(s) > len(x) and s.startswith(x):
ret[s] = self.map[s]
else:
break
bs += 1
return ret
def supported_type(self, f, st, verbose=False):
if stat.S_ISREG(st.st_mode) or stat.S_ISLNK(st.st_mode):
return True
if verbose:
kind = 'unknown'
if stat.S_ISCHR(st.st_mode): kind = _('character device')
elif stat.S_ISBLK(st.st_mode): kind = _('block device')
elif stat.S_ISFIFO(st.st_mode): kind = _('fifo')
elif stat.S_ISSOCK(st.st_mode): kind = _('socket')
elif stat.S_ISDIR(st.st_mode): kind = _('directory')
self.ui.warn(_('%s: unsupported file type (type is %s)\n')
% (self.pathto(f), kind))
return False
def walk(self, files=None, match=util.always, badmatch=None):
# filter out the stat
for src, f, st in self.statwalk(files, match, badmatch=badmatch):
yield src, f
def statwalk(self, files=None, match=util.always, ignored=False,
badmatch=None, directories=False):
'''
walk recursively through the directory tree, finding all files
matched by the match function
results are yielded in a tuple (src, filename, st), where src
is one of:
'f' the file was found in the directory tree
'd' the file is a directory of the tree
'm' the file was only in the dirstate and not in the tree
'b' file was not found and matched badmatch
and st is the stat result if the file was found in the directory.
'''
# walk all files by default
if not files:
files = ['.']
dc = self.map.copy()
else:
files = util.unique(files)
dc = self.filterfiles(files)
def imatch(file_):
if file_ not in dc and self._ignore(file_):
return False
return match(file_)
ignore = self._ignore
if ignored:
imatch = match
ignore = util.never
# self.root may end with a path separator when self.root == '/'
common_prefix_len = len(self.root)
if not self.root.endswith(os.sep):
common_prefix_len += 1
# recursion free walker, faster than os.walk.
def findfiles(s):
work = [s]
if directories:
yield 'd', util.normpath(s[common_prefix_len:]), os.lstat(s)
while work:
top = work.pop()
names = os.listdir(top)
names.sort()
# nd is the top of the repository dir tree
nd = util.normpath(top[common_prefix_len:])
if nd == '.':
nd = ''
else:
# do not recurse into a repo contained in this
# one. use bisect to find .hg directory so speed
# is good on big directory.
hg = bisect.bisect_left(names, '.hg')
if hg < len(names) and names[hg] == '.hg':
if os.path.isdir(os.path.join(top, '.hg')):
continue
for f in names:
np = util.pconvert(os.path.join(nd, f))
if seen(np):
continue
p = os.path.join(top, f)
# don't trip over symlinks
st = os.lstat(p)
if stat.S_ISDIR(st.st_mode):
if not ignore(np):
work.append(p)
if directories:
yield 'd', np, st
if imatch(np) and np in dc:
yield 'm', np, st
elif imatch(np):
if self.supported_type(np, st):
yield 'f', np, st
elif np in dc:
yield 'm', np, st
known = {'.hg': 1}
def seen(fn):
if fn in known: return True
known[fn] = 1
# step one, find all files that match our criteria
files.sort()
for ff in files:
nf = util.normpath(ff)
f = self.wjoin(ff)
try:
st = os.lstat(f)
except OSError, inst:
found = False
for fn in dc:
if nf == fn or (fn.startswith(nf) and fn[len(nf)] == '/'):
found = True
break
if not found:
if inst.errno != errno.ENOENT or not badmatch:
self.ui.warn('%s: %s\n' % (self.pathto(ff),
inst.strerror))
elif badmatch and badmatch(ff) and imatch(nf):
yield 'b', ff, None
continue
if stat.S_ISDIR(st.st_mode):
cmp1 = (lambda x, y: cmp(x[1], y[1]))
sorted_ = [ x for x in findfiles(f) ]
sorted_.sort(cmp1)
for e in sorted_:
yield e
else:
if not seen(nf) and match(nf):
if self.supported_type(ff, st, verbose=True):
yield 'f', nf, st
elif ff in dc:
yield 'm', nf, st
# step two run through anything left in the dc hash and yield
# if we haven't already seen it
ks = dc.keys()
ks.sort()
for k in ks:
if not seen(k) and imatch(k):
yield 'm', k, None
def status(self, files=None, match=util.always, list_ignored=False,
list_clean=False):
lookup, modified, added, unknown, ignored = [], [], [], [], []
removed, deleted, clean = [], [], []
for src, fn, st in self.statwalk(files, match, ignored=list_ignored):
try:
type_, mode, size, time = self[fn]
except KeyError:
if list_ignored and self._ignore(fn):
ignored.append(fn)
else:
unknown.append(fn)
continue
if src == 'm':
nonexistent = True
if not st:
try:
st = os.lstat(self.wjoin(fn))
except OSError, inst:
if inst.errno != errno.ENOENT:
raise
st = None
# We need to re-check that it is a valid file
if st and self.supported_type(fn, st):
nonexistent = False
# XXX: what to do with file no longer present in the fs
# who are not removed in the dirstate ?
if nonexistent and type_ in "nm":
deleted.append(fn)
continue
# check the common case first
if type_ == 'n':
if not st:
st = os.lstat(self.wjoin(fn))
if size >= 0 and (size != st.st_size
or (mode ^ st.st_mode) & 0100):
modified.append(fn)
elif time != int(st.st_mtime):
lookup.append(fn)
elif list_clean:
clean.append(fn)
elif type_ == 'm':
modified.append(fn)
elif type_ == 'a':
added.append(fn)
elif type_ == 'r':
removed.append(fn)
return (lookup, modified, added, removed, deleted, unknown, ignored,
clean)