phases: avoid N² behavior in `advanceboundary`
We allowed duplicated entries in the deque, which each entry could potentially
insert all its ancestors. So advancing boundary for the full repository would
mean each revision would walk all its ancestors, resulting in O(N²) iteration.
For repository of any decent size, N² is quickly insane.
We introduce a simple set to avoid this and get back to reasonable performance.
# Import something from Mercurial, so the module loader gets initialized.
from mercurial import pycompat
del pycompat # unused for now
from hgext.lfs import pointer
def tryparse(text):
r = {}
try:
r = pointer.deserialize(text)
print('ok')
except Exception as ex:
print((b'%s' % ex).decode('ascii'))
if r:
text2 = r.serialize()
if text2 != text:
print('reconstructed text differs')
return r
t = (
b'version https://git-lfs.github.com/spec/v1\n'
b'oid sha256:4d7a214614ab2935c943f9e0ff69d22eadbb8f32b1'
b'258daaa5e2ca24d17e2393\n'
b'size 12345\n'
b'x-foo extra-information\n'
)
tryparse(b'')
tryparse(t)
tryparse(t.replace(b'git-lfs', b'unknown'))
tryparse(t.replace(b'v1\n', b'v1\n\n'))
tryparse(t.replace(b'sha256', b'ahs256'))
tryparse(t.replace(b'sha256:', b''))
tryparse(t.replace(b'12345', b'0x12345'))
tryparse(t.replace(b'extra-information', b'extra\0information'))
tryparse(t.replace(b'extra-information', b'extra\ninformation'))
tryparse(t.replace(b'x-foo', b'x_foo'))
tryparse(t.replace(b'oid', b'blobid'))
tryparse(t.replace(b'size', b'size-bytes').replace(b'oid', b'object-id'))