diff options
Diffstat (limited to 'lib/git/objects')
| -rw-r--r-- | lib/git/objects/base.py | 43 | ||||
| -rw-r--r-- | lib/git/objects/commit.py | 91 | ||||
| -rw-r--r-- | lib/git/objects/tag.py | 7 | ||||
| -rw-r--r-- | lib/git/objects/tree.py | 252 | ||||
| -rw-r--r-- | lib/git/objects/utils.py (renamed from lib/git/objects/util.py) | 0 |
5 files changed, 268 insertions, 125 deletions
diff --git a/lib/git/objects/base.py b/lib/git/objects/base.py index 43aa8dd1..07538ada 100644 --- a/lib/git/objects/base.py +++ b/lib/git/objects/base.py @@ -4,32 +4,10 @@ # This module is part of GitPython and is released under # the BSD License: http://www.opensource.org/licenses/bsd-license.php import os - -class LazyMixin(object): - lazy_properties = [] - __slots__ = tuple() +from git.utils import LazyMixin - def __getattr__(self, attr): - """ - Whenever an attribute is requested that we do not know, we allow it - to be created and set. Next time the same attribute is reqeusted, it is simply - returned from our dict/slots. - """ - self._set_cache_(attr) - # will raise in case the cache was not created - return object.__getattribute__(self, attr) +_assertion_msg_format = "Created object %r whose python type %r disagrees with the acutal git object type %r" - def _set_cache_(self, attr): - """ This method should be overridden in the derived class. - It should check whether the attribute named by attr can be created - and cached. Do nothing if you do not know the attribute or call your subclass - - The derived class may create as many additional attributes as it deems - necessary in case a git command returns more information than represented - in the single attribute.""" - pass - - class Object(LazyMixin): """ Implements an Object which may be Blobs, Trees, Commits and Tags @@ -71,9 +49,13 @@ class Object(LazyMixin): Retrieve object information """ if attr == "size": - self.size = int(self.repo.git.cat_file(self.id, s=True).rstrip()) + hexsha, typename, self.size = self.repo.git.get_object_header(self.id) + assert typename == self.type, _assertion_msg_format % (self.id, typename, self.type) elif attr == "data": - self.data = self.repo.git.cat_file(self.id, p=True, with_raw_output=True) + hexsha, typename, self.size, self.data = self.repo.git.get_object_data(self.id) + assert typename == self.type, _assertion_msg_format % (self.id, typename, self.type) + else: + super(Object,self)._set_cache_(attr) def __eq__(self, other): """ @@ -143,8 +125,15 @@ class IndexObject(Object): if isinstance(mode, basestring): self.mode = self._mode_str_to_int(mode) + def _set_cache_(self, attr): + if attr in IndexObject.__slots__: + # they cannot be retrieved lateron ( not without searching for them ) + raise AttributeError( "path and mode attributes must have been set during %s object creation" % type(self).__name__ ) + else: + super(IndexObject, self)._set_cache_(attr) + @classmethod - def _mode_str_to_int( cls, modestr ): + def _mode_str_to_int(cls, modestr): """ ``modestr`` string like 755 or 644 or 100644 - only the last 3 chars will be used diff --git a/lib/git/objects/commit.py b/lib/git/objects/commit.py index c3e97bf9..101014ab 100644 --- a/lib/git/objects/commit.py +++ b/lib/git/objects/commit.py @@ -6,14 +6,14 @@ import re import time - +from git.utils import Iterable from git.actor import Actor -from tree import Tree import git.diff as diff import git.stats as stats +from tree import Tree import base -class Commit(base.Object): +class Commit(base.Object, Iterable): """ Wraps a git Commit object. @@ -37,7 +37,7 @@ class Commit(base.Object): The parameter documentation indicates the type of the argument after a colon ':'. ``id`` - is the sha id of the commit + is the sha id of the commit or a ref ``parents`` : tuple( Commit, ... ) is a tuple of commit ids or actual Commits @@ -71,7 +71,7 @@ class Commit(base.Object): # END for each parent to convert if self.id and tree is not None: - self.tree = Tree(repo, id=tree) + self.tree = Tree(repo, id=tree, path='') # END id to tree conversion def _set_cache_(self, attr): @@ -80,8 +80,11 @@ class Commit(base.Object): to be set. We set all values at once. """ - if attr in self.__slots__: - temp = Commit.find_all(self.repo, self.id, max_count=1)[0] + if attr in Commit.__slots__: + # prepare our data lines to match rev-list + data_lines = self.data.splitlines() + data_lines.insert(0, "commit %s" % self.id) + temp = self._iter_from_process_or_stream(self.repo, iter(data_lines)).next() self.parents = temp.parents self.tree = temp.tree self.author = temp.author @@ -120,7 +123,7 @@ class Commit(base.Object): return len(repo.git.rev_list(ref, '--', path).strip().splitlines()) @classmethod - def find_all(cls, repo, ref, path='', **kwargs): + def iter_items(cls, repo, ref, path='', **kwargs): """ Find all commits matching the given criteria. @@ -128,7 +131,7 @@ class Commit(base.Object): is the Repo ``ref`` - is the ref from which to begin (SHA1 or name) + is the ref from which to begin (SHA1, Head or name) ``path`` is an optinal path, if set only Commits that include the path @@ -140,55 +143,67 @@ class Commit(base.Object): ``skip`` is the number of commits to skip Returns - git.Commit[] + iterator yielding Commit items """ - options = {'pretty': 'raw'} + options = {'pretty': 'raw', 'as_process' : True } options.update(kwargs) - output = repo.git.rev_list(ref, '--', path, **options) - return cls._list_from_string(repo, output) + # the test system might confront us with string values - + proc = repo.git.rev_list(ref, '--', path, **options) + return cls._iter_from_process_or_stream(repo, proc) @classmethod - def _list_from_string(cls, repo, text): + def _iter_from_process_or_stream(cls, repo, proc_or_stream): """ Parse out commit information into a list of Commit objects ``repo`` is the Repo - ``text`` - is the text output from the git-rev-list command (raw format) + ``proc`` + git-rev-list process instance (raw format) Returns - git.Commit[] + iterator returning Commit objects """ - lines =text.splitlines(False) - commits = [] - - while lines: - id = lines.pop(0).split()[1] - tree = lines.pop(0).split()[1] + stream = proc_or_stream + if not hasattr(stream,'next'): + stream = proc_or_stream.stdout + + for line in stream: + id = line.split()[1] + assert line.split()[0] == "commit" + tree = stream.next().split()[1] parents = [] - while lines and lines[0].startswith('parent'): - parents.append(lines.pop(0).split()[-1]) - # END while there are parent lines - author, authored_date = cls._actor(lines.pop(0)) - committer, committed_date = cls._actor(lines.pop(0)) + next_line = None + for parent_line in stream: + if not parent_line.startswith('parent'): + next_line = parent_line + break + # END abort reading parents + parents.append(parent_line.split()[-1]) + # END for each parent line + + author, authored_date = cls._actor(next_line) + committer, committed_date = cls._actor(stream.next()) - # free line - lines.pop(0) + # empty line + stream.next() message_lines = [] - while lines and not lines[0].startswith('commit'): - message_lines.append(lines.pop(0).strip()) + next_line = None + for msg_line in stream: + if not msg_line.startswith(' '): + break + # END abort message reading + message_lines.append(msg_line.strip()) # END while there are message lines - message = '\n'.join(message_lines[:-1]) # last line is empty - - commits.append(Commit(repo, id=id, parents=parents, tree=tree, author=author, authored_date=authored_date, - committer=committer, committed_date=committed_date, message=message)) - # END while lines - return commits + message = '\n'.join(message_lines) + + yield Commit(repo, id=id, parents=parents, tree=tree, author=author, authored_date=authored_date, + committer=committer, committed_date=committed_date, message=message) + # END for each line in stream @classmethod def diff(cls, repo, a, b=None, paths=None): diff --git a/lib/git/objects/tag.py b/lib/git/objects/tag.py index af1022f0..ecf6349d 100644 --- a/lib/git/objects/tag.py +++ b/lib/git/objects/tag.py @@ -8,7 +8,7 @@ Module containing all object based types. """ import base import commit -from util import get_object_type_by_name +from utils import get_object_type_by_name class TagObject(base.Object): """ @@ -48,9 +48,8 @@ class TagObject(base.Object): """ Cache all our attributes at once """ - if attr in self.__slots__: - output = self.repo.git.cat_file(self.type,self.id) - lines = output.split("\n") + if attr in TagObject.__slots__: + lines = self.data.splitlines() obj, hexsha = lines[0].split(" ") # object <hexsha> type_token, type_name = lines[1].split(" ") # type <type_name> diff --git a/lib/git/objects/tree.py b/lib/git/objects/tree.py index 273384a3..abfa9622 100644 --- a/lib/git/objects/tree.py +++ b/lib/git/objects/tree.py @@ -7,53 +7,125 @@ import os import blob import base +import binascii + +def sha_to_hex(sha): + """Takes a string and returns the hex of the sha within""" + hexsha = binascii.hexlify(sha) + assert len(hexsha) == 40, "Incorrect length of sha1 string: %d" % hexsha + return hexsha class Tree(base.IndexObject): + """ + Tress represent a ordered list of Blobs and other Trees. Hence it can be + accessed like a list. + + Tree's will cache their contents after first retrieval to improve efficiency. + + ``Tree as a list``:: + + Access a specific blob using the + tree['filename'] notation. + + You may as well access by index + blob = tree[0] + + + """ type = "tree" - __slots__ = "_contents" + __slots__ = "_cache" + + # using ascii codes for comparison + ascii_commit_id = (0x31 << 4) + 0x36 + ascii_blob_id = (0x31 << 4) + 0x30 + ascii_tree_id = (0x34 << 4) + 0x30 - def __init__(self, repo, id, mode=None, path=None): + + def __init__(self, repo, id, mode=0, path=None): super(Tree, self).__init__(repo, id, mode, path) def _set_cache_(self, attr): - if attr == "_contents": - # Read the tree contents. - self._contents = {} - for line in self.repo.git.ls_tree(self.id).splitlines(): - obj = self.content__from_string(self.repo, line) - if obj is not None: - self._contents[obj.path] = obj + if attr == "_cache": + # Set the data when we need it + self._cache = self._get_tree_cache() else: super(Tree, self)._set_cache_(attr) - @staticmethod - def content__from_string(repo, text): + def _get_tree_cache(self): """ - Parse a content item and create the appropriate object - - ``repo`` - is the Repo - - ``text`` - is the single line containing the items data in `git ls-tree` format - + Return + list(object_instance, ...) + + ``treeish`` + sha or ref identifying a tree + """ + out = list() + for obj in self._iter_from_data(): + if obj is not None: + out.append(obj) + # END if object was handled + # END for each line from ls-tree + return out + + + def _iter_from_data(self): + """ + Reads the binary non-pretty printed representation of a tree and converts + it into Blob, Tree or Commit objects. + + Note: This method was inspired by the parse_tree method in dulwich. + Returns - ``git.Blob`` or ``git.Tree`` + list(IndexObject, ...) """ - try: - mode, typ, id, path = text.expandtabs(1).split(" ", 3) - except: - return None + ord_zero = ord('0') + data = self.data + len_data = len(data) + i = 0 + while i < len_data: + mode = 0 + mode_boundary = i + 6 + + # keep it ascii - we compare against the respective values + type_id = (ord(data[i])<<4) + ord(data[i+1]) + i += 2 + + while data[i] != ' ': + # move existing mode integer up one level being 3 bits + # and add the actual ordinal value of the character + mode = (mode << 3) + (ord(data[i]) - ord_zero) + i += 1 + # END while reading mode + + # byte is space now, skip it + i += 1 + + # parse name, it is NULL separated + + ns = i + while data[i] != '\0': + i += 1 + # END while not reached NULL + name = data[ns:i] + + # byte is NULL, get next 20 + i += 1 + sha = data[i:i+20] + i = i + 20 + + hexsha = sha_to_hex(sha) + if type_id == self.ascii_blob_id: + yield blob.Blob(self.repo, hexsha, mode, name) + elif type_id == self.ascii_tree_id: + yield Tree(self.repo, hexsha, mode, name) + elif type_id == self.ascii_commit_id: + # todo + yield None + else: + raise TypeError( "Unknown type found in tree data: %i" % type_id ) + # END for each byte in data stream - if typ == "tree": - return Tree(repo, id, mode, path) - elif typ == "blob": - return blob.Blob(repo, id, mode, path) - elif typ == "commit": - return None - else: - raise(TypeError, "Invalid type: %s" % typ) def __div__(self, file): """ @@ -67,36 +139,104 @@ class Tree(base.IndexObject): <git.Blob "8b1e02c0fb554eed2ce2ef737a68bb369d7527df"> Returns - ``git.Blob`` or ``git.Tree`` or ``None`` if not found + ``git.Blob`` or ``git.Tree`` + + Raise + KeyError if given file or tree does not exist in tree """ - return self.get(file) + return self[file] def __repr__(self): return '<git.Tree "%s">' % self.id + + @classmethod + def _iter_recursive(cls, repo, tree, cur_depth, max_depth, predicate ): + + for obj in tree: + # adjust path to be complete + obj.path = os.path.join(tree.path, obj.path) + if not predicate(obj): + continue + yield obj + if obj.type == "tree" and ( max_depth < 0 or cur_depth+1 <= max_depth ): + for recursive_obj in cls._iter_recursive( repo, obj, cur_depth+1, max_depth, predicate ): + yield recursive_obj + # END for each recursive object + # END if we may enter recursion + # END for each object + + def traverse(self, max_depth=-1, predicate = lambda i: True): + """ + Returns + Iterator to traverse the tree recursively up to the given level. + The iterator returns Blob and Tree objects + + ``max_depth`` + + if -1, the whole tree will be traversed + if 0, only the first level will be traversed which is the same as + the default non-recursive iterator + + ``predicate`` + + If predicate(item) returns True, item will be returned by iterator + """ + return self._iter_recursive( self.repo, self, 0, max_depth, predicate ) + + @property + def trees(self): + """ + Returns + list(Tree, ...) list of trees directly below this tree + """ + return [ i for i in self if i.type == "tree" ] + + @property + def blobs(self): + """ + Returns + list(Blob, ...) list of blobs directly below this tree + """ + return [ i for i in self if i.type == "blob" ] - # Implement the basics of the dict protocol: - # directories/trees can be seen as object dicts. - def __getitem__(self, key): - return self._contents[key] + # List protocol + def __getslice__(self,i,j): + return self._cache[i:j] + def __iter__(self): - return iter(self._contents) - + return iter(self._cache) + def __len__(self): - return len(self._contents) - - def __contains__(self, key): - return key in self._contents - - def get(self, key): - return self._contents.get(key) - - def items(self): - return self._contents.items() - - def keys(self): - return self._contents.keys() - - def values(self): - return self._contents.values() + return len(self._cache) + + def __getitem__(self,item): + if isinstance(item, int): + return self._cache[item] + + if isinstance(item, basestring): + # compatability + for obj in self._cache: + if obj.path == item: + return obj + # END for each obj + raise KeyError( "Blob or Tree named %s not found" % item ) + # END index is basestring + + raise TypeError( "Invalid index type: %r" % item ) + + + def __contains__(self,item): + if isinstance(item, base.IndexObject): + return item in self._cache + + # compatability + for obj in self._cache: + if item == obj.path: + return True + # END for each item + return False + + def __reversed__(self): + return reversed(self._cache) diff --git a/lib/git/objects/util.py b/lib/git/objects/utils.py index 15c1d114..15c1d114 100644 --- a/lib/git/objects/util.py +++ b/lib/git/objects/utils.py |