X-Git-Url: https://arthur.barton.de/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=lib%2Fbup%2Fgit.py;h=13913a56328a418f551d7c727a22798ac4920793;hb=e861537c86a6b25b39a093342408d0e222f32afc;hp=798bea2cb7328c9ca9356379843e2f5029194952;hpb=b22a2d815444007637ee1dc676f618439b3f160f;p=bup.git diff --git a/lib/bup/git.py b/lib/bup/git.py index 798bea2..13913a5 100644 --- a/lib/bup/git.py +++ b/lib/bup/git.py @@ -3,20 +3,25 @@ bup repositories are in Git format. This library allows us to interact with the Git data structures. """ +from __future__ import absolute_import import errno, os, sys, zlib, time, subprocess, struct, stat, re, tempfile, glob from collections import namedtuple from itertools import islice from numbers import Integral from bup import _helpers, compat, hashsplit, path, midx, bloom, xstat +from bup.compat import range from bup.helpers import (Sha1, add_error, chunkyreader, debug1, debug2, fdatasync, - hostname, localtime, log, merge_iter, + hostname, localtime, log, + merge_dict, + merge_iter, mmap_read, mmap_readwrite, parse_num, progress, qprogress, shstr, stat_if_exists, - unlink, username, userfullname, + unlink, utc_offset_str) +from bup.pwdgrp import username, userfullname verbose = 0 ignore_midx = 0 @@ -33,13 +38,18 @@ class GitError(Exception): pass +def _gitenv(repo_dir=None): + if not repo_dir: + repo_dir = repo() + return merge_dict(os.environ, {'GIT_DIR': os.path.abspath(repo_dir)}) + def _git_wait(cmd, p): rv = p.wait() if rv != 0: raise GitError('%s returned %d' % (shstr(cmd), rv)) def _git_capture(argv): - p = subprocess.Popen(argv, stdout=subprocess.PIPE, preexec_fn = _gitenv()) + p = subprocess.Popen(argv, stdout=subprocess.PIPE, env=_gitenv()) r = p.stdout.read() _git_wait(repr(argv), p) return r @@ -47,7 +57,7 @@ def _git_capture(argv): 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)) + env=_gitenv(repo_dir=repo_dir)) r = p.stdout.read() rc = p.wait() if rc == 0: @@ -74,17 +84,21 @@ _safe_str_rx = '(?:%s{1,2}|(?:%s%s*%s))' \ _start_end_char, _content_char, _start_end_char) _tz_rx = r'[-+]\d\d[0-5]\d' _parent_rx = r'(?:parent [abcdefABCDEF0123456789]{40}\n)' +# Assumes every following line starting with a space is part of the +# mergetag. Is there a formal commit blob spec? +_mergetag_rx = r'(?:\nmergetag object [abcdefABCDEF0123456789]{40}(?:\n [^\0\n]*)*)' _commit_rx = re.compile(r'''tree (?P[abcdefABCDEF0123456789]{40}) (?P%s*)author (?P%s) <(?P%s)> (?P\d+) (?P%s) -committer (?P%s) <(?P%s)> (?P\d+) (?P%s) +committer (?P%s) <(?P%s)> (?P\d+) (?P%s)(?P%s?) (?P(?:.|\n)*)''' % (_parent_rx, _safe_str_rx, _safe_str_rx, _tz_rx, - _safe_str_rx, _safe_str_rx, _tz_rx)) + _safe_str_rx, _safe_str_rx, _tz_rx, + _mergetag_rx)) _parent_hash_rx = re.compile(r'\s*parent ([abcdefABCDEF0123456789]{40})\s*') - -# Note that the author_sec and committer_sec values are (UTC) epoch seconds. +# Note that the author_sec and committer_sec values are (UTC) epoch +# seconds, and for now the mergetag is not included. CommitInfo = namedtuple('CommitInfo', ['tree', 'parents', 'author_name', 'author_mail', 'author_sec', 'author_offset', @@ -110,13 +124,14 @@ def parse_commit(content): message=matches['message']) -def get_commit_items(id, cp): - commit_it = cp.get(id) - _, typ, _ = next(commit_it) - assert(typ == 'commit') - commit_content = ''.join(commit_it) - return parse_commit(commit_content) +def get_cat_data(cat_iterator, expected_type): + _, kind, _ = next(cat_iterator) + if kind != expected_type: + raise Exception('expected %r, saw %r' % (expected_type, kind)) + return ''.join(cat_iterator) +def get_commit_items(id, cp): + return parse_commit(get_cat_data(cp.get(id), 'commit')) def _local_git_date_str(epoch_sec): return '%d %s' % (epoch_sec, utc_offset_str(epoch_sec)) @@ -133,7 +148,6 @@ def _git_date_str(epoch_sec, tz_offset_sec): def repo(sub = '', repo_dir=None): """Get the path to the git repository or one of its subdirectories.""" - global repodir repo_dir = repo_dir or repodir if not repo_dir: raise GitError('You should call check_repo_or_die()') @@ -141,7 +155,7 @@ def repo(sub = '', repo_dir=None): # If there's a .git subdirectory, then the actual repo is in there. gd = os.path.join(repo_dir, '.git') if os.path.exists(gd): - repodir = gd + repo_dir = gd return os.path.join(repo_dir, sub) @@ -381,22 +395,25 @@ class PackIdxV1(PackIdx): self.name = filename self.idxnames = [self.name] self.map = mmap_read(f) - self.fanout = list(struct.unpack('!256I', - str(buffer(self.map, 0, 256*4)))) + self.fanout = list(struct.unpack('!256I', buffer(self.map, 0, 256 * 4))) self.fanout.append(0) # entry "-1" nsha = self.fanout[255] self.sha_ofs = 256*4 self.shatable = buffer(self.map, self.sha_ofs, nsha*24) def _ofs_from_idx(self, idx): - return struct.unpack('!I', str(self.shatable[idx*24 : idx*24+4]))[0] + ofs = idx * 24 + return struct.unpack('!I', self.shatable[ofs : ofs + 4])[0] def _idx_to_hash(self, idx): - return str(self.shatable[idx*24+4 : idx*24+24]) + ofs = idx * 24 + 4 + return self.shatable[ofs : ofs + 20] def __iter__(self): - for i in xrange(self.fanout[255]): - yield buffer(self.map, 256*4 + 24*i + 4, 20) + count = self.fanout[255] + start = 256 * 4 + 4 + for ofs in range(start, start + (24 * count), 24): + yield self.map[ofs : ofs + 20] class PackIdxV2(PackIdx): @@ -405,9 +422,9 @@ class PackIdxV2(PackIdx): self.name = filename self.idxnames = [self.name] self.map = mmap_read(f) - assert(str(self.map[0:8]) == '\377tOc\0\0\0\2') + assert self.map[0:8] == b'\377tOc\0\0\0\2' self.fanout = list(struct.unpack('!256I', - str(buffer(self.map, 8, 256*4)))) + buffer(self.map[8 : 8 + 256 * 4]))) self.fanout.append(0) # entry "-1" nsha = self.fanout[255] self.sha_ofs = 8 + 256*4 @@ -419,19 +436,22 @@ class PackIdxV2(PackIdx): 8 + 256*4 + nsha*20 + nsha*4 + nsha*4) def _ofs_from_idx(self, idx): - ofs = struct.unpack('!I', str(buffer(self.ofstable, idx*4, 4)))[0] + i = idx * 4 + ofs = struct.unpack('!I', self.ofstable[i : i + 4])[0] if ofs & 0x80000000: idx64 = ofs & 0x7fffffff - ofs = struct.unpack('!Q', - str(buffer(self.ofs64table, idx64*8, 8)))[0] + idx64_i = idx64 * 8 + ofs = struct.unpack('!Q', self.ofs64table[idx64_i : idx64_i + 8])[0] return ofs def _idx_to_hash(self, idx): - return str(self.shatable[idx*20:(idx+1)*20]) + return self.shatable[idx * 20 : (idx + 1) * 20] def __iter__(self): - for i in xrange(self.fanout[255]): - yield buffer(self.map, 8 + 256*4 + 20*i, 20) + count = self.fanout[255] + start = 8 + 256 * 4 + for ofs in range(start, start + (20 * count), 20): + yield self.map[ofs : ofs + 20] _mpi_count = 0 @@ -552,7 +572,7 @@ class PackIdxList: if self.bloom is None and os.path.exists(bfull): self.bloom = bloom.ShaBloom(bfull) self.packs = list(set(d.values())) - self.packs.sort(lambda x,y: -cmp(len(x),len(y))) + self.packs.sort(reverse=True, key=lambda x: len(x)) if self.bloom and self.bloom.valid() and len(self.bloom) >= len(self): self.do_bloom = True else: @@ -608,8 +628,8 @@ 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, - max_pack_size=None, max_pack_objects=None): - self.repo_dir = repo() + max_pack_size=None, max_pack_objects=None, repo_dir=None): + self.repo_dir = repo_dir or repo() self.file = None self.parentfd = None self.count = 0 @@ -637,6 +657,12 @@ class PackWriter: def __del__(self): self.close() + def __enter__(self): + return self + + def __exit__(self, type, value, traceback): + self.close() + def _open(self): if not self.file: objdir = dir = os.path.join(self.repo_dir, 'objects') @@ -715,17 +741,18 @@ class PackWriter: return self.objcache.exists(id, want_source=want_source) def just_write(self, sha, type, content): - """Write an object to the pack file, bypassing the objcache. Fails if - sha exists().""" + """Write an object to the pack file without checking for duplication.""" self._write(sha, type, content) + # If nothing else, gc doesn't have/want an objcache + if self.objcache is not None: + self.objcache.add(sha) def maybe_write(self, type, content): """Write an object to the pack file if not present and return its id.""" sha = calc_hash(type, content) if not self.exists(sha): - self.just_write(sha, type, content) self._require_objcache() - self.objcache.add(sha) + self.just_write(sha, type, content) return sha def new_blob(self, blob): @@ -874,14 +901,6 @@ class PackWriter: idx_f.close() -def _gitenv(repo_dir = None): - if not repo_dir: - repo_dir = repo() - def env(): - os.environ['GIT_DIR'] = os.path.abspath(repo_dir) - return env - - def list_refs(patterns=None, repo_dir=None, limit_to_heads=False, limit_to_tags=False): """Yield (refname, hash) tuples for all repository refs unless @@ -899,9 +918,7 @@ def list_refs(patterns=None, repo_dir=None, argv.append('--') if patterns: argv.extend(patterns) - p = subprocess.Popen(argv, - preexec_fn = _gitenv(repo_dir), - stdout = subprocess.PIPE) + p = subprocess.Popen(argv, env=_gitenv(repo_dir), stdout=subprocess.PIPE) out = p.stdout.read().strip() rv = p.wait() # not fatal if rv: @@ -955,7 +972,7 @@ def rev_list(ref_or_refs, count=None, parse=None, format=None, repo_dir=None): assert bool(parse) == bool(format) p = subprocess.Popen(rev_list_invocation(ref_or_refs, count=count, format=format), - preexec_fn = _gitenv(repo_dir), + env=_gitenv(repo_dir), stdout = subprocess.PIPE) if not format: for line in p.stdout: @@ -966,7 +983,9 @@ def rev_list(ref_or_refs, count=None, parse=None, format=None, repo_dir=None): s = line.strip() if not s.startswith('commit '): raise Exception('unexpected line ' + s) - yield s[7:], parse(p.stdout) + s = s[7:] + assert len(s) == 40 + yield s, parse(p.stdout) line = p.stdout.readline() rv = p.wait() # not fatal @@ -1020,7 +1039,7 @@ def update_ref(refname, newval, oldval, repo_dir=None): or refname.startswith('refs/tags/')) p = subprocess.Popen(['git', 'update-ref', refname, newval.encode('hex'), oldval.encode('hex')], - preexec_fn = _gitenv(repo_dir)) + env=_gitenv(repo_dir)) _git_wait('git update-ref', p) @@ -1029,7 +1048,7 @@ def delete_ref(refname, oldvalue=None): assert(refname.startswith('refs/')) oldvalue = [] if not oldvalue else [oldvalue] p = subprocess.Popen(['git', 'update-ref', '-d', refname] + oldvalue, - preexec_fn = _gitenv()) + env=_gitenv()) _git_wait('git update-ref', p) @@ -1059,16 +1078,16 @@ def init_repo(path=None): if os.path.exists(d) and not os.path.isdir(os.path.join(d, '.')): raise GitError('"%s" exists but is not a directory\n' % d) p = subprocess.Popen(['git', '--bare', 'init'], stdout=sys.stderr, - preexec_fn = _gitenv()) + env=_gitenv()) _git_wait('git init', p) # Force the index version configuration in order to ensure bup works # regardless of the version of the installed Git binary. p = subprocess.Popen(['git', 'config', 'pack.indexVersion', '2'], - stdout=sys.stderr, preexec_fn = _gitenv()) + stdout=sys.stderr, env=_gitenv()) _git_wait('git config', p) # Enable the reflog p = subprocess.Popen(['git', 'config', 'core.logAllRefUpdates', 'true'], - stdout=sys.stderr, preexec_fn = _gitenv()) + stdout=sys.stderr, env=_gitenv()) _git_wait('git config', p) @@ -1172,7 +1191,7 @@ class CatPipe: stdout=subprocess.PIPE, close_fds = True, bufsize = 4096, - preexec_fn = _gitenv(self.repo_dir)) + env=_gitenv(self.repo_dir)) def get(self, ref): """Yield (oidx, type, size), followed by the data referred to by ref. @@ -1200,7 +1219,7 @@ class CatPipe: return info = hdr.split(' ') if len(info) != 3 or len(info[0]) != 40: - raise GitError('expected object (id, type, size), got %r' % spl) + raise GitError('expected object (id, type, size), got %r' % info) oidx, typ, size = info size = int(size) it = _AbortableIter(chunkyreader(self.p.stdout, size), @@ -1295,14 +1314,13 @@ WalkItem = namedtuple('WalkItem', ['oid', 'type', 'mode', # ... -def walk_object(cat_pipe, oidx, - stop_at=None, - include_data=None): - """Yield everything reachable from oidx via cat_pipe as a WalkItem, - stopping whenever stop_at(oidx) returns true. Throw MissingObject - if a hash encountered is missing from the repository, and don't - read or return blob content in the data field unless include_data - is set. +def walk_object(get_ref, oidx, stop_at=None, include_data=None): + """Yield everything reachable from oidx via get_ref (which must behave + like CatPipe get) as a WalkItem, stopping whenever stop_at(oidx) + returns true. Throw MissingObject if a hash encountered is + missing from the repository, and don't read or return blob content + in the data field unless include_data is set. + """ # Maintain the pending stack on the heap to avoid stack overflow pending = [(oidx, [], [], None)] @@ -1322,7 +1340,7 @@ def walk_object(cat_pipe, oidx, data=None) continue - item_it = cat_pipe.get(oidx) + item_it = get_ref(oidx) get_oidx, typ, _ = next(item_it) if not get_oidx: raise MissingObject(oidx.decode('hex'))