fdatasync,
hostname, localtime, log, merge_iter,
mmap_read, mmap_readwrite,
+ parse_num,
progress, qprogress, stat_if_exists,
unlink, username, userfullname,
utc_offset_str)
-
-max_pack_size = 1000*1000*1000 # larger packs will slow down pruning
-max_pack_objects = 200*1000 # cache memory usage is about 83 bytes per object
-
verbose = 0
ignore_midx = 0
repodir = None # The default repository, once initialized
_git_wait(repr(argv), p)
return r
+def git_config_get(option, repo_dir=None):
+ cmd = ('git', 'config', '--get', option)
+ p = subprocess.Popen(cmd, stdout=subprocess.PIPE,
+ preexec_fn=_gitenv(repo_dir=repo_dir))
+ r = p.stdout.read()
+ rc = p.wait()
+ if rc == 0:
+ return r
+ if rc != 1:
+ raise GitError('%s returned %d' % (cmd, rc))
+ return None
+
def parse_tz_offset(s):
"""UTC offset in seconds."""
class PackWriter:
"""Writes Git objects inside a pack file."""
def __init__(self, objcache_maker=_make_objcache, compression_level=1,
- run_midx=True, on_pack_finish=None):
+ run_midx=True, on_pack_finish=None,
+ max_pack_size=None, max_pack_objects=None):
+ self.repo_dir = repo()
self.file = None
self.parentfd = None
self.count = 0
self.compression_level = compression_level
self.run_midx=run_midx
self.on_pack_finish = on_pack_finish
+ if not max_pack_size:
+ max_pack_size = git_config_get('pack.packSizeLimit',
+ repo_dir=self.repo_dir)
+ if max_pack_size is not None:
+ max_pack_size = parse_num(max_pack_size)
+ if not max_pack_size:
+ # larger packs slow down pruning
+ max_pack_size = 1000 * 1000 * 1000
+ self.max_pack_size = max_pack_size
+ # cache memory usage is about 83 bytes per object
+ self.max_pack_objects = max_pack_objects if max_pack_objects \
+ else max(1, self.max_pack_size // 5000)
def __del__(self):
self.close()
def _open(self):
if not self.file:
- objdir = dir=repo('objects')
+ objdir = dir = os.path.join(self.repo_dir, 'objects')
fd, name = tempfile.mkstemp(suffix='.pack', dir=objdir)
try:
self.file = os.fdopen(fd, 'w+b')
size, crc = self._raw_write(_encode_packobj(type, content,
self.compression_level),
sha=sha)
- if self.outbytes >= max_pack_size or self.count >= max_pack_objects:
+ if self.outbytes >= self.max_pack_size \
+ or self.count >= self.max_pack_objects:
self.breakpoint()
return sha
f.close()
obj_list_sha = self._write_pack_idx_v2(self.filename + '.idx', idx, packbin)
-
- nameprefix = repo('objects/pack/pack-%s' % obj_list_sha)
+ nameprefix = os.path.join(self.repo_dir,
+ 'objects/pack/pack-' + obj_list_sha)
if os.path.exists(self.filename + '.map'):
os.unlink(self.filename + '.map')
os.rename(self.filename + '.pack', nameprefix + '.pack')
os.close(self.parentfd)
if run_midx:
- auto_midx(repo('objects/pack'))
+ auto_midx(os.path.join(self.repo_dir, 'objects/pack'))
if self.on_pack_finish:
self.on_pack_finish(nameprefix)
self.repo_dir = repo_dir
wanted = ('1','5','6')
if ver() < wanted:
- if not _ver_warned:
- log('warning: git version < %s; bup will be slow.\n'
- % '.'.join(wanted))
- _ver_warned = 1
- self.get = self._slow_get
- else:
- self.p = self.inprogress = None
- self.get = self._fast_get
+ log('error: git version must be at least 1.5.6\n')
+ sys.exit(1)
+ self.p = self.inprogress = None
def _abort(self):
if self.p:
bufsize = 4096,
preexec_fn = _gitenv(self.repo_dir))
- def _fast_get(self, id):
+ def get(self, id, size=False):
+ """Yield the object type, and then an iterator over the data referred
+ to by the id ref. If size is true, yield (obj_type, obj_size)
+ instead of just the type.
+
+ """
if not self.p or self.p.poll() != None:
self.restart()
assert(self.p)
poll_result = self.p.poll()
assert(poll_result == None)
if self.inprogress:
- log('_fast_get: opening %r while %r is open\n'
- % (id, self.inprogress))
+ log('get: opening %r while %r is open\n' % (id, self.inprogress))
assert(not self.inprogress)
assert(id.find('\n') < 0)
assert(id.find('\r') < 0)
spl = hdr.split(' ')
if len(spl) != 3 or len(spl[0]) != 40:
raise GitError('expected blob, got %r' % spl)
- (hex, type, size) = spl
-
- it = _AbortableIter(chunkyreader(self.p.stdout, int(spl[2])),
- onabort = self._abort)
+ hex, typ, sz = spl
+ sz = int(sz)
+ it = _AbortableIter(chunkyreader(self.p.stdout, sz),
+ onabort=self._abort)
try:
- yield type
+ if size:
+ yield typ, sz
+ else:
+ yield typ
for blob in it:
yield blob
readline_result = self.p.stdout.readline()
it.abort()
raise
- def _slow_get(self, id):
- assert(id.find('\n') < 0)
- assert(id.find('\r') < 0)
- assert(id[0] != '-')
- type = _git_capture(['git', 'cat-file', '-t', id]).strip()
- yield type
-
- p = subprocess.Popen(['git', 'cat-file', type, id],
- stdout=subprocess.PIPE,
- preexec_fn = _gitenv(self.repo_dir))
- for blob in chunkyreader(p.stdout):
- yield blob
- _git_wait('git cat-file', p)
-
def _join(self, it):
type = it.next()
if type == 'blob':
projects / bup.git / blobdiff