You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

712 lines
16 KiB

#include "cache.h"
#include "tag.h"
#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>
20 years ago
int save_commit_buffer = 1;
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;
};
const char *commit_type = "commit";
enum cmit_fmt get_commit_format(const char *arg)
{
if (!*arg)
return CMIT_FMT_DEFAULT;
if (!strcmp(arg, "=raw"))
return CMIT_FMT_RAW;
if (!strcmp(arg, "=medium"))
return CMIT_FMT_MEDIUM;
if (!strcmp(arg, "=short"))
return CMIT_FMT_SHORT;
if (!strcmp(arg, "=full"))
return CMIT_FMT_FULL;
if (!strcmp(arg, "=fuller"))
return CMIT_FMT_FULLER;
if (!strcmp(arg, "=oneline"))
return CMIT_FMT_ONELINE;
die("invalid --pretty format");
}
static struct commit *check_commit(struct object *obj,
const unsigned char *sha1,
int quiet)
{
if (obj->type != commit_type) {
if (!quiet)
error("Object %s is a %s, not a commit",
sha1_to_hex(sha1), obj->type);
return NULL;
}
return (struct commit *) obj;
}
struct commit *lookup_commit_reference_gently(const unsigned char *sha1,
int quiet)
{
struct object *obj = deref_tag(parse_object(sha1), NULL, 0);
if (!obj)
return NULL;
return check_commit(obj, sha1, quiet);
}
struct commit *lookup_commit_reference(const unsigned char *sha1)
{
return lookup_commit_reference_gently(sha1, 0);
}
struct commit *lookup_commit(const unsigned char *sha1)
{
struct object *obj = lookup_object(sha1);
if (!obj) {
struct commit *ret = xcalloc(1, sizeof(struct commit));
created_object(sha1, &ret->object);
ret->object.type = commit_type;
return ret;
}
if (!obj->type)
obj->type = commit_type;
return check_commit(obj, sha1, 0);
}
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;
}
static struct commit_graft **commit_graft;
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;
}
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;
if (buf[0] == '#' || buf[0] == '\0')
return NULL;
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)
{
FILE *fp = fopen(graft_file, "r");
char buf[1024];
if (!fp)
return -1;
while (fgets(buf, sizeof(buf), fp)) {
/* The format is just "Commit Parent1 Parent2 ...\n" */
int len = strlen(buf);
struct commit_graft *graft = read_graft_line(buf, len);
if (!graft)
continue;
if (register_commit_graft(graft, 1))
error("duplicate graft data: %s", buf);
}
fclose(fp);
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;
}
static struct commit_graft *lookup_commit_graft(const unsigned char *sha1)
{
int pos;
prepare_commit_graft();
pos = commit_graft_pos(sha1);
if (pos < 0)
return NULL;
return commit_graft[pos];
}
int parse_commit_buffer(struct commit *item, void *buffer, unsigned long size)
{
char *bufptr = buffer;
unsigned char parent[20];
struct commit_list **pptr;
struct commit_graft *graft;
unsigned n_refs = 0;
if (item->object.parsed)
return 0;
item->object.parsed = 1;
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)
return error("bad tree pointer in commit %s",
sha1_to_hex(item->object.sha1));
item->tree = lookup_tree(parent);
if (item->tree)
n_refs++;
bufptr += 46; /* "tree " + "hex sha1" + "\n" */
pptr = &item->parents;
graft = lookup_commit_graft(item->object.sha1);
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));
bufptr += 48;
if (graft)
continue;
new_parent = lookup_commit(parent);
if (new_parent) {
pptr = &commit_list_insert(new_parent, pptr)->next;
n_refs++;
}
}
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;
n_refs++;
}
}
item->date = parse_commit_date(bufptr);
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);
}
return 0;
}
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>
20 years ago
if (save_commit_buffer && !ret) {
item->buffer = buffer;
return 0;
}
free(buffer);
return ret;
}
struct commit_list *commit_list_insert(struct commit *item, struct commit_list **list_p)
{
struct commit_list *new_list = xmalloc(sizeof(struct commit_list));
new_list->item = item;
new_list->next = *list_p;
*list_p = new_list;
return new_list;
}
void free_commit_list(struct commit_list *list)
{
while (list) {
struct commit_list *temp = list;
list = temp->next;
free(temp);
}
}
struct commit_list * insert_by_date(struct commit *item, struct commit_list **list)
{
struct commit_list **pp = list;
struct commit_list *p;
while ((p = *pp) != NULL) {
if (p->item->date < item->date) {
break;
}
pp = &p->next;
}
return commit_list_insert(item, pp);
}
void sort_by_date(struct commit_list **list)
{
struct commit_list *ret = NULL;
while (*list) {
insert_by_date((*list)->item, &ret);
*list = (*list)->next;
}
*list = ret;
}
struct commit *pop_most_recent_commit(struct commit_list **list,
unsigned int mark)
{
struct commit *ret = (*list)->item;
struct commit_list *parents = ret->parents;
struct commit_list *old = *list;
*list = (*list)->next;
free(old);
while (parents) {
struct commit *commit = parents->item;
parse_commit(commit);
if (!(commit->object.flags & mark)) {
commit->object.flags |= mark;
insert_by_date(commit, list);
}
parents = parents->next;
}
return ret;
}
void clear_commit_marks(struct commit *commit, unsigned int mark)
{
struct commit_list *parents;
parents = commit->parents;
commit->object.flags &= ~mark;
while (parents) {
struct commit *parent = parents->item;
if (parent && parent->object.parsed &&
(parent->object.flags & mark))
clear_commit_marks(parent, mark);
parents = parents->next;
}
}
/*
* 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++;
if (!c)
break;
ret++;
if (c == '\n')
break;
}
return ret;
}
static int add_user_info(const char *what, enum cmit_fmt fmt, char *buf, const char *line)
{
char *date;
int namelen;
unsigned long time;
int tz, ret;
const char *filler = " ";
if (fmt == CMIT_FMT_ONELINE)
return 0;
date = strchr(line, '>');
if (!date)
return 0;
namelen = ++date - line;
time = strtoul(date, &date, 10);
tz = strtol(date, NULL, 10);
ret = sprintf(buf, "%s: %.*s%.*s\n", what,
(fmt == CMIT_FMT_FULLER) ? 4 : 0,
filler, namelen, line);
switch (fmt) {
case CMIT_FMT_MEDIUM:
ret += sprintf(buf + ret, "Date: %s\n", show_date(time, tz));
break;
case CMIT_FMT_FULLER:
ret += sprintf(buf + ret, "%sDate: %s\n", what, show_date(time, tz));
break;
default:
/* notin' */
break;
}
return ret;
}
static int is_empty_line(const char *line, int len)
{
while (len && isspace(line[len-1]))
len--;
return !len;
}
static int add_merge_info(enum cmit_fmt fmt, char *buf, const struct commit *commit, int abbrev)
{
struct commit_list *parent = commit->parents;
int offset;
if ((fmt == CMIT_FMT_ONELINE) || !parent || !parent->next)
return 0;
offset = sprintf(buf, "Merge:");
while (parent) {
struct commit *p = parent->item;
const char *hex = abbrev
? find_unique_abbrev(p->object.sha1, abbrev)
: sha1_to_hex(p->object.sha1);
char *dots = (abbrev && strlen(hex) != 40) ? "..." : "";
parent = parent->next;
offset += sprintf(buf + offset, " %s%s", hex, dots);
}
buf[offset++] = '\n';
return offset;
}
unsigned long pretty_print_commit(enum cmit_fmt fmt, const struct commit *commit, unsigned long len, char *buf, unsigned long space, int abbrev)
{
int hdr = 1, body = 0;
unsigned long offset = 0;
int indent = (fmt == CMIT_FMT_ONELINE) ? 0 : 4;
int parents_shown = 0;
const char *msg = commit->buffer;
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) {
if (linelen == 1) {
hdr = 0;
if (fmt != CMIT_FMT_ONELINE)
buf[offset++] = '\n';
continue;
}
if (fmt == CMIT_FMT_RAW) {
memcpy(buf + offset, line, linelen);
offset += linelen;
continue;
}
if (!memcmp(line, "parent ", 7)) {
if (linelen != 48)
die("bad parent line in commit");
continue;
}
if (!parents_shown) {
offset += add_merge_info(fmt, buf + offset,
commit, abbrev);
parents_shown = 1;
continue;
}
/*
* MEDIUM == DEFAULT shows only author with dates.
* FULL shows both authors but not dates.
* FULLER shows both authors and dates.
*/
if (!memcmp(line, "author ", 7))
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);
continue;
}
if (is_empty_line(line, linelen)) {
if (!body)
continue;
if (fmt == CMIT_FMT_SHORT)
break;
} else {
body = 1;
}
memset(buf + offset, ' ', indent);
memcpy(buf + offset + indent, line, linelen);
offset += linelen + indent;
if (fmt == CMIT_FMT_ONELINE)
break;
}
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';
buf[offset] = '\0';
return offset;
}
[PATCH] Modify git-rev-list to linearise the commit history in merge order. This patch linearises the GIT commit history graph into merge order which is defined by invariants specified in Documentation/git-rev-list.txt. The linearisation produced by this patch is superior in an objective sense to that produced by the existing git-rev-list implementation in that the linearisation produced is guaranteed to have the minimum number of discontinuities, where a discontinuity is defined as an adjacent pair of commits in the output list which are not related in a direct child-parent relationship. With this patch a graph like this: a4 --- | \ \ | b4 | |/ | | a3 | | | | | a2 | | | | c3 | | | | | c2 | b3 | | | /| | b2 | | | c1 | | / | b1 a1 | | | a0 | | / root Sorts like this: = a4 | c3 | c2 | c1 ^ b4 | b3 | b2 | b1 ^ a3 | a2 | a1 | a0 = root Instead of this: = a4 | c3 ^ b4 | a3 ^ c2 ^ b3 ^ a2 ^ b2 ^ c1 ^ a1 ^ b1 ^ a0 = root A test script, t/t6000-rev-list.sh, includes a test which demonstrates that the linearisation produced by --merge-order has less discontinuities than the linearisation produced by git-rev-list without the --merge-order flag specified. To see this, do the following: cd t ./t6000-rev-list.sh cd trash cat actual-default-order cat actual-merge-order The existing behaviour of git-rev-list is preserved, by default. To obtain the modified behaviour, specify --merge-order or --merge-order --show-breaks on the command line. This version of the patch has been tested on the git repository and also on the linux-2.6 repository and has reasonable performance on both - ~50-100% slower than the original algorithm. This version of the patch has incorporated a functional equivalent of the Linus' output limiting algorithm into the merge-order algorithm itself. This operates per the notes associated with Linus' commit 337cb3fb8da45f10fe9a0c3cf571600f55ead2ce. This version has incorporated Linus' feedback regarding proposed changes to rev-list.c. (see: [PATCH] Factor out filtering in rev-list.c) This version has improved the way sort_first_epoch marks commits as uninteresting. For more details about this change, refer to Documentation/git-rev-list.txt and http://blackcubes.dyndns.org/epoch/. Signed-off-by: Jon Seymour <jon.seymour@gmail.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago
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;
}
void topo_sort_default_setter(struct commit *c, void *data)
{
c->object.util = data;
}
void *topo_sort_default_getter(struct commit *c)
{
return c->object.util;
}
/*
* Performs an in-place topological sort on the list supplied.
*/
void sort_in_topological_order(struct commit_list ** list, int lifo)
{
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)
{
struct commit_list * next = *list;
struct commit_list * work = NULL, **insert;
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++;
}
if (!count)
return;
/* 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;
setter(next->item, next_nodes);
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;
struct sort_node * pn = (struct sort_node *) getter(parent);
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;
insert = &work;
while (next) {
struct sort_node * node = (struct sort_node *) getter(next->item);
if (node->indegree == 0) {
insert = &commit_list_insert(next->item, insert)->next;
}
next=next->next;
}
/* process the list in topological order */
if (!lifo)
sort_by_date(&work);
while (work) {
struct commit * work_item = pop_commit(&work);
struct sort_node * work_node = (struct sort_node *) getter(work_item);
struct commit_list * parents = work_item->parents;
while (parents) {
struct commit * parent=parents->item;
struct sort_node * pn = (struct sort_node *) getter(parent);
if (pn) {
/*
* parents are only enqueued for emission
* when all their children have been emitted thereby
* guaranteeing topological order.
*/
pn->indegree--;
if (!pn->indegree) {
if (!lifo)
insert_by_date(parent, &work);
else
commit_list_insert(parent, &work);
}
}
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;
setter(work_item, NULL);
}
free(nodes);
}