##// END OF EJS Templates
phases: large rewrite on retract boundary...
phases: large rewrite on retract boundary The new code is still pure Python, so we still have room to going significantly faster. However its complexity of the complex part is `O(|[min_new_draft, tip]|)` instead of `O(|[min_draft, tip]|` which should help tremendously one repository with old draft (like mercurial-devel or mozilla-try). This is especially useful as the most common "retract boundary" operation happens when we commit/rewrite new drafts or when we push new draft to a non-publishing server. In this case, the smallest new_revs is very close to the tip and there is very few work to do. A few smaller optimisation could be done for these cases and will be introduced in later changesets. We still have iterate over large sets of roots, but this is already a great improvement for a very small amount of work. We gather information on the affected changeset as we go as we can put it to use in the next changesets. This extra data collection might slowdown the `register_new` case a bit, however for register_new, it should not really matters. The set of new nodes is either small, so the impact is negligible, or the set of new nodes is large, and the amount of work to do to had them will dominate the overhead the collecting information in `changed_revs`. As this new code compute the changes on the fly, it unlock other interesting improvement to be done in later changeset.

File last commit:

r50493:594fc56c default
r52302:2f39c7ae default
Show More
bdiff.py
108 lines | 2.6 KiB | text/x-python | PythonLexer
# bdiff.py - Python implementation of bdiff.c
#
# Copyright 2009 Olivia Mackall <olivia@selenic.com> and others
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
import difflib
import re
import struct
from typing import (
List,
Tuple,
)
def splitnewlines(text: bytes) -> List[bytes]:
'''like str.splitlines, but only split on newlines.'''
lines = [l + b'\n' for l in text.split(b'\n')]
if lines:
if lines[-1] == b'\n':
lines.pop()
else:
lines[-1] = lines[-1][:-1]
return lines
def _normalizeblocks(
a: List[bytes], b: List[bytes], blocks
) -> List[Tuple[int, int, int]]:
prev = None
r = []
for curr in blocks:
if prev is None:
prev = curr
continue
shift = 0
a1, b1, l1 = prev
a1end = a1 + l1
b1end = b1 + l1
a2, b2, l2 = curr
a2end = a2 + l2
b2end = b2 + l2
if a1end == a2:
while (
a1end + shift < a2end and a[a1end + shift] == b[b1end + shift]
):
shift += 1
elif b1end == b2:
while (
b1end + shift < b2end and a[a1end + shift] == b[b1end + shift]
):
shift += 1
r.append((a1, b1, l1 + shift))
prev = a2 + shift, b2 + shift, l2 - shift
if prev is not None:
r.append(prev)
return r
def bdiff(a: bytes, b: bytes) -> bytes:
a = bytes(a).splitlines(True)
b = bytes(b).splitlines(True)
if not a:
s = b"".join(b)
return s and (struct.pack(b">lll", 0, 0, len(s)) + s)
bin = []
p = [0]
for i in a:
p.append(p[-1] + len(i))
d = difflib.SequenceMatcher(None, a, b).get_matching_blocks()
d = _normalizeblocks(a, b, d)
la = 0
lb = 0
for am, bm, size in d:
s = b"".join(b[lb:bm])
if am > la or s:
bin.append(struct.pack(b">lll", p[la], p[am], len(s)) + s)
la = am + size
lb = bm + size
return b"".join(bin)
def blocks(a: bytes, b: bytes) -> List[Tuple[int, int, int, int]]:
an = splitnewlines(a)
bn = splitnewlines(b)
d = difflib.SequenceMatcher(None, an, bn).get_matching_blocks()
d = _normalizeblocks(an, bn, d)
return [(i, i + n, j, j + n) for (i, j, n) in d]
def fixws(text: bytes, allws: bool) -> bytes:
if allws:
text = re.sub(b'[ \t\r]+', b'', text)
else:
text = re.sub(b'[ \t\r]+', b' ', text)
text = text.replace(b' \n', b'\n')
return text