undef/git-commit-vandalism
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

tree_entry_interesting: do basedir compare on wildcard patterns when possible

Browse Source 
Currently we treat "*.c" and "path/to/*.c" the same way. Which means
we check all possible paths in repo against "path/to/*.c". One could
see that "path/elsewhere/foo.c" obviously cannot match "path/to/*.c"
and we only need to check all paths _inside_ "path/to/" against that
pattern.

This patch checks the leading fixed part of a pathspec against base
directory and exit early if possible. We could even optimize further
in "path/to/something*.c" case (i.e. check the fixed part against
name_entry as well) but that's more complicated and probably does not
gain us much.

-O2 build on linux-2.6, without and with this patch respectively:

$ time git rev-list --quiet HEAD -- 'drivers/*.c'

real    1m9.484s
user    1m9.128s
sys     0m0.181s

$ time ~/w/git/git rev-list --quiet HEAD -- 'drivers/*.c'

real    0m15.710s
user    0m15.564s
sys     0m0.107s

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This commit is contained in:
Nguyễn Thái Ngọc Duy 2012-11-24 11:33:51 +07:00 committed by Junio C Hamano
parent 8c6abbcd27
commit c904cd89e4
1 changed files with 64 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

65
tree-walk.c
View File

@ -572,6 +572,54 @@ static int match_dir_prefix(const char *base,
return 0;
}
/*
* Perform matching on the leading non-wildcard part of
* pathspec. item->nowildcard_len must be greater than zero. Return
* non-zero if base is matched.
*/
static int match_wildcard_base(const struct pathspec_item *item,
const char *base, int baselen,
int *matched)
{
const char *match = item->match;
/* the wildcard part is not considered in this function */
int matchlen = item->nowildcard_len;
if (baselen) {
int dirlen;
/*
* Return early if base is longer than the
* non-wildcard part but it does not match.
*/
if (baselen >= matchlen) {
*matched = matchlen;
return !strncmp(base, match, matchlen);
}
dirlen = matchlen;
while (dirlen && match[dirlen - 1] != '/')
dirlen--;
/*
* Return early if base is shorter than the
* non-wildcard part but it does not match. Note that
* base ends with '/' so we are sure it really matches
* directory
*/
if (strncmp(base, match, baselen))
return 0;
*matched = baselen;
} else
*matched = 0;
/*
* we could have checked entry against the non-wildcard part
* that is not in base and does similar never_interesting
* optimization as in match_entry. For now just be happy with
* base comparison.
*/
return entry_interesting;
}
/*
* Is a tree entry interesting given the pathspec we have?
*
@ -602,7 +650,7 @@ enum interesting tree_entry_interesting(const struct name_entry *entry,
const struct pathspec_item *item = ps->items+i;
const char *match = item->match;
const char *base_str = base->buf + base_offset;
int matchlen = item->len;
int matchlen = item->len, matched = 0;
if (baselen >= matchlen) {
/* If it doesn't match, move along... */
@ -647,9 +695,24 @@ match_wildcards:
if (item->nowildcard_len == item->len)
continue;
if (item->nowildcard_len &&
!match_wildcard_base(item, base_str, baselen, &matched))
return entry_not_interesting;
/*
* Concatenate base and entry->path into one and do
* fnmatch() on it.
*
* While we could avoid concatenation in certain cases
* [1], which saves a memcpy and potentially a
* realloc, it turns out not worth it. Measurement on
* linux-2.6 does not show any clear improvements,
* partly because of the nowildcard_len optimization
* in git_fnmatch(). Avoid micro-optimizations here.
*
* [1] if match_wildcard_base() says the base
* directory is already matched, we only need to match
* the rest, which is shorter so _in theory_ faster.
*/
strbuf_add(base, entry->path, pathlen);