2006-01-07 09:33:54 +00:00
|
|
|
#include "cache.h"
|
2005-05-18 23:14:22 +00:00
|
|
|
#include "tag.h"
|
2005-04-18 18:39:48 +00:00
|
|
|
#include "commit.h"
|
|
|
|
|
[PATCH] Avoid wasting memory in git-rev-list
As pointed out on the list, git-rev-list can use a lot of memory.
One low-hanging fruit is to free the commit buffer for commits that we
parse. By default, parse_commit() will save away the buffer, since a lot
of cases do want it, and re-reading it continually would be unnecessary.
However, in many cases the buffer isn't actually necessary and saving it
just wastes memory.
We could just free the buffer ourselves, but especially in git-rev-list,
we actually end up using the helper functions that automatically add
parent commits to the commit lists, so we don't actually control the
commit parsing directly.
Instead, just make this behaviour of "parse_commit()" a global flag.
Maybe this is a bit tasteless, but it's very simple, and it makes a
noticable difference in memory usage.
Before the change:
[torvalds@g5 linux]$ /usr/bin/time git-rev-list v2.6.12..HEAD > /dev/null
0.26user 0.02system 0:00.28elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+3714minor)pagefaults 0swaps
after the change:
[torvalds@g5 linux]$ /usr/bin/time git-rev-list v2.6.12..HEAD > /dev/null
0.26user 0.00system 0:00.27elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2433minor)pagefaults 0swaps
note how the minor faults have decreased from 3714 pages to 2433 pages.
That's all due to the fewer anonymous pages allocated to hold the comment
buffers and their metadata.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-09-15 21:43:17 +00:00
|
|
|
int save_commit_buffer = 1;
|
|
|
|
|
2005-07-06 16:39:34 +00:00
|
|
|
struct sort_node
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* the number of children of the associated commit
|
|
|
|
* that also occur in the list being sorted.
|
|
|
|
*/
|
|
|
|
unsigned int indegree;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* reference to original list item that we will re-use
|
|
|
|
* on output.
|
|
|
|
*/
|
|
|
|
struct commit_list * list_item;
|
|
|
|
|
|
|
|
};
|
|
|
|
|
2005-04-18 18:39:48 +00:00
|
|
|
const char *commit_type = "commit";
|
|
|
|
|
2006-05-15 00:20:46 +00:00
|
|
|
struct cmt_fmt_map {
|
|
|
|
const char *n;
|
|
|
|
size_t cmp_len;
|
|
|
|
enum cmit_fmt v;
|
|
|
|
} cmt_fmts[] = {
|
|
|
|
{ "raw", 1, CMIT_FMT_RAW },
|
|
|
|
{ "medium", 1, CMIT_FMT_MEDIUM },
|
|
|
|
{ "short", 1, CMIT_FMT_SHORT },
|
2006-05-21 08:34:54 +00:00
|
|
|
{ "email", 1, CMIT_FMT_EMAIL },
|
2006-05-15 00:20:46 +00:00
|
|
|
{ "full", 5, CMIT_FMT_FULL },
|
|
|
|
{ "fuller", 5, CMIT_FMT_FULLER },
|
|
|
|
{ "oneline", 1, CMIT_FMT_ONELINE },
|
|
|
|
};
|
|
|
|
|
2005-06-27 00:50:46 +00:00
|
|
|
enum cmit_fmt get_commit_format(const char *arg)
|
|
|
|
{
|
2006-05-15 00:20:46 +00:00
|
|
|
int i;
|
|
|
|
|
|
|
|
if (!arg || !*arg)
|
2005-06-27 00:50:46 +00:00
|
|
|
return CMIT_FMT_DEFAULT;
|
2006-05-15 00:20:46 +00:00
|
|
|
if (*arg == '=')
|
|
|
|
arg++;
|
|
|
|
for (i = 0; i < ARRAY_SIZE(cmt_fmts); i++) {
|
|
|
|
if (!strncmp(arg, cmt_fmts[i].n, cmt_fmts[i].cmp_len))
|
|
|
|
return cmt_fmts[i].v;
|
|
|
|
}
|
|
|
|
|
|
|
|
die("invalid --pretty format: %s", arg);
|
2005-06-27 00:50:46 +00:00
|
|
|
}
|
|
|
|
|
2005-08-21 09:51:10 +00:00
|
|
|
static struct commit *check_commit(struct object *obj,
|
|
|
|
const unsigned char *sha1,
|
|
|
|
int quiet)
|
2005-05-18 23:14:22 +00:00
|
|
|
{
|
Shrink "struct object" a bit
This shrinks "struct object" by a small amount, by getting rid of the
"struct type *" pointer and replacing it with a 3-bit bitfield instead.
In addition, we merge the bitfields and the "flags" field, which
incidentally should also remove a useless 4-byte padding from the object
when in 64-bit mode.
Now, our "struct object" is still too damn large, but it's now less
obviously bloated, and of the remaining fields, only the "util" (which is
not used by most things) is clearly something that should be eventually
discarded.
This shrinks the "git-rev-list --all" memory use by about 2.5% on the
kernel archive (and, perhaps more importantly, on the larger mozilla
archive). That may not sound like much, but I suspect it's more on a
64-bit platform.
There are other remaining inefficiencies (the parent lists, for example,
probably have horrible malloc overhead), but this was pretty obvious.
Most of the patch is just changing the comparison of the "type" pointer
from one of the constant string pointers to the appropriate new TYPE_xxx
small integer constant.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-14 23:45:13 +00:00
|
|
|
if (obj->type != TYPE_COMMIT) {
|
2005-08-21 09:51:10 +00:00
|
|
|
if (!quiet)
|
|
|
|
error("Object %s is a %s, not a commit",
|
Shrink "struct object" a bit
This shrinks "struct object" by a small amount, by getting rid of the
"struct type *" pointer and replacing it with a 3-bit bitfield instead.
In addition, we merge the bitfields and the "flags" field, which
incidentally should also remove a useless 4-byte padding from the object
when in 64-bit mode.
Now, our "struct object" is still too damn large, but it's now less
obviously bloated, and of the remaining fields, only the "util" (which is
not used by most things) is clearly something that should be eventually
discarded.
This shrinks the "git-rev-list --all" memory use by about 2.5% on the
kernel archive (and, perhaps more importantly, on the larger mozilla
archive). That may not sound like much, but I suspect it's more on a
64-bit platform.
There are other remaining inefficiencies (the parent lists, for example,
probably have horrible malloc overhead), but this was pretty obvious.
Most of the patch is just changing the comparison of the "type" pointer
from one of the constant string pointers to the appropriate new TYPE_xxx
small integer constant.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-14 23:45:13 +00:00
|
|
|
sha1_to_hex(sha1), typename(obj->type));
|
2005-05-18 23:14:22 +00:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
return (struct commit *) obj;
|
|
|
|
}
|
|
|
|
|
2005-08-21 09:51:10 +00:00
|
|
|
struct commit *lookup_commit_reference_gently(const unsigned char *sha1,
|
|
|
|
int quiet)
|
2005-05-18 23:14:22 +00:00
|
|
|
{
|
2005-11-02 23:19:13 +00:00
|
|
|
struct object *obj = deref_tag(parse_object(sha1), NULL, 0);
|
2005-05-18 23:14:22 +00:00
|
|
|
|
|
|
|
if (!obj)
|
|
|
|
return NULL;
|
2005-08-21 09:51:10 +00:00
|
|
|
return check_commit(obj, sha1, quiet);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct commit *lookup_commit_reference(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
return lookup_commit_reference_gently(sha1, 0);
|
2005-05-18 23:14:22 +00:00
|
|
|
}
|
|
|
|
|
2005-06-03 15:05:39 +00:00
|
|
|
struct commit *lookup_commit(const unsigned char *sha1)
|
2005-04-18 18:39:48 +00:00
|
|
|
{
|
|
|
|
struct object *obj = lookup_object(sha1);
|
|
|
|
if (!obj) {
|
Add specialized object allocator
This creates a simple specialized object allocator for basic
objects.
This avoids wasting space with malloc overhead (metadata and
extra alignment), since the specialized allocator knows the
alignment, and that objects, once allocated, are never freed.
It also allows us to track some basic statistics about object
allocations. For example, for the mozilla import, it shows
object usage as follows:
blobs: 627629 (14710 kB)
trees: 1119035 (34969 kB)
commits: 196423 (8440 kB)
tags: 1336 (46 kB)
and the simpler allocator shaves off about 2.5% off the memory
footprint off a "git-rev-list --all --objects", and is a bit
faster too.
[ Side note: this concludes the series of "save memory in object storage".
The thing is, there simply isn't much more to be saved on the objects.
Doing "git-rev-list --all --objects" on the mozilla archive has a final
total RSS of 131498 pages for me: that's about 513MB. Of that, the
object overhead is now just 56MB, the rest is going somewhere else (put
another way: the fact that this patch shaves off 2.5% of the total
memory overhead, considering that objects are now not much more than 10%
of the total shows how big the wasted space really was: this makes
object allocations much more memory- and time-efficient).
I haven't looked at where the rest is, but I suspect the bulk of it is
just the pack-file loading. It may be that we should pack the tree
objects separately from the blob objects: for git-rev-list --objects, we
don't actually ever need to even look at the blobs, but since trees and
blobs are interspersed in the pack-file, we end up not being dense in
the tree accesses, so we end up looking at more pages than we strictly
need to.
So with a 535MB pack-file, it's entirely possible - even likely - that
most of the remaining RSS is just the mmap of the pack-file itself. We
don't need to map in _all_ of it, but we do end up mapping a fair
amount. ]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-19 17:44:15 +00:00
|
|
|
struct commit *ret = alloc_commit_node();
|
2005-04-18 18:39:48 +00:00
|
|
|
created_object(sha1, &ret->object);
|
Shrink "struct object" a bit
This shrinks "struct object" by a small amount, by getting rid of the
"struct type *" pointer and replacing it with a 3-bit bitfield instead.
In addition, we merge the bitfields and the "flags" field, which
incidentally should also remove a useless 4-byte padding from the object
when in 64-bit mode.
Now, our "struct object" is still too damn large, but it's now less
obviously bloated, and of the remaining fields, only the "util" (which is
not used by most things) is clearly something that should be eventually
discarded.
This shrinks the "git-rev-list --all" memory use by about 2.5% on the
kernel archive (and, perhaps more importantly, on the larger mozilla
archive). That may not sound like much, but I suspect it's more on a
64-bit platform.
There are other remaining inefficiencies (the parent lists, for example,
probably have horrible malloc overhead), but this was pretty obvious.
Most of the patch is just changing the comparison of the "type" pointer
from one of the constant string pointers to the appropriate new TYPE_xxx
small integer constant.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-14 23:45:13 +00:00
|
|
|
ret->object.type = TYPE_COMMIT;
|
2005-04-18 18:39:48 +00:00
|
|
|
return ret;
|
|
|
|
}
|
2005-05-20 20:59:17 +00:00
|
|
|
if (!obj->type)
|
Shrink "struct object" a bit
This shrinks "struct object" by a small amount, by getting rid of the
"struct type *" pointer and replacing it with a 3-bit bitfield instead.
In addition, we merge the bitfields and the "flags" field, which
incidentally should also remove a useless 4-byte padding from the object
when in 64-bit mode.
Now, our "struct object" is still too damn large, but it's now less
obviously bloated, and of the remaining fields, only the "util" (which is
not used by most things) is clearly something that should be eventually
discarded.
This shrinks the "git-rev-list --all" memory use by about 2.5% on the
kernel archive (and, perhaps more importantly, on the larger mozilla
archive). That may not sound like much, but I suspect it's more on a
64-bit platform.
There are other remaining inefficiencies (the parent lists, for example,
probably have horrible malloc overhead), but this was pretty obvious.
Most of the patch is just changing the comparison of the "type" pointer
from one of the constant string pointers to the appropriate new TYPE_xxx
small integer constant.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-06-14 23:45:13 +00:00
|
|
|
obj->type = TYPE_COMMIT;
|
2005-08-21 09:51:10 +00:00
|
|
|
return check_commit(obj, sha1, 0);
|
2005-04-18 18:39:48 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long parse_commit_date(const char *buf)
|
|
|
|
{
|
|
|
|
unsigned long date;
|
|
|
|
|
|
|
|
if (memcmp(buf, "author", 6))
|
|
|
|
return 0;
|
|
|
|
while (*buf++ != '\n')
|
|
|
|
/* nada */;
|
|
|
|
if (memcmp(buf, "committer", 9))
|
|
|
|
return 0;
|
|
|
|
while (*buf++ != '>')
|
|
|
|
/* nada */;
|
|
|
|
date = strtoul(buf, NULL, 10);
|
|
|
|
if (date == ULONG_MAX)
|
|
|
|
date = 0;
|
|
|
|
return date;
|
|
|
|
}
|
|
|
|
|
2006-04-07 06:58:51 +00:00
|
|
|
static struct commit_graft **commit_graft;
|
2005-07-30 07:58:28 +00:00
|
|
|
static int commit_graft_alloc, commit_graft_nr;
|
|
|
|
|
|
|
|
static int commit_graft_pos(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
int lo, hi;
|
|
|
|
lo = 0;
|
|
|
|
hi = commit_graft_nr;
|
|
|
|
while (lo < hi) {
|
|
|
|
int mi = (lo + hi) / 2;
|
|
|
|
struct commit_graft *graft = commit_graft[mi];
|
|
|
|
int cmp = memcmp(sha1, graft->sha1, 20);
|
|
|
|
if (!cmp)
|
|
|
|
return mi;
|
|
|
|
if (cmp < 0)
|
|
|
|
hi = mi;
|
|
|
|
else
|
|
|
|
lo = mi + 1;
|
|
|
|
}
|
|
|
|
return -lo - 1;
|
|
|
|
}
|
|
|
|
|
2006-04-07 06:58:51 +00:00
|
|
|
int register_commit_graft(struct commit_graft *graft, int ignore_dups)
|
|
|
|
{
|
|
|
|
int pos = commit_graft_pos(graft->sha1);
|
|
|
|
|
|
|
|
if (0 <= pos) {
|
|
|
|
if (ignore_dups)
|
|
|
|
free(graft);
|
|
|
|
else {
|
|
|
|
free(commit_graft[pos]);
|
|
|
|
commit_graft[pos] = graft;
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
pos = -pos - 1;
|
|
|
|
if (commit_graft_alloc <= ++commit_graft_nr) {
|
|
|
|
commit_graft_alloc = alloc_nr(commit_graft_alloc);
|
|
|
|
commit_graft = xrealloc(commit_graft,
|
|
|
|
sizeof(*commit_graft) *
|
|
|
|
commit_graft_alloc);
|
|
|
|
}
|
|
|
|
if (pos < commit_graft_nr)
|
|
|
|
memmove(commit_graft + pos + 1,
|
|
|
|
commit_graft + pos,
|
|
|
|
(commit_graft_nr - pos - 1) *
|
|
|
|
sizeof(*commit_graft));
|
|
|
|
commit_graft[pos] = graft;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct commit_graft *read_graft_line(char *buf, int len)
|
|
|
|
{
|
|
|
|
/* The format is just "Commit Parent1 Parent2 ...\n" */
|
|
|
|
int i;
|
|
|
|
struct commit_graft *graft = NULL;
|
|
|
|
|
|
|
|
if (buf[len-1] == '\n')
|
|
|
|
buf[--len] = 0;
|
2006-04-17 11:41:49 +00:00
|
|
|
if (buf[0] == '#' || buf[0] == '\0')
|
2006-04-16 21:24:56 +00:00
|
|
|
return NULL;
|
2006-04-07 06:58:51 +00:00
|
|
|
if ((len + 1) % 41) {
|
|
|
|
bad_graft_data:
|
|
|
|
error("bad graft data: %s", buf);
|
|
|
|
free(graft);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
i = (len + 1) / 41 - 1;
|
|
|
|
graft = xmalloc(sizeof(*graft) + 20 * i);
|
|
|
|
graft->nr_parent = i;
|
|
|
|
if (get_sha1_hex(buf, graft->sha1))
|
|
|
|
goto bad_graft_data;
|
|
|
|
for (i = 40; i < len; i += 41) {
|
|
|
|
if (buf[i] != ' ')
|
|
|
|
goto bad_graft_data;
|
|
|
|
if (get_sha1_hex(buf + i + 1, graft->parent[i/41]))
|
|
|
|
goto bad_graft_data;
|
|
|
|
}
|
|
|
|
return graft;
|
|
|
|
}
|
|
|
|
|
|
|
|
int read_graft_file(const char *graft_file)
|
2005-07-30 07:58:28 +00:00
|
|
|
{
|
|
|
|
FILE *fp = fopen(graft_file, "r");
|
|
|
|
char buf[1024];
|
2006-04-07 06:58:51 +00:00
|
|
|
if (!fp)
|
|
|
|
return -1;
|
2005-07-30 07:58:28 +00:00
|
|
|
while (fgets(buf, sizeof(buf), fp)) {
|
|
|
|
/* The format is just "Commit Parent1 Parent2 ...\n" */
|
|
|
|
int len = strlen(buf);
|
2006-04-07 06:58:51 +00:00
|
|
|
struct commit_graft *graft = read_graft_line(buf, len);
|
2006-04-16 21:24:56 +00:00
|
|
|
if (!graft)
|
|
|
|
continue;
|
2006-04-07 06:58:51 +00:00
|
|
|
if (register_commit_graft(graft, 1))
|
2005-07-30 07:58:28 +00:00
|
|
|
error("duplicate graft data: %s", buf);
|
|
|
|
}
|
|
|
|
fclose(fp);
|
2006-04-07 06:58:51 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void prepare_commit_graft(void)
|
|
|
|
{
|
|
|
|
static int commit_graft_prepared;
|
|
|
|
char *graft_file;
|
|
|
|
|
|
|
|
if (commit_graft_prepared)
|
|
|
|
return;
|
|
|
|
graft_file = get_graft_file();
|
|
|
|
read_graft_file(graft_file);
|
|
|
|
commit_graft_prepared = 1;
|
2005-07-30 07:58:28 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static struct commit_graft *lookup_commit_graft(const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
int pos;
|
2006-04-07 06:58:51 +00:00
|
|
|
prepare_commit_graft();
|
2005-07-30 07:58:28 +00:00
|
|
|
pos = commit_graft_pos(sha1);
|
|
|
|
if (pos < 0)
|
|
|
|
return NULL;
|
|
|
|
return commit_graft[pos];
|
|
|
|
}
|
|
|
|
|
2005-05-06 17:48:34 +00:00
|
|
|
int parse_commit_buffer(struct commit *item, void *buffer, unsigned long size)
|
2005-04-18 18:39:48 +00:00
|
|
|
{
|
2005-07-27 22:12:48 +00:00
|
|
|
char *bufptr = buffer;
|
2005-04-18 18:39:48 +00:00
|
|
|
unsigned char parent[20];
|
2005-06-21 03:26:03 +00:00
|
|
|
struct commit_list **pptr;
|
2005-07-30 07:58:28 +00:00
|
|
|
struct commit_graft *graft;
|
2005-11-17 05:32:44 +00:00
|
|
|
unsigned n_refs = 0;
|
2005-05-06 17:48:34 +00:00
|
|
|
|
2005-04-18 18:39:48 +00:00
|
|
|
if (item->object.parsed)
|
|
|
|
return 0;
|
|
|
|
item->object.parsed = 1;
|
2005-07-27 22:12:48 +00:00
|
|
|
if (memcmp(bufptr, "tree ", 5))
|
|
|
|
return error("bogus commit object %s", sha1_to_hex(item->object.sha1));
|
|
|
|
if (get_sha1_hex(bufptr + 5, parent) < 0)
|
2006-02-23 01:47:10 +00:00
|
|
|
return error("bad tree pointer in commit %s",
|
|
|
|
sha1_to_hex(item->object.sha1));
|
2005-04-18 18:39:48 +00:00
|
|
|
item->tree = lookup_tree(parent);
|
2005-04-24 21:31:57 +00:00
|
|
|
if (item->tree)
|
2005-11-17 05:32:44 +00:00
|
|
|
n_refs++;
|
2005-04-18 18:39:48 +00:00
|
|
|
bufptr += 46; /* "tree " + "hex sha1" + "\n" */
|
2005-06-21 03:26:03 +00:00
|
|
|
pptr = &item->parents;
|
2005-07-30 07:58:28 +00:00
|
|
|
|
|
|
|
graft = lookup_commit_graft(item->object.sha1);
|
2005-07-27 22:12:48 +00:00
|
|
|
while (!memcmp(bufptr, "parent ", 7)) {
|
|
|
|
struct commit *new_parent;
|
|
|
|
|
|
|
|
if (get_sha1_hex(bufptr + 7, parent) || bufptr[47] != '\n')
|
|
|
|
return error("bad parents in commit %s", sha1_to_hex(item->object.sha1));
|
2005-07-30 07:58:28 +00:00
|
|
|
bufptr += 48;
|
|
|
|
if (graft)
|
|
|
|
continue;
|
2005-07-27 22:12:48 +00:00
|
|
|
new_parent = lookup_commit(parent);
|
2005-04-24 21:31:57 +00:00
|
|
|
if (new_parent) {
|
2005-06-21 03:26:03 +00:00
|
|
|
pptr = &commit_list_insert(new_parent, pptr)->next;
|
2005-11-17 05:32:44 +00:00
|
|
|
n_refs++;
|
2005-04-24 21:31:57 +00:00
|
|
|
}
|
2005-07-30 07:58:28 +00:00
|
|
|
}
|
|
|
|
if (graft) {
|
|
|
|
int i;
|
|
|
|
struct commit *new_parent;
|
|
|
|
for (i = 0; i < graft->nr_parent; i++) {
|
|
|
|
new_parent = lookup_commit(graft->parent[i]);
|
|
|
|
if (!new_parent)
|
|
|
|
continue;
|
|
|
|
pptr = &commit_list_insert(new_parent, pptr)->next;
|
2005-11-17 05:32:44 +00:00
|
|
|
n_refs++;
|
2005-07-30 07:58:28 +00:00
|
|
|
}
|
2005-04-18 18:39:48 +00:00
|
|
|
}
|
|
|
|
item->date = parse_commit_date(bufptr);
|
2005-11-17 05:32:44 +00:00
|
|
|
|
|
|
|
if (track_object_refs) {
|
|
|
|
unsigned i = 0;
|
|
|
|
struct commit_list *p;
|
|
|
|
struct object_refs *refs = alloc_object_refs(n_refs);
|
|
|
|
if (item->tree)
|
|
|
|
refs->ref[i++] = &item->tree->object;
|
|
|
|
for (p = item->parents; p; p = p->next)
|
|
|
|
refs->ref[i++] = &p->item->object;
|
|
|
|
set_object_refs(&item->object, refs);
|
|
|
|
}
|
|
|
|
|
2005-04-18 18:39:48 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-05-06 17:48:34 +00:00
|
|
|
int parse_commit(struct commit *item)
|
|
|
|
{
|
|
|
|
char type[20];
|
|
|
|
void *buffer;
|
|
|
|
unsigned long size;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (item->object.parsed)
|
|
|
|
return 0;
|
|
|
|
buffer = read_sha1_file(item->object.sha1, type, &size);
|
|
|
|
if (!buffer)
|
|
|
|
return error("Could not read %s",
|
|
|
|
sha1_to_hex(item->object.sha1));
|
|
|
|
if (strcmp(type, commit_type)) {
|
|
|
|
free(buffer);
|
|
|
|
return error("Object %s not a commit",
|
|
|
|
sha1_to_hex(item->object.sha1));
|
|
|
|
}
|
|
|
|
ret = parse_commit_buffer(item, buffer, size);
|
[PATCH] Avoid wasting memory in git-rev-list
As pointed out on the list, git-rev-list can use a lot of memory.
One low-hanging fruit is to free the commit buffer for commits that we
parse. By default, parse_commit() will save away the buffer, since a lot
of cases do want it, and re-reading it continually would be unnecessary.
However, in many cases the buffer isn't actually necessary and saving it
just wastes memory.
We could just free the buffer ourselves, but especially in git-rev-list,
we actually end up using the helper functions that automatically add
parent commits to the commit lists, so we don't actually control the
commit parsing directly.
Instead, just make this behaviour of "parse_commit()" a global flag.
Maybe this is a bit tasteless, but it's very simple, and it makes a
noticable difference in memory usage.
Before the change:
[torvalds@g5 linux]$ /usr/bin/time git-rev-list v2.6.12..HEAD > /dev/null
0.26user 0.02system 0:00.28elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+3714minor)pagefaults 0swaps
after the change:
[torvalds@g5 linux]$ /usr/bin/time git-rev-list v2.6.12..HEAD > /dev/null
0.26user 0.00system 0:00.27elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (0major+2433minor)pagefaults 0swaps
note how the minor faults have decreased from 3714 pages to 2433 pages.
That's all due to the fewer anonymous pages allocated to hold the comment
buffers and their metadata.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-09-15 21:43:17 +00:00
|
|
|
if (save_commit_buffer && !ret) {
|
2005-05-26 01:27:14 +00:00
|
|
|
item->buffer = buffer;
|
|
|
|
return 0;
|
|
|
|
}
|
2005-05-06 17:48:34 +00:00
|
|
|
free(buffer);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2005-05-31 01:44:02 +00:00
|
|
|
struct commit_list *commit_list_insert(struct commit *item, struct commit_list **list_p)
|
2005-04-24 01:47:23 +00:00
|
|
|
{
|
2005-04-26 19:00:58 +00:00
|
|
|
struct commit_list *new_list = xmalloc(sizeof(struct commit_list));
|
2005-04-24 01:47:23 +00:00
|
|
|
new_list->item = item;
|
|
|
|
new_list->next = *list_p;
|
|
|
|
*list_p = new_list;
|
2005-05-31 01:44:02 +00:00
|
|
|
return new_list;
|
2005-04-24 01:47:23 +00:00
|
|
|
}
|
|
|
|
|
2005-04-18 18:39:48 +00:00
|
|
|
void free_commit_list(struct commit_list *list)
|
|
|
|
{
|
|
|
|
while (list) {
|
|
|
|
struct commit_list *temp = list;
|
|
|
|
list = temp->next;
|
|
|
|
free(temp);
|
|
|
|
}
|
|
|
|
}
|
2005-04-24 01:47:23 +00:00
|
|
|
|
2005-07-06 16:31:17 +00:00
|
|
|
struct commit_list * insert_by_date(struct commit *item, struct commit_list **list)
|
2005-04-24 01:47:23 +00:00
|
|
|
{
|
|
|
|
struct commit_list **pp = list;
|
|
|
|
struct commit_list *p;
|
|
|
|
while ((p = *pp) != NULL) {
|
|
|
|
if (p->item->date < item->date) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
pp = &p->next;
|
|
|
|
}
|
2005-07-06 16:31:17 +00:00
|
|
|
return commit_list_insert(item, pp);
|
2005-04-24 01:47:23 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void sort_by_date(struct commit_list **list)
|
|
|
|
{
|
|
|
|
struct commit_list *ret = NULL;
|
|
|
|
while (*list) {
|
2005-07-06 16:31:17 +00:00
|
|
|
insert_by_date((*list)->item, &ret);
|
2005-04-24 01:47:23 +00:00
|
|
|
*list = (*list)->next;
|
|
|
|
}
|
|
|
|
*list = ret;
|
|
|
|
}
|
|
|
|
|
2005-04-24 03:29:22 +00:00
|
|
|
struct commit *pop_most_recent_commit(struct commit_list **list,
|
|
|
|
unsigned int mark)
|
2005-04-24 01:47:23 +00:00
|
|
|
{
|
|
|
|
struct commit *ret = (*list)->item;
|
|
|
|
struct commit_list *parents = ret->parents;
|
|
|
|
struct commit_list *old = *list;
|
|
|
|
|
|
|
|
*list = (*list)->next;
|
|
|
|
free(old);
|
|
|
|
|
|
|
|
while (parents) {
|
2005-04-24 02:21:28 +00:00
|
|
|
struct commit *commit = parents->item;
|
2005-04-24 03:29:22 +00:00
|
|
|
parse_commit(commit);
|
|
|
|
if (!(commit->object.flags & mark)) {
|
|
|
|
commit->object.flags |= mark;
|
2005-07-06 16:31:17 +00:00
|
|
|
insert_by_date(commit, list);
|
2005-04-24 02:21:28 +00:00
|
|
|
}
|
2005-04-24 01:47:23 +00:00
|
|
|
parents = parents->next;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
2005-06-01 15:34:23 +00:00
|
|
|
|
2006-01-08 02:52:42 +00:00
|
|
|
void clear_commit_marks(struct commit *commit, unsigned int mark)
|
|
|
|
{
|
|
|
|
struct commit_list *parents;
|
|
|
|
|
|
|
|
parents = commit->parents;
|
|
|
|
commit->object.flags &= ~mark;
|
|
|
|
while (parents) {
|
2006-01-16 06:15:37 +00:00
|
|
|
struct commit *parent = parents->item;
|
|
|
|
if (parent && parent->object.parsed &&
|
|
|
|
(parent->object.flags & mark))
|
|
|
|
clear_commit_marks(parent, mark);
|
2006-01-08 02:52:42 +00:00
|
|
|
parents = parents->next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-06-01 15:34:23 +00:00
|
|
|
/*
|
|
|
|
* Generic support for pretty-printing the header
|
|
|
|
*/
|
|
|
|
static int get_one_line(const char *msg, unsigned long len)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
while (len--) {
|
|
|
|
char c = *msg++;
|
2006-04-12 18:31:23 +00:00
|
|
|
if (!c)
|
|
|
|
break;
|
2005-06-01 15:34:23 +00:00
|
|
|
ret++;
|
|
|
|
if (c == '\n')
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2006-05-16 09:29:42 +00:00
|
|
|
static int is_rfc2047_special(char ch)
|
|
|
|
{
|
|
|
|
return ((ch & 0x80) || (ch == '=') || (ch == '?') || (ch == '_'));
|
|
|
|
}
|
|
|
|
|
|
|
|
static int add_rfc2047(char *buf, const char *line, int len)
|
|
|
|
{
|
|
|
|
char *bp = buf;
|
|
|
|
int i, needquote;
|
|
|
|
static const char q_utf8[] = "=?utf-8?q?";
|
|
|
|
|
|
|
|
for (i = needquote = 0; !needquote && i < len; i++) {
|
|
|
|
unsigned ch = line[i];
|
|
|
|
if (ch & 0x80)
|
|
|
|
needquote++;
|
|
|
|
if ((i + 1 < len) &&
|
|
|
|
(ch == '=' && line[i+1] == '?'))
|
|
|
|
needquote++;
|
|
|
|
}
|
|
|
|
if (!needquote)
|
|
|
|
return sprintf(buf, "%.*s", len, line);
|
|
|
|
|
|
|
|
memcpy(bp, q_utf8, sizeof(q_utf8)-1);
|
|
|
|
bp += sizeof(q_utf8)-1;
|
|
|
|
for (i = 0; i < len; i++) {
|
2006-06-19 22:00:17 +00:00
|
|
|
unsigned ch = line[i] & 0xFF;
|
2006-05-16 09:29:42 +00:00
|
|
|
if (is_rfc2047_special(ch)) {
|
|
|
|
sprintf(bp, "=%02X", ch);
|
|
|
|
bp += 3;
|
|
|
|
}
|
|
|
|
else if (ch == ' ')
|
|
|
|
*bp++ = '_';
|
|
|
|
else
|
|
|
|
*bp++ = ch;
|
|
|
|
}
|
|
|
|
memcpy(bp, "?=", 2);
|
|
|
|
bp += 2;
|
|
|
|
return bp - buf;
|
|
|
|
}
|
|
|
|
|
2005-06-27 00:50:46 +00:00
|
|
|
static int add_user_info(const char *what, enum cmit_fmt fmt, char *buf, const char *line)
|
2005-06-01 15:34:23 +00:00
|
|
|
{
|
|
|
|
char *date;
|
2005-08-30 04:17:21 +00:00
|
|
|
int namelen;
|
2005-06-01 15:34:23 +00:00
|
|
|
unsigned long time;
|
2005-06-05 16:02:03 +00:00
|
|
|
int tz, ret;
|
2005-11-10 06:15:27 +00:00
|
|
|
const char *filler = " ";
|
2005-06-01 15:34:23 +00:00
|
|
|
|
2005-08-09 05:15:40 +00:00
|
|
|
if (fmt == CMIT_FMT_ONELINE)
|
|
|
|
return 0;
|
2005-06-01 15:34:23 +00:00
|
|
|
date = strchr(line, '>');
|
|
|
|
if (!date)
|
|
|
|
return 0;
|
|
|
|
namelen = ++date - line;
|
|
|
|
time = strtoul(date, &date, 10);
|
|
|
|
tz = strtol(date, NULL, 10);
|
|
|
|
|
2006-04-18 23:45:27 +00:00
|
|
|
if (fmt == CMIT_FMT_EMAIL) {
|
2006-05-16 09:29:42 +00:00
|
|
|
char *name_tail = strchr(line, '<');
|
|
|
|
int display_name_length;
|
|
|
|
if (!name_tail)
|
|
|
|
return 0;
|
|
|
|
while (line < name_tail && isspace(name_tail[-1]))
|
|
|
|
name_tail--;
|
|
|
|
display_name_length = name_tail - line;
|
2006-04-18 23:45:27 +00:00
|
|
|
filler = "";
|
2006-05-16 09:29:42 +00:00
|
|
|
strcpy(buf, "From: ");
|
|
|
|
ret = strlen(buf);
|
|
|
|
ret += add_rfc2047(buf + ret, line, display_name_length);
|
|
|
|
memcpy(buf + ret, name_tail, namelen - display_name_length);
|
|
|
|
ret += namelen - display_name_length;
|
|
|
|
buf[ret++] = '\n';
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
ret = sprintf(buf, "%s: %.*s%.*s\n", what,
|
|
|
|
(fmt == CMIT_FMT_FULLER) ? 4 : 0,
|
|
|
|
filler, namelen, line);
|
2006-04-18 23:45:27 +00:00
|
|
|
}
|
2005-11-10 06:15:27 +00:00
|
|
|
switch (fmt) {
|
|
|
|
case CMIT_FMT_MEDIUM:
|
2005-06-05 16:02:03 +00:00
|
|
|
ret += sprintf(buf + ret, "Date: %s\n", show_date(time, tz));
|
2005-11-10 06:15:27 +00:00
|
|
|
break;
|
2006-04-18 23:45:27 +00:00
|
|
|
case CMIT_FMT_EMAIL:
|
2006-05-01 08:44:33 +00:00
|
|
|
ret += sprintf(buf + ret, "Date: %s\n",
|
|
|
|
show_rfc2822_date(time, tz));
|
2006-04-18 23:45:27 +00:00
|
|
|
break;
|
2005-11-10 06:15:27 +00:00
|
|
|
case CMIT_FMT_FULLER:
|
|
|
|
ret += sprintf(buf + ret, "%sDate: %s\n", what, show_date(time, tz));
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
/* notin' */
|
|
|
|
break;
|
|
|
|
}
|
2005-06-05 16:02:03 +00:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2006-04-18 23:45:27 +00:00
|
|
|
static int is_empty_line(const char *line, int *len_p)
|
2005-06-05 16:02:03 +00:00
|
|
|
{
|
2006-04-18 23:45:27 +00:00
|
|
|
int len = *len_p;
|
2005-06-05 16:02:03 +00:00
|
|
|
while (len && isspace(line[len-1]))
|
|
|
|
len--;
|
2006-04-18 23:45:27 +00:00
|
|
|
*len_p = len;
|
2005-06-05 16:02:03 +00:00
|
|
|
return !len;
|
2005-06-01 15:34:23 +00:00
|
|
|
}
|
|
|
|
|
2006-01-27 10:17:19 +00:00
|
|
|
static int add_merge_info(enum cmit_fmt fmt, char *buf, const struct commit *commit, int abbrev)
|
2005-06-18 20:52:05 +00:00
|
|
|
{
|
2006-01-27 10:17:19 +00:00
|
|
|
struct commit_list *parent = commit->parents;
|
|
|
|
int offset;
|
2005-08-09 05:15:40 +00:00
|
|
|
|
2006-04-18 23:45:27 +00:00
|
|
|
if ((fmt == CMIT_FMT_ONELINE) || (fmt == CMIT_FMT_EMAIL) ||
|
|
|
|
!parent || !parent->next)
|
2006-01-27 10:17:19 +00:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
offset = sprintf(buf, "Merge:");
|
|
|
|
|
|
|
|
while (parent) {
|
|
|
|
struct commit *p = parent->item;
|
2006-02-03 01:52:19 +00:00
|
|
|
const char *hex = abbrev
|
|
|
|
? find_unique_abbrev(p->object.sha1, abbrev)
|
|
|
|
: sha1_to_hex(p->object.sha1);
|
2006-06-28 09:04:39 +00:00
|
|
|
const char *dots = (abbrev && strlen(hex) != 40) ? "..." : "";
|
2006-01-27 10:17:19 +00:00
|
|
|
parent = parent->next;
|
|
|
|
|
2006-02-03 01:52:19 +00:00
|
|
|
offset += sprintf(buf + offset, " %s%s", hex, dots);
|
2005-06-18 20:52:05 +00:00
|
|
|
}
|
2006-01-27 10:17:19 +00:00
|
|
|
buf[offset++] = '\n';
|
2005-06-18 20:52:05 +00:00
|
|
|
return offset;
|
|
|
|
}
|
|
|
|
|
2006-05-20 13:40:29 +00:00
|
|
|
unsigned long pretty_print_commit(enum cmit_fmt fmt, const struct commit *commit, unsigned long len, char *buf, unsigned long space, int abbrev, const char *subject, const char *after_subject)
|
2005-06-01 15:34:23 +00:00
|
|
|
{
|
2005-06-05 16:02:03 +00:00
|
|
|
int hdr = 1, body = 0;
|
2005-06-01 15:34:23 +00:00
|
|
|
unsigned long offset = 0;
|
2006-04-18 23:45:27 +00:00
|
|
|
int indent = 4;
|
2006-01-27 10:17:19 +00:00
|
|
|
int parents_shown = 0;
|
2006-01-27 09:54:59 +00:00
|
|
|
const char *msg = commit->buffer;
|
2006-05-22 06:55:00 +00:00
|
|
|
int plain_non_ascii = 0;
|
2006-04-18 23:45:27 +00:00
|
|
|
|
|
|
|
if (fmt == CMIT_FMT_ONELINE || fmt == CMIT_FMT_EMAIL)
|
|
|
|
indent = 0;
|
2005-06-01 15:34:23 +00:00
|
|
|
|
2006-05-22 06:55:00 +00:00
|
|
|
/* After-subject is used to pass in Content-Type: multipart
|
|
|
|
* MIME header; in that case we do not have to do the
|
|
|
|
* plaintext content type even if the commit message has
|
|
|
|
* non 7-bit ASCII character. Otherwise, check if we need
|
|
|
|
* to say this is not a 7-bit ASCII.
|
|
|
|
*/
|
|
|
|
if (fmt == CMIT_FMT_EMAIL && !after_subject) {
|
2006-06-19 22:00:17 +00:00
|
|
|
int i, ch, in_body;
|
|
|
|
|
|
|
|
for (in_body = i = 0; (ch = msg[i]) && i < len; i++) {
|
|
|
|
if (!in_body) {
|
|
|
|
/* author could be non 7-bit ASCII but
|
|
|
|
* the log may so; skip over the
|
|
|
|
* header part first.
|
|
|
|
*/
|
|
|
|
if (ch == '\n' &&
|
|
|
|
i + 1 < len && msg[i+1] == '\n')
|
|
|
|
in_body = 1;
|
|
|
|
}
|
|
|
|
else if (ch & 0x80) {
|
2006-05-22 06:55:00 +00:00
|
|
|
plain_non_ascii = 1;
|
2006-06-19 22:00:17 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2006-05-22 06:55:00 +00:00
|
|
|
}
|
|
|
|
|
2005-06-01 15:34:23 +00:00
|
|
|
for (;;) {
|
|
|
|
const char *line = msg;
|
|
|
|
int linelen = get_one_line(msg, len);
|
|
|
|
|
|
|
|
if (!linelen)
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We want some slop for indentation and a possible
|
|
|
|
* final "...". Thus the "+ 20".
|
|
|
|
*/
|
|
|
|
if (offset + linelen + 20 > space) {
|
|
|
|
memcpy(buf + offset, " ...\n", 8);
|
|
|
|
offset += 8;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
msg += linelen;
|
|
|
|
len -= linelen;
|
|
|
|
if (hdr) {
|
2005-06-05 16:02:03 +00:00
|
|
|
if (linelen == 1) {
|
|
|
|
hdr = 0;
|
2006-04-18 23:45:27 +00:00
|
|
|
if ((fmt != CMIT_FMT_ONELINE) && !subject)
|
2005-08-09 05:15:40 +00:00
|
|
|
buf[offset++] = '\n';
|
2005-06-05 16:02:03 +00:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (fmt == CMIT_FMT_RAW) {
|
|
|
|
memcpy(buf + offset, line, linelen);
|
|
|
|
offset += linelen;
|
|
|
|
continue;
|
|
|
|
}
|
2005-06-18 20:52:05 +00:00
|
|
|
if (!memcmp(line, "parent ", 7)) {
|
|
|
|
if (linelen != 48)
|
|
|
|
die("bad parent line in commit");
|
2006-01-27 10:17:19 +00:00
|
|
|
continue;
|
2005-06-18 20:52:05 +00:00
|
|
|
}
|
2005-11-10 06:15:27 +00:00
|
|
|
|
2006-01-27 10:17:19 +00:00
|
|
|
if (!parents_shown) {
|
|
|
|
offset += add_merge_info(fmt, buf + offset,
|
|
|
|
commit, abbrev);
|
|
|
|
parents_shown = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2005-11-10 06:15:27 +00:00
|
|
|
/*
|
|
|
|
* MEDIUM == DEFAULT shows only author with dates.
|
|
|
|
* FULL shows both authors but not dates.
|
|
|
|
* FULLER shows both authors and dates.
|
|
|
|
*/
|
2005-06-01 15:34:23 +00:00
|
|
|
if (!memcmp(line, "author ", 7))
|
2005-11-10 06:15:27 +00:00
|
|
|
offset += add_user_info("Author", fmt,
|
|
|
|
buf + offset,
|
|
|
|
line + 7);
|
|
|
|
if (!memcmp(line, "committer ", 10) &&
|
|
|
|
(fmt == CMIT_FMT_FULL || fmt == CMIT_FMT_FULLER))
|
|
|
|
offset += add_user_info("Commit", fmt,
|
|
|
|
buf + offset,
|
|
|
|
line + 10);
|
2005-06-01 15:34:23 +00:00
|
|
|
continue;
|
|
|
|
}
|
2005-06-05 16:02:03 +00:00
|
|
|
|
2006-04-18 23:45:27 +00:00
|
|
|
if (is_empty_line(line, &linelen)) {
|
2005-06-05 16:02:03 +00:00
|
|
|
if (!body)
|
|
|
|
continue;
|
2006-04-18 23:45:27 +00:00
|
|
|
if (subject)
|
|
|
|
continue;
|
2005-06-05 16:02:03 +00:00
|
|
|
if (fmt == CMIT_FMT_SHORT)
|
|
|
|
break;
|
|
|
|
} else {
|
|
|
|
body = 1;
|
|
|
|
}
|
2005-08-09 05:15:40 +00:00
|
|
|
|
2006-04-18 23:45:27 +00:00
|
|
|
if (subject) {
|
2006-04-22 10:06:13 +00:00
|
|
|
int slen = strlen(subject);
|
|
|
|
memcpy(buf + offset, subject, slen);
|
|
|
|
offset += slen;
|
2006-05-16 09:29:42 +00:00
|
|
|
offset += add_rfc2047(buf + offset, line, linelen);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
memset(buf + offset, ' ', indent);
|
|
|
|
memcpy(buf + offset + indent, line, linelen);
|
|
|
|
offset += linelen + indent;
|
2006-04-18 23:45:27 +00:00
|
|
|
}
|
|
|
|
buf[offset++] = '\n';
|
2005-08-09 05:15:40 +00:00
|
|
|
if (fmt == CMIT_FMT_ONELINE)
|
|
|
|
break;
|
2006-05-22 06:55:00 +00:00
|
|
|
if (subject && plain_non_ascii) {
|
2006-05-16 09:29:42 +00:00
|
|
|
static const char header[] =
|
|
|
|
"Content-Type: text/plain; charset=UTF-8\n"
|
|
|
|
"Content-Transfer-Encoding: 8bit\n";
|
|
|
|
memcpy(buf + offset, header, sizeof(header)-1);
|
|
|
|
offset += sizeof(header)-1;
|
|
|
|
}
|
2006-05-20 13:40:29 +00:00
|
|
|
if (after_subject) {
|
|
|
|
int slen = strlen(after_subject);
|
|
|
|
if (slen > space - offset - 1)
|
|
|
|
slen = space - offset - 1;
|
|
|
|
memcpy(buf + offset, after_subject, slen);
|
|
|
|
offset += slen;
|
|
|
|
after_subject = NULL;
|
|
|
|
}
|
2006-04-18 23:45:27 +00:00
|
|
|
subject = NULL;
|
2005-08-09 05:15:40 +00:00
|
|
|
}
|
2006-04-15 04:20:51 +00:00
|
|
|
while (offset && isspace(buf[offset-1]))
|
|
|
|
offset--;
|
|
|
|
/* Make sure there is an EOLN for the non-oneline case */
|
|
|
|
if (fmt != CMIT_FMT_ONELINE)
|
|
|
|
buf[offset++] = '\n';
|
2005-06-01 15:34:23 +00:00
|
|
|
buf[offset] = '\0';
|
|
|
|
return offset;
|
|
|
|
}
|
2005-06-06 15:39:40 +00:00
|
|
|
|
|
|
|
struct commit *pop_commit(struct commit_list **stack)
|
|
|
|
{
|
|
|
|
struct commit_list *top = *stack;
|
|
|
|
struct commit *item = top ? top->item : NULL;
|
|
|
|
|
|
|
|
if (top) {
|
|
|
|
*stack = top->next;
|
|
|
|
free(top);
|
|
|
|
}
|
|
|
|
return item;
|
|
|
|
}
|
|
|
|
|
|
|
|
int count_parents(struct commit * commit)
|
|
|
|
{
|
|
|
|
int count = 0;
|
|
|
|
struct commit_list * parents = commit->parents;
|
|
|
|
for (count=0;parents; parents=parents->next,count++)
|
|
|
|
;
|
|
|
|
return count;
|
|
|
|
}
|
|
|
|
|
2006-03-10 09:21:37 +00:00
|
|
|
void topo_sort_default_setter(struct commit *c, void *data)
|
|
|
|
{
|
2006-06-18 01:26:18 +00:00
|
|
|
c->util = data;
|
2006-03-10 09:21:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void *topo_sort_default_getter(struct commit *c)
|
|
|
|
{
|
2006-06-18 01:26:18 +00:00
|
|
|
return c->util;
|
2006-03-10 09:21:37 +00:00
|
|
|
}
|
|
|
|
|
2005-07-06 16:39:34 +00:00
|
|
|
/*
|
|
|
|
* Performs an in-place topological sort on the list supplied.
|
|
|
|
*/
|
2006-02-16 06:05:33 +00:00
|
|
|
void sort_in_topological_order(struct commit_list ** list, int lifo)
|
2006-03-10 09:21:37 +00:00
|
|
|
{
|
|
|
|
sort_in_topological_order_fn(list, lifo, topo_sort_default_setter,
|
|
|
|
topo_sort_default_getter);
|
|
|
|
}
|
|
|
|
|
|
|
|
void sort_in_topological_order_fn(struct commit_list ** list, int lifo,
|
|
|
|
topo_sort_set_fn_t setter,
|
|
|
|
topo_sort_get_fn_t getter)
|
2005-07-06 16:39:34 +00:00
|
|
|
{
|
|
|
|
struct commit_list * next = *list;
|
2005-11-14 18:01:26 +00:00
|
|
|
struct commit_list * work = NULL, **insert;
|
2005-07-06 16:39:34 +00:00
|
|
|
struct commit_list ** pptr = list;
|
|
|
|
struct sort_node * nodes;
|
|
|
|
struct sort_node * next_nodes;
|
|
|
|
int count = 0;
|
|
|
|
|
|
|
|
/* determine the size of the list */
|
|
|
|
while (next) {
|
|
|
|
next = next->next;
|
|
|
|
count++;
|
|
|
|
}
|
2005-12-24 12:12:43 +00:00
|
|
|
|
|
|
|
if (!count)
|
|
|
|
return;
|
2005-07-06 16:39:34 +00:00
|
|
|
/* allocate an array to help sort the list */
|
|
|
|
nodes = xcalloc(count, sizeof(*nodes));
|
|
|
|
/* link the list to the array */
|
|
|
|
next_nodes = nodes;
|
|
|
|
next=*list;
|
|
|
|
while (next) {
|
|
|
|
next_nodes->list_item = next;
|
2006-03-10 09:21:37 +00:00
|
|
|
setter(next->item, next_nodes);
|
2005-07-06 16:39:34 +00:00
|
|
|
next_nodes++;
|
|
|
|
next = next->next;
|
|
|
|
}
|
|
|
|
/* update the indegree */
|
|
|
|
next=*list;
|
|
|
|
while (next) {
|
|
|
|
struct commit_list * parents = next->item->parents;
|
|
|
|
while (parents) {
|
|
|
|
struct commit * parent=parents->item;
|
2006-03-10 09:21:37 +00:00
|
|
|
struct sort_node * pn = (struct sort_node *) getter(parent);
|
|
|
|
|
2005-07-06 16:39:34 +00:00
|
|
|
if (pn)
|
|
|
|
pn->indegree++;
|
|
|
|
parents=parents->next;
|
|
|
|
}
|
|
|
|
next=next->next;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* find the tips
|
|
|
|
*
|
|
|
|
* tips are nodes not reachable from any other node in the list
|
|
|
|
*
|
|
|
|
* the tips serve as a starting set for the work queue.
|
|
|
|
*/
|
|
|
|
next=*list;
|
2005-11-14 18:01:26 +00:00
|
|
|
insert = &work;
|
2005-07-06 16:39:34 +00:00
|
|
|
while (next) {
|
2006-03-10 09:21:37 +00:00
|
|
|
struct sort_node * node = (struct sort_node *) getter(next->item);
|
2005-07-06 16:39:34 +00:00
|
|
|
|
|
|
|
if (node->indegree == 0) {
|
2005-11-14 18:01:26 +00:00
|
|
|
insert = &commit_list_insert(next->item, insert)->next;
|
2005-07-06 16:39:34 +00:00
|
|
|
}
|
|
|
|
next=next->next;
|
|
|
|
}
|
2006-02-16 06:05:33 +00:00
|
|
|
|
2005-07-06 16:39:34 +00:00
|
|
|
/* process the list in topological order */
|
2006-02-16 06:05:33 +00:00
|
|
|
if (!lifo)
|
|
|
|
sort_by_date(&work);
|
2005-07-06 16:39:34 +00:00
|
|
|
while (work) {
|
|
|
|
struct commit * work_item = pop_commit(&work);
|
2006-03-10 09:21:37 +00:00
|
|
|
struct sort_node * work_node = (struct sort_node *) getter(work_item);
|
2005-07-06 16:39:34 +00:00
|
|
|
struct commit_list * parents = work_item->parents;
|
|
|
|
|
|
|
|
while (parents) {
|
|
|
|
struct commit * parent=parents->item;
|
2006-03-10 09:21:37 +00:00
|
|
|
struct sort_node * pn = (struct sort_node *) getter(parent);
|
|
|
|
|
2005-07-06 16:39:34 +00:00
|
|
|
if (pn) {
|
2006-03-10 09:21:37 +00:00
|
|
|
/*
|
2005-07-06 16:39:34 +00:00
|
|
|
* parents are only enqueued for emission
|
|
|
|
* when all their children have been emitted thereby
|
|
|
|
* guaranteeing topological order.
|
|
|
|
*/
|
|
|
|
pn->indegree--;
|
2006-02-16 06:05:33 +00:00
|
|
|
if (!pn->indegree) {
|
|
|
|
if (!lifo)
|
|
|
|
insert_by_date(parent, &work);
|
|
|
|
else
|
|
|
|
commit_list_insert(parent, &work);
|
|
|
|
}
|
2005-07-06 16:39:34 +00:00
|
|
|
}
|
|
|
|
parents=parents->next;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* work_item is a commit all of whose children
|
|
|
|
* have already been emitted. we can emit it now.
|
|
|
|
*/
|
|
|
|
*pptr = work_node->list_item;
|
|
|
|
pptr = &(*pptr)->next;
|
|
|
|
*pptr = NULL;
|
2006-03-10 09:21:37 +00:00
|
|
|
setter(work_item, NULL);
|
2005-07-06 16:39:34 +00:00
|
|
|
}
|
|
|
|
free(nodes);
|
|
|
|
}
|