def inventory_to_tree_and_blobs(repo, mapping, revision_id):
stack = []
cur = ""
tree = Tree()
inv = repo.get_inventory(revision_id)
for path, entry in inv.iter_entries():
while stack and not path.startswith(cur):
tree.serialize()
sha = tree.sha().hexdigest()
yield sha, tree
t = (stat.S_IFDIR, splitpath(cur)[-1:][0].encode('UTF-8'), sha)
cur, tree = stack.pop()
tree.add(*t)
if type(entry) == InventoryDirectory:
stack.append((cur, tree))
cur = path
tree = Tree()
if type(entry) == InventoryFile:
#FIXME: We can make potentially make this Lazy to avoid shaing lots of stuff
# and having all these objects in memory at once
blob = Blob()
_, blob._text = repo.iter_files_bytes([(entry.file_id, revision_id,
path)]).next()
sha = blob.sha().hexdigest()
yield sha, blob
name = splitpath(path)[-1:][0].encode('UTF-8')
mode = stat.S_IFREG | 0644
if entry.executable:
mode |= 0111
tree.add(mode, name, sha)
while len(stack) > 1:
tree.serialize()
sha = tree.sha().hexdigest()
yield sha, tree
t = (stat.S_IFDIR, splitpath(cur)[-1:][0].encode('UTF-8'), sha)
cur, tree = stack.pop()
tree.add(*t)
tree.serialize()
yield tree.sha().hexdigest(), tree
def inventory_to_tree_and_blobs(repo, mapping, revision_id):
stack = []
cur = ""
tree = Tree()
inv = repo.get_inventory(revision_id)
for path, entry in inv.iter_entries():
while stack and not path.startswith(cur):
tree.serialize()
sha = tree.sha().hexdigest()
yield sha, tree
t = (stat.S_IFDIR, splitpath(cur)[-1:][0].encode('UTF-8'), sha)
cur, tree = stack.pop()
tree.add(*t)
if type(entry) == InventoryDirectory:
stack.append((cur, tree))
cur = path
tree = Tree()
if type(entry) == InventoryFile:
#FIXME: We can make potentially make this Lazy to avoid shaing lots of stuff
# and having all these objects in memory at once
blob = Blob()
_, blob._text = repo.iter_files_bytes([(entry.file_id, revision_id, path)]).next()
sha = blob.sha().hexdigest()
yield sha, blob
name = splitpath(path)[-1:][0].encode('UTF-8')
mode = stat.S_IFREG | 0644
if entry.executable:
mode |= 0111
tree.add(mode, name, sha)
while len(stack) > 1:
tree.serialize()
sha = tree.sha().hexdigest()
yield sha, tree
t = (stat.S_IFDIR, splitpath(cur)[-1:][0].encode('UTF-8'), sha)
cur, tree = stack.pop()
tree.add(*t)
tree.serialize()
yield tree.sha().hexdigest(), tree
store = repo.object_store
print("Object store: {}".format(store))
store.add_object(blob)
print("Added one object")
# notes:
# - single object file in .git/objects now
# - "git cat-file -p <hash>": see our file!
# - file is compressed (hexdump -C .git/objects/*)
# now lets create a real "tree" object so we can give our Blob a
# filename
tree = Tree()
tree.add(b"a_file.txt", 0o100644, blob.id)
print("Tree: {}".format(tree.sha().hexdigest()))
# note: still not "in" the object repository
store.add_object(tree)
# From a user perpective, Trees live in Commit objects
commit = Commit()
commit.tree = tree
commit.author = commit.committer = b"meejah <*****@*****.**>"
commit.commit_time = commit.author_time = int(time()) # seconds since epoch
commit.commit_timezone = commit.author_timezone = -7 * (
60 * 60) # seconds offset; MST
commit.encoding = b"utf8"
commit.message = b"The beginning"
# no commit.parents because this is the first Commit
# again, not in the object store yet
