At http://bzr.arbash-meinel.com/plugins/groupcompress_rabin
------------------------------------------------------------
revno: 58
revision-id: [email protected]
parent: [email protected]
committer: John Arbash Meinel <[email protected]>
branch nick: groupcompress_rabin
timestamp: Fri 2009-02-27 14:18:47 -0600
message:
Allowing the source bytes to be longer than expected.
This makes a huge difference for extraction speed.
10s versus 45s. Versus 17s for the original groupcompress code.
Also, the compiled version in _groupcompress_c seems ~ the same speed as
the patch-delta.c version.
At the very least, the extra memory copy overhead negates any benefit.
=== modified file '_groupcompress_c.pyx'
--- a/_groupcompress_c.pyx 2009-02-27 18:21:04 +0000
+++ b/_groupcompress_c.pyx 2009-02-27 20:18:47 +0000
@@ -144,9 +144,10 @@
# make sure the orig file size matches what we expect
# XXX: gcc warns because data isn't defined as 'const'
size = get_delta_hdr_size(&data, top)
- if (size != source_size):
+ if (size > source_size):
# XXX: mismatched source size
return None
+ source_size = size
# now the result size
size = get_delta_hdr_size(&data, top)
=== modified file 'groupcompress.py'
--- a/groupcompress.py 2009-02-27 19:54:27 +0000
+++ b/groupcompress.py 2009-02-27 20:18:47 +0000
@@ -475,8 +475,11 @@
else:
# TODO: relax apply_delta so that it can allow source to be
# longer than expected
- chunks = [_groupcompress_c.apply_delta(
- plain[0:index_memo[3]], delta)]
+ bytes = _groupcompress_c.apply_delta(plain, delta)
+ if bytes is None:
+ import pdb; pdb.set_trace()
+ chunks = [bytes]
+ del bytes
if sha_strings(chunks) != sha1:
raise AssertionError('sha1 sum did not match')
yield ChunkedContentFactory(key, parents, sha1, chunks)
=== modified file 'patch-delta.c'
--- a/patch-delta.c 2009-02-27 17:32:04 +0000
+++ b/patch-delta.c 2009-02-27 20:18:47 +0000
@@ -27,8 +27,11 @@
/* make sure the orig file size matches what we expect */
size = get_delta_hdr_size(&data, top);
- if (size != src_size)
+ /* MOD: We allow a bigger source, assuming we only compressed
+ against the first bytes. */
+ if (size > src_size)
return NULL;
+ src_size = size;
/* now the result size */
size = get_delta_hdr_size(&data, top);
--
bazaar-commits mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/bazaar-commits
