2006-02-26 01:19:46 +01:00
|
|
|
#include "cache.h"
|
|
|
|
#include "tag.h"
|
|
|
|
#include "blob.h"
|
|
|
|
#include "tree.h"
|
|
|
|
#include "commit.h"
|
2006-02-28 20:24:00 +01:00
|
|
|
#include "diff.h"
|
2006-02-26 01:19:46 +01:00
|
|
|
#include "refs.h"
|
|
|
|
#include "revision.h"
|
|
|
|
|
|
|
|
static char *path_name(struct name_path *path, const char *name)
|
|
|
|
{
|
|
|
|
struct name_path *p;
|
|
|
|
char *n, *m;
|
|
|
|
int nlen = strlen(name);
|
|
|
|
int len = nlen + 1;
|
|
|
|
|
|
|
|
for (p = path; p; p = p->up) {
|
|
|
|
if (p->elem_len)
|
|
|
|
len += p->elem_len + 1;
|
|
|
|
}
|
|
|
|
n = xmalloc(len);
|
|
|
|
m = n + len - (nlen + 1);
|
|
|
|
strcpy(m, name);
|
|
|
|
for (p = path; p; p = p->up) {
|
|
|
|
if (p->elem_len) {
|
|
|
|
m -= p->elem_len + 1;
|
|
|
|
memcpy(m, p->elem, p->elem_len);
|
|
|
|
m[p->elem_len] = '/';
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return n;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct object_list **add_object(struct object *obj,
|
|
|
|
struct object_list **p,
|
|
|
|
struct name_path *path,
|
|
|
|
const char *name)
|
|
|
|
{
|
|
|
|
struct object_list *entry = xmalloc(sizeof(*entry));
|
|
|
|
entry->item = obj;
|
|
|
|
entry->next = *p;
|
|
|
|
entry->name = path_name(path, name);
|
|
|
|
*p = entry;
|
|
|
|
return &entry->next;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mark_blob_uninteresting(struct blob *blob)
|
|
|
|
{
|
|
|
|
if (blob->object.flags & UNINTERESTING)
|
|
|
|
return;
|
|
|
|
blob->object.flags |= UNINTERESTING;
|
|
|
|
}
|
|
|
|
|
|
|
|
void mark_tree_uninteresting(struct tree *tree)
|
|
|
|
{
|
|
|
|
struct object *obj = &tree->object;
|
|
|
|
struct tree_entry_list *entry;
|
|
|
|
|
|
|
|
if (obj->flags & UNINTERESTING)
|
|
|
|
return;
|
|
|
|
obj->flags |= UNINTERESTING;
|
|
|
|
if (!has_sha1_file(obj->sha1))
|
|
|
|
return;
|
|
|
|
if (parse_tree(tree) < 0)
|
|
|
|
die("bad tree %s", sha1_to_hex(obj->sha1));
|
|
|
|
entry = tree->entries;
|
|
|
|
tree->entries = NULL;
|
|
|
|
while (entry) {
|
|
|
|
struct tree_entry_list *next = entry->next;
|
|
|
|
if (entry->directory)
|
|
|
|
mark_tree_uninteresting(entry->item.tree);
|
|
|
|
else
|
|
|
|
mark_blob_uninteresting(entry->item.blob);
|
|
|
|
free(entry);
|
|
|
|
entry = next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void mark_parents_uninteresting(struct commit *commit)
|
|
|
|
{
|
|
|
|
struct commit_list *parents = commit->parents;
|
|
|
|
|
|
|
|
while (parents) {
|
|
|
|
struct commit *commit = parents->item;
|
2006-03-09 05:04:36 +01:00
|
|
|
if (!(commit->object.flags & UNINTERESTING)) {
|
|
|
|
commit->object.flags |= UNINTERESTING;
|
2006-02-26 01:19:46 +01:00
|
|
|
|
2006-03-09 05:04:36 +01:00
|
|
|
/*
|
|
|
|
* Normally we haven't parsed the parent
|
|
|
|
* yet, so we won't have a parent of a parent
|
|
|
|
* here. However, it may turn out that we've
|
|
|
|
* reached this commit some other way (where it
|
|
|
|
* wasn't uninteresting), in which case we need
|
|
|
|
* to mark its parents recursively too..
|
|
|
|
*/
|
|
|
|
if (commit->parents)
|
|
|
|
mark_parents_uninteresting(commit);
|
|
|
|
}
|
2006-02-26 01:19:46 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
* A missing commit is ok iff its parent is marked
|
|
|
|
* uninteresting.
|
|
|
|
*
|
|
|
|
* We just mark such a thing parsed, so that when
|
|
|
|
* it is popped next time around, we won't be trying
|
|
|
|
* to parse it and get an error.
|
|
|
|
*/
|
|
|
|
if (!has_sha1_file(commit->object.sha1))
|
|
|
|
commit->object.parsed = 1;
|
|
|
|
parents = parents->next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void add_pending_object(struct rev_info *revs, struct object *obj, const char *name)
|
|
|
|
{
|
|
|
|
add_object(obj, &revs->pending_objects, NULL, name);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct commit *get_commit_reference(struct rev_info *revs, const char *name, const unsigned char *sha1, unsigned int flags)
|
|
|
|
{
|
|
|
|
struct object *object;
|
|
|
|
|
|
|
|
object = parse_object(sha1);
|
|
|
|
if (!object)
|
|
|
|
die("bad object %s", name);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Tag object? Look what it points to..
|
|
|
|
*/
|
|
|
|
while (object->type == tag_type) {
|
|
|
|
struct tag *tag = (struct tag *) object;
|
|
|
|
object->flags |= flags;
|
|
|
|
if (revs->tag_objects && !(object->flags & UNINTERESTING))
|
|
|
|
add_pending_object(revs, object, tag->tag);
|
|
|
|
object = parse_object(tag->tagged->sha1);
|
|
|
|
if (!object)
|
|
|
|
die("bad object %s", sha1_to_hex(tag->tagged->sha1));
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Commit object? Just return it, we'll do all the complex
|
|
|
|
* reachability crud.
|
|
|
|
*/
|
|
|
|
if (object->type == commit_type) {
|
|
|
|
struct commit *commit = (struct commit *)object;
|
|
|
|
object->flags |= flags;
|
|
|
|
if (parse_commit(commit) < 0)
|
|
|
|
die("unable to parse commit %s", name);
|
2006-02-27 17:54:36 +01:00
|
|
|
if (flags & UNINTERESTING) {
|
2006-02-26 01:19:46 +01:00
|
|
|
mark_parents_uninteresting(commit);
|
2006-02-27 17:54:36 +01:00
|
|
|
revs->limited = 1;
|
|
|
|
}
|
2006-02-26 01:19:46 +01:00
|
|
|
return commit;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Tree object? Either mark it uniniteresting, or add it
|
|
|
|
* to the list of objects to look at later..
|
|
|
|
*/
|
|
|
|
if (object->type == tree_type) {
|
|
|
|
struct tree *tree = (struct tree *)object;
|
|
|
|
if (!revs->tree_objects)
|
|
|
|
return NULL;
|
|
|
|
if (flags & UNINTERESTING) {
|
|
|
|
mark_tree_uninteresting(tree);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
add_pending_object(revs, object, "");
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Blob object? You know the drill by now..
|
|
|
|
*/
|
|
|
|
if (object->type == blob_type) {
|
|
|
|
struct blob *blob = (struct blob *)object;
|
|
|
|
if (!revs->blob_objects)
|
|
|
|
return NULL;
|
|
|
|
if (flags & UNINTERESTING) {
|
|
|
|
mark_blob_uninteresting(blob);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
add_pending_object(revs, object, "");
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
die("%s is unknown object", name);
|
|
|
|
}
|
|
|
|
|
2006-02-28 20:24:00 +01:00
|
|
|
static int everybody_uninteresting(struct commit_list *orig)
|
|
|
|
{
|
|
|
|
struct commit_list *list = orig;
|
|
|
|
while (list) {
|
|
|
|
struct commit *commit = list->item;
|
|
|
|
list = list->next;
|
|
|
|
if (commit->object.flags & UNINTERESTING)
|
|
|
|
continue;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2006-03-10 10:21:39 +01:00
|
|
|
static int tree_difference = REV_TREE_SAME;
|
2006-02-28 20:24:00 +01:00
|
|
|
|
|
|
|
static void file_add_remove(struct diff_options *options,
|
|
|
|
int addremove, unsigned mode,
|
|
|
|
const unsigned char *sha1,
|
|
|
|
const char *base, const char *path)
|
|
|
|
{
|
2006-03-10 10:21:39 +01:00
|
|
|
int diff = REV_TREE_DIFFERENT;
|
2006-02-28 20:24:00 +01:00
|
|
|
|
|
|
|
/*
|
2006-03-10 10:21:39 +01:00
|
|
|
* Is it an add of a new file? It means that the old tree
|
|
|
|
* didn't have it at all, so we will turn "REV_TREE_SAME" ->
|
|
|
|
* "REV_TREE_NEW", but leave any "REV_TREE_DIFFERENT" alone
|
|
|
|
* (and if it already was "REV_TREE_NEW", we'll keep it
|
|
|
|
* "REV_TREE_NEW" of course).
|
2006-02-28 20:24:00 +01:00
|
|
|
*/
|
|
|
|
if (addremove == '+') {
|
|
|
|
diff = tree_difference;
|
2006-03-10 10:21:39 +01:00
|
|
|
if (diff != REV_TREE_SAME)
|
2006-02-28 20:24:00 +01:00
|
|
|
return;
|
2006-03-10 10:21:39 +01:00
|
|
|
diff = REV_TREE_NEW;
|
2006-02-28 20:24:00 +01:00
|
|
|
}
|
|
|
|
tree_difference = diff;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void file_change(struct diff_options *options,
|
|
|
|
unsigned old_mode, unsigned new_mode,
|
|
|
|
const unsigned char *old_sha1,
|
|
|
|
const unsigned char *new_sha1,
|
|
|
|
const char *base, const char *path)
|
|
|
|
{
|
2006-03-10 10:21:39 +01:00
|
|
|
tree_difference = REV_TREE_DIFFERENT;
|
2006-02-28 20:24:00 +01:00
|
|
|
}
|
|
|
|
|
2006-04-11 03:14:54 +02:00
|
|
|
int rev_compare_tree(struct rev_info *revs, struct tree *t1, struct tree *t2)
|
2006-02-28 20:24:00 +01:00
|
|
|
{
|
|
|
|
if (!t1)
|
2006-03-10 10:21:39 +01:00
|
|
|
return REV_TREE_NEW;
|
2006-02-28 20:24:00 +01:00
|
|
|
if (!t2)
|
2006-03-10 10:21:39 +01:00
|
|
|
return REV_TREE_DIFFERENT;
|
|
|
|
tree_difference = REV_TREE_SAME;
|
2006-04-11 03:14:54 +02:00
|
|
|
if (diff_tree_sha1(t1->object.sha1, t2->object.sha1, "",
|
|
|
|
&revs->diffopt) < 0)
|
2006-03-10 10:21:39 +01:00
|
|
|
return REV_TREE_DIFFERENT;
|
2006-02-28 20:24:00 +01:00
|
|
|
return tree_difference;
|
|
|
|
}
|
|
|
|
|
2006-04-11 03:14:54 +02:00
|
|
|
int rev_same_tree_as_empty(struct rev_info *revs, struct tree *t1)
|
2006-02-28 20:24:00 +01:00
|
|
|
{
|
|
|
|
int retval;
|
|
|
|
void *tree;
|
|
|
|
struct tree_desc empty, real;
|
|
|
|
|
|
|
|
if (!t1)
|
|
|
|
return 0;
|
|
|
|
|
2006-04-02 14:44:09 +02:00
|
|
|
tree = read_object_with_reference(t1->object.sha1, tree_type, &real.size, NULL);
|
2006-02-28 20:24:00 +01:00
|
|
|
if (!tree)
|
|
|
|
return 0;
|
|
|
|
real.buf = tree;
|
|
|
|
|
|
|
|
empty.buf = "";
|
|
|
|
empty.size = 0;
|
|
|
|
|
|
|
|
tree_difference = 0;
|
2006-04-11 03:14:54 +02:00
|
|
|
retval = diff_tree(&empty, &real, "", &revs->diffopt);
|
2006-02-28 20:24:00 +01:00
|
|
|
free(tree);
|
|
|
|
|
|
|
|
return retval >= 0 && !tree_difference;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void try_to_simplify_commit(struct rev_info *revs, struct commit *commit)
|
|
|
|
{
|
|
|
|
struct commit_list **pp, *parent;
|
2006-03-11 06:59:37 +01:00
|
|
|
int tree_changed = 0;
|
2006-02-28 20:24:00 +01:00
|
|
|
|
|
|
|
if (!commit->tree)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (!commit->parents) {
|
2006-04-11 03:14:54 +02:00
|
|
|
if (!rev_same_tree_as_empty(revs, commit->tree))
|
2006-02-28 20:24:00 +01:00
|
|
|
commit->object.flags |= TREECHANGE;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
pp = &commit->parents;
|
|
|
|
while ((parent = *pp) != NULL) {
|
|
|
|
struct commit *p = parent->item;
|
|
|
|
|
|
|
|
parse_commit(p);
|
2006-04-11 03:14:54 +02:00
|
|
|
switch (rev_compare_tree(revs, p->tree, commit->tree)) {
|
2006-03-10 10:21:39 +01:00
|
|
|
case REV_TREE_SAME:
|
2006-03-11 06:59:37 +01:00
|
|
|
if (p->object.flags & UNINTERESTING) {
|
|
|
|
/* Even if a merge with an uninteresting
|
|
|
|
* side branch brought the entire change
|
|
|
|
* we are interested in, we do not want
|
|
|
|
* to lose the other branches of this
|
|
|
|
* merge, so we just keep going.
|
|
|
|
*/
|
|
|
|
pp = &parent->next;
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-28 20:24:00 +01:00
|
|
|
parent->next = NULL;
|
|
|
|
commit->parents = parent;
|
|
|
|
return;
|
|
|
|
|
2006-03-10 10:21:39 +01:00
|
|
|
case REV_TREE_NEW:
|
|
|
|
if (revs->remove_empty_trees &&
|
2006-04-11 03:14:54 +02:00
|
|
|
rev_same_tree_as_empty(revs, p->tree)) {
|
2006-03-12 22:39:31 +01:00
|
|
|
/* We are adding all the specified
|
|
|
|
* paths from this parent, so the
|
|
|
|
* history beyond this parent is not
|
|
|
|
* interesting. Remove its parents
|
|
|
|
* (they are grandparents for us).
|
|
|
|
* IOW, we pretend this parent is a
|
|
|
|
* "root" commit.
|
2006-03-12 22:39:31 +01:00
|
|
|
*/
|
2006-03-12 22:39:31 +01:00
|
|
|
parse_commit(p);
|
|
|
|
p->parents = NULL;
|
2006-02-28 20:24:00 +01:00
|
|
|
}
|
|
|
|
/* fallthrough */
|
2006-03-10 10:21:39 +01:00
|
|
|
case REV_TREE_DIFFERENT:
|
2006-03-11 06:59:37 +01:00
|
|
|
tree_changed = 1;
|
2006-02-28 20:24:00 +01:00
|
|
|
pp = &parent->next;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
die("bad tree compare for commit %s", sha1_to_hex(commit->object.sha1));
|
|
|
|
}
|
2006-03-11 06:59:37 +01:00
|
|
|
if (tree_changed)
|
|
|
|
commit->object.flags |= TREECHANGE;
|
2006-02-28 20:24:00 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void add_parents_to_list(struct rev_info *revs, struct commit *commit, struct commit_list **list)
|
|
|
|
{
|
|
|
|
struct commit_list *parent = commit->parents;
|
|
|
|
|
Make "--parents" logs also be incremental
The parent rewriting feature caused us to create the whole history in one
go, and then simplify it later, because of how rewrite_parents() had been
written. However, with a little tweaking, it's perfectly possible to do
even that one incrementally.
Right now, this doesn't really much matter, because every user of
"--parents" will probably generally _also_ use "--topo-order", which will
cause the old non-incremental behaviour anyway. However, I'm hopeful that
we could make even the topological sort incremental, or at least
_partially_ so (for example, make it incremental up to the first merge).
In the meantime, this at least moves things in the right direction, and
removes a strange special case.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-04-09 02:05:58 +02:00
|
|
|
if (commit->object.flags & ADDED)
|
|
|
|
return;
|
|
|
|
commit->object.flags |= ADDED;
|
|
|
|
|
2006-02-28 20:24:00 +01:00
|
|
|
/*
|
|
|
|
* If the commit is uninteresting, don't try to
|
|
|
|
* prune parents - we want the maximal uninteresting
|
|
|
|
* set.
|
|
|
|
*
|
|
|
|
* Normally we haven't parsed the parent
|
|
|
|
* yet, so we won't have a parent of a parent
|
|
|
|
* here. However, it may turn out that we've
|
|
|
|
* reached this commit some other way (where it
|
|
|
|
* wasn't uninteresting), in which case we need
|
|
|
|
* to mark its parents recursively too..
|
|
|
|
*/
|
|
|
|
if (commit->object.flags & UNINTERESTING) {
|
|
|
|
while (parent) {
|
|
|
|
struct commit *p = parent->item;
|
|
|
|
parent = parent->next;
|
|
|
|
parse_commit(p);
|
|
|
|
p->object.flags |= UNINTERESTING;
|
|
|
|
if (p->parents)
|
|
|
|
mark_parents_uninteresting(p);
|
|
|
|
if (p->object.flags & SEEN)
|
|
|
|
continue;
|
|
|
|
p->object.flags |= SEEN;
|
|
|
|
insert_by_date(p, list);
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Ok, the commit wasn't uninteresting. Try to
|
|
|
|
* simplify the commit history and find the parent
|
|
|
|
* that has no differences in the path set if one exists.
|
|
|
|
*/
|
2006-03-10 10:21:39 +01:00
|
|
|
if (revs->prune_fn)
|
|
|
|
revs->prune_fn(revs, commit);
|
2006-02-28 20:24:00 +01:00
|
|
|
|
|
|
|
parent = commit->parents;
|
|
|
|
while (parent) {
|
|
|
|
struct commit *p = parent->item;
|
|
|
|
|
|
|
|
parent = parent->next;
|
|
|
|
|
|
|
|
parse_commit(p);
|
|
|
|
if (p->object.flags & SEEN)
|
|
|
|
continue;
|
|
|
|
p->object.flags |= SEEN;
|
|
|
|
insert_by_date(p, list);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void limit_list(struct rev_info *revs)
|
|
|
|
{
|
|
|
|
struct commit_list *list = revs->commits;
|
|
|
|
struct commit_list *newlist = NULL;
|
|
|
|
struct commit_list **p = &newlist;
|
|
|
|
|
|
|
|
while (list) {
|
|
|
|
struct commit_list *entry = list;
|
|
|
|
struct commit *commit = list->item;
|
|
|
|
struct object *obj = &commit->object;
|
|
|
|
|
|
|
|
list = list->next;
|
|
|
|
free(entry);
|
|
|
|
|
|
|
|
if (revs->max_age != -1 && (commit->date < revs->max_age))
|
|
|
|
obj->flags |= UNINTERESTING;
|
|
|
|
if (revs->unpacked && has_sha1_pack(obj->sha1))
|
|
|
|
obj->flags |= UNINTERESTING;
|
|
|
|
add_parents_to_list(revs, commit, &list);
|
|
|
|
if (obj->flags & UNINTERESTING) {
|
|
|
|
mark_parents_uninteresting(commit);
|
|
|
|
if (everybody_uninteresting(list))
|
|
|
|
break;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (revs->min_age != -1 && (commit->date > revs->min_age))
|
|
|
|
continue;
|
|
|
|
p = &commit_list_insert(commit, p)->next;
|
|
|
|
}
|
2006-03-28 09:58:34 +02:00
|
|
|
if (revs->boundary) {
|
2006-03-31 09:59:19 +02:00
|
|
|
/* mark the ones that are on the result list first */
|
|
|
|
for (list = newlist; list; list = list->next) {
|
|
|
|
struct commit *commit = list->item;
|
|
|
|
commit->object.flags |= TMP_MARK;
|
|
|
|
}
|
|
|
|
for (list = newlist; list; list = list->next) {
|
2006-03-28 09:58:34 +02:00
|
|
|
struct commit *commit = list->item;
|
|
|
|
struct object *obj = &commit->object;
|
2006-03-31 09:59:19 +02:00
|
|
|
struct commit_list *parent;
|
|
|
|
if (obj->flags & UNINTERESTING)
|
2006-03-28 09:58:34 +02:00
|
|
|
continue;
|
2006-03-31 09:59:19 +02:00
|
|
|
for (parent = commit->parents;
|
|
|
|
parent;
|
|
|
|
parent = parent->next) {
|
2006-03-28 09:58:34 +02:00
|
|
|
struct commit *pcommit = parent->item;
|
|
|
|
if (!(pcommit->object.flags & UNINTERESTING))
|
|
|
|
continue;
|
|
|
|
pcommit->object.flags |= BOUNDARY;
|
2006-03-31 09:59:19 +02:00
|
|
|
if (pcommit->object.flags & TMP_MARK)
|
|
|
|
continue;
|
|
|
|
pcommit->object.flags |= TMP_MARK;
|
2006-03-28 09:58:34 +02:00
|
|
|
p = &commit_list_insert(pcommit, p)->next;
|
|
|
|
}
|
2006-03-31 09:59:19 +02:00
|
|
|
}
|
|
|
|
for (list = newlist; list; list = list->next) {
|
|
|
|
struct commit *commit = list->item;
|
|
|
|
commit->object.flags &= ~TMP_MARK;
|
2006-03-28 09:58:34 +02:00
|
|
|
}
|
|
|
|
}
|
2006-02-28 20:24:00 +01:00
|
|
|
revs->commits = newlist;
|
|
|
|
}
|
|
|
|
|
2006-02-26 01:19:46 +01:00
|
|
|
static void add_one_commit(struct commit *commit, struct rev_info *revs)
|
|
|
|
{
|
|
|
|
if (!commit || (commit->object.flags & SEEN))
|
|
|
|
return;
|
|
|
|
commit->object.flags |= SEEN;
|
|
|
|
commit_list_insert(commit, &revs->commits);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int all_flags;
|
|
|
|
static struct rev_info *all_revs;
|
|
|
|
|
|
|
|
static int handle_one_ref(const char *path, const unsigned char *sha1)
|
|
|
|
{
|
|
|
|
struct commit *commit = get_commit_reference(all_revs, path, sha1, all_flags);
|
|
|
|
add_one_commit(commit, all_revs);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void handle_all(struct rev_info *revs, unsigned flags)
|
|
|
|
{
|
|
|
|
all_revs = revs;
|
|
|
|
all_flags = flags;
|
|
|
|
for_each_ref(handle_one_ref);
|
|
|
|
}
|
|
|
|
|
2006-03-10 10:21:39 +01:00
|
|
|
void init_revisions(struct rev_info *revs)
|
|
|
|
{
|
|
|
|
memset(revs, 0, sizeof(*revs));
|
2006-04-11 03:14:54 +02:00
|
|
|
revs->diffopt.recursive = 1;
|
|
|
|
revs->diffopt.add_remove = file_add_remove;
|
|
|
|
revs->diffopt.change = file_change;
|
2006-03-10 10:21:39 +01:00
|
|
|
revs->lifo = 1;
|
|
|
|
revs->dense = 1;
|
|
|
|
revs->prefix = setup_git_directory();
|
|
|
|
revs->max_age = -1;
|
|
|
|
revs->min_age = -1;
|
|
|
|
revs->max_count = -1;
|
|
|
|
|
|
|
|
revs->prune_fn = NULL;
|
|
|
|
revs->prune_data = NULL;
|
|
|
|
|
|
|
|
revs->topo_setter = topo_sort_default_setter;
|
|
|
|
revs->topo_getter = topo_sort_default_getter;
|
|
|
|
}
|
|
|
|
|
2006-02-26 01:19:46 +01:00
|
|
|
/*
|
|
|
|
* Parse revision information, filling in the "rev_info" structure,
|
|
|
|
* and removing the used arguments from the argument list.
|
|
|
|
*
|
2006-03-01 00:07:20 +01:00
|
|
|
* Returns the number of arguments left that weren't recognized
|
|
|
|
* (which are also moved to the head of the argument list)
|
2006-02-26 01:19:46 +01:00
|
|
|
*/
|
2006-02-28 20:24:00 +01:00
|
|
|
int setup_revisions(int argc, const char **argv, struct rev_info *revs, const char *def)
|
2006-02-26 01:19:46 +01:00
|
|
|
{
|
|
|
|
int i, flags, seen_dashdash;
|
2006-03-01 00:07:20 +01:00
|
|
|
const char **unrecognized = argv + 1;
|
2006-02-26 01:19:46 +01:00
|
|
|
int left = 1;
|
|
|
|
|
2006-03-10 10:21:39 +01:00
|
|
|
init_revisions(revs);
|
2006-02-26 01:19:46 +01:00
|
|
|
|
|
|
|
/* First, search for "--" */
|
|
|
|
seen_dashdash = 0;
|
|
|
|
for (i = 1; i < argc; i++) {
|
|
|
|
const char *arg = argv[i];
|
|
|
|
if (strcmp(arg, "--"))
|
|
|
|
continue;
|
|
|
|
argv[i] = NULL;
|
|
|
|
argc = i;
|
2006-03-10 10:21:39 +01:00
|
|
|
revs->prune_data = get_pathspec(revs->prefix, argv + i + 1);
|
2006-02-26 01:19:46 +01:00
|
|
|
seen_dashdash = 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
flags = 0;
|
|
|
|
for (i = 1; i < argc; i++) {
|
|
|
|
struct commit *commit;
|
|
|
|
const char *arg = argv[i];
|
|
|
|
unsigned char sha1[20];
|
|
|
|
char *dotdot;
|
|
|
|
int local_flags;
|
|
|
|
|
|
|
|
if (*arg == '-') {
|
|
|
|
if (!strncmp(arg, "--max-count=", 12)) {
|
|
|
|
revs->max_count = atoi(arg + 12);
|
|
|
|
continue;
|
|
|
|
}
|
2006-03-03 00:24:01 +01:00
|
|
|
/* accept -<digit>, like traditilnal "head" */
|
|
|
|
if ((*arg == '-') && isdigit(arg[1])) {
|
|
|
|
revs->max_count = atoi(arg + 1);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "-n")) {
|
|
|
|
if (argc <= i + 1)
|
|
|
|
die("-n requires an argument");
|
|
|
|
revs->max_count = atoi(argv[++i]);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strncmp(arg,"-n",2)) {
|
|
|
|
revs->max_count = atoi(arg + 2);
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-26 01:19:46 +01:00
|
|
|
if (!strncmp(arg, "--max-age=", 10)) {
|
|
|
|
revs->max_age = atoi(arg + 10);
|
|
|
|
continue;
|
|
|
|
}
|
2006-03-01 09:12:37 +01:00
|
|
|
if (!strncmp(arg, "--since=", 8)) {
|
|
|
|
revs->max_age = approxidate(arg + 8);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strncmp(arg, "--after=", 8)) {
|
|
|
|
revs->max_age = approxidate(arg + 8);
|
2006-04-02 04:38:25 +02:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strncmp(arg, "--min-age=", 10)) {
|
|
|
|
revs->min_age = atoi(arg + 10);
|
2006-03-01 09:12:37 +01:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strncmp(arg, "--before=", 9)) {
|
|
|
|
revs->min_age = approxidate(arg + 9);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strncmp(arg, "--until=", 8)) {
|
|
|
|
revs->min_age = approxidate(arg + 8);
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-26 01:19:46 +01:00
|
|
|
if (!strcmp(arg, "--all")) {
|
|
|
|
handle_all(revs, flags);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "--not")) {
|
|
|
|
flags ^= UNINTERESTING;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "--default")) {
|
|
|
|
if (++i >= argc)
|
|
|
|
die("bad --default argument");
|
|
|
|
def = argv[i];
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "--topo-order")) {
|
|
|
|
revs->topo_order = 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "--date-order")) {
|
|
|
|
revs->lifo = 0;
|
|
|
|
revs->topo_order = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-03-31 02:52:42 +02:00
|
|
|
if (!strcmp(arg, "--parents")) {
|
|
|
|
revs->parents = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-26 01:19:46 +01:00
|
|
|
if (!strcmp(arg, "--dense")) {
|
|
|
|
revs->dense = 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "--sparse")) {
|
|
|
|
revs->dense = 0;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "--remove-empty")) {
|
|
|
|
revs->remove_empty_trees = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-03-28 10:04:50 +02:00
|
|
|
if (!strcmp(arg, "--no-merges")) {
|
2006-03-01 00:07:20 +01:00
|
|
|
revs->no_merges = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-03-28 09:58:34 +02:00
|
|
|
if (!strcmp(arg, "--boundary")) {
|
|
|
|
revs->boundary = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-26 01:19:46 +01:00
|
|
|
if (!strcmp(arg, "--objects")) {
|
|
|
|
revs->tag_objects = 1;
|
|
|
|
revs->tree_objects = 1;
|
|
|
|
revs->blob_objects = 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (!strcmp(arg, "--objects-edge")) {
|
|
|
|
revs->tag_objects = 1;
|
|
|
|
revs->tree_objects = 1;
|
|
|
|
revs->blob_objects = 1;
|
|
|
|
revs->edge_hint = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-27 17:54:36 +01:00
|
|
|
if (!strcmp(arg, "--unpacked")) {
|
|
|
|
revs->unpacked = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2006-02-26 01:19:46 +01:00
|
|
|
*unrecognized++ = arg;
|
|
|
|
left++;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
dotdot = strstr(arg, "..");
|
|
|
|
if (dotdot) {
|
|
|
|
unsigned char from_sha1[20];
|
2006-03-30 09:30:52 +02:00
|
|
|
const char *next = dotdot + 2;
|
|
|
|
const char *this = arg;
|
2006-02-26 01:19:46 +01:00
|
|
|
*dotdot = 0;
|
|
|
|
if (!*next)
|
2006-03-30 09:30:52 +02:00
|
|
|
next = "HEAD";
|
2006-03-30 05:41:37 +02:00
|
|
|
if (dotdot == arg)
|
2006-03-30 09:30:52 +02:00
|
|
|
this = "HEAD";
|
2006-03-30 05:41:37 +02:00
|
|
|
if (!get_sha1(this, from_sha1) &&
|
|
|
|
!get_sha1(next, sha1)) {
|
2006-02-26 01:19:46 +01:00
|
|
|
struct commit *exclude;
|
|
|
|
struct commit *include;
|
|
|
|
|
2006-03-30 05:41:37 +02:00
|
|
|
exclude = get_commit_reference(revs, this, from_sha1, flags ^ UNINTERESTING);
|
2006-02-26 01:19:46 +01:00
|
|
|
include = get_commit_reference(revs, next, sha1, flags);
|
|
|
|
if (!exclude || !include)
|
|
|
|
die("Invalid revision range %s..%s", arg, next);
|
|
|
|
add_one_commit(exclude, revs);
|
|
|
|
add_one_commit(include, revs);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
*dotdot = '.';
|
|
|
|
}
|
|
|
|
local_flags = 0;
|
|
|
|
if (*arg == '^') {
|
|
|
|
local_flags = UNINTERESTING;
|
|
|
|
arg++;
|
|
|
|
}
|
|
|
|
if (get_sha1(arg, sha1) < 0) {
|
|
|
|
struct stat st;
|
|
|
|
int j;
|
|
|
|
|
|
|
|
if (seen_dashdash || local_flags)
|
|
|
|
die("bad revision '%s'", arg);
|
|
|
|
|
|
|
|
/* If we didn't have a "--", all filenames must exist */
|
|
|
|
for (j = i; j < argc; j++) {
|
|
|
|
if (lstat(argv[j], &st) < 0)
|
2006-03-27 02:28:20 +02:00
|
|
|
die("'%s': %s", argv[j], strerror(errno));
|
2006-02-26 01:19:46 +01:00
|
|
|
}
|
2006-03-10 10:21:39 +01:00
|
|
|
revs->prune_data = get_pathspec(revs->prefix, argv + i);
|
2006-02-26 01:19:46 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
commit = get_commit_reference(revs, arg, sha1, flags ^ local_flags);
|
|
|
|
add_one_commit(commit, revs);
|
|
|
|
}
|
|
|
|
if (def && !revs->commits) {
|
|
|
|
unsigned char sha1[20];
|
|
|
|
struct commit *commit;
|
|
|
|
if (get_sha1(def, sha1) < 0)
|
|
|
|
die("bad default revision '%s'", def);
|
|
|
|
commit = get_commit_reference(revs, def, sha1, 0);
|
|
|
|
add_one_commit(commit, revs);
|
|
|
|
}
|
2006-03-10 10:21:39 +01:00
|
|
|
|
2006-04-02 05:13:22 +02:00
|
|
|
if (revs->topo_order || revs->unpacked)
|
2006-04-02 04:38:25 +02:00
|
|
|
revs->limited = 1;
|
|
|
|
|
2006-03-10 10:21:39 +01:00
|
|
|
if (revs->prune_data) {
|
2006-04-11 03:14:54 +02:00
|
|
|
diff_tree_setup_paths(revs->prune_data, &revs->diffopt);
|
2006-03-10 10:21:39 +01:00
|
|
|
revs->prune_fn = try_to_simplify_commit;
|
|
|
|
}
|
|
|
|
|
2006-02-26 01:19:46 +01:00
|
|
|
return left;
|
|
|
|
}
|
2006-02-28 20:24:00 +01:00
|
|
|
|
|
|
|
void prepare_revision_walk(struct rev_info *revs)
|
|
|
|
{
|
|
|
|
sort_by_date(&revs->commits);
|
|
|
|
if (revs->limited)
|
|
|
|
limit_list(revs);
|
|
|
|
if (revs->topo_order)
|
2006-03-10 10:21:39 +01:00
|
|
|
sort_in_topological_order_fn(&revs->commits, revs->lifo,
|
|
|
|
revs->topo_setter,
|
|
|
|
revs->topo_getter);
|
2006-02-28 20:24:00 +01:00
|
|
|
}
|
|
|
|
|
Make "--parents" logs also be incremental
The parent rewriting feature caused us to create the whole history in one
go, and then simplify it later, because of how rewrite_parents() had been
written. However, with a little tweaking, it's perfectly possible to do
even that one incrementally.
Right now, this doesn't really much matter, because every user of
"--parents" will probably generally _also_ use "--topo-order", which will
cause the old non-incremental behaviour anyway. However, I'm hopeful that
we could make even the topological sort incremental, or at least
_partially_ so (for example, make it incremental up to the first merge).
In the meantime, this at least moves things in the right direction, and
removes a strange special case.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-04-09 02:05:58 +02:00
|
|
|
static int rewrite_one(struct rev_info *revs, struct commit **pp)
|
2006-03-01 00:07:20 +01:00
|
|
|
{
|
|
|
|
for (;;) {
|
|
|
|
struct commit *p = *pp;
|
Make "--parents" logs also be incremental
The parent rewriting feature caused us to create the whole history in one
go, and then simplify it later, because of how rewrite_parents() had been
written. However, with a little tweaking, it's perfectly possible to do
even that one incrementally.
Right now, this doesn't really much matter, because every user of
"--parents" will probably generally _also_ use "--topo-order", which will
cause the old non-incremental behaviour anyway. However, I'm hopeful that
we could make even the topological sort incremental, or at least
_partially_ so (for example, make it incremental up to the first merge).
In the meantime, this at least moves things in the right direction, and
removes a strange special case.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-04-09 02:05:58 +02:00
|
|
|
if (!revs->limited)
|
|
|
|
add_parents_to_list(revs, p, &revs->commits);
|
2006-03-01 00:07:20 +01:00
|
|
|
if (p->object.flags & (TREECHANGE | UNINTERESTING))
|
|
|
|
return 0;
|
|
|
|
if (!p->parents)
|
|
|
|
return -1;
|
|
|
|
*pp = p->parents->item;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Make "--parents" logs also be incremental
The parent rewriting feature caused us to create the whole history in one
go, and then simplify it later, because of how rewrite_parents() had been
written. However, with a little tweaking, it's perfectly possible to do
even that one incrementally.
Right now, this doesn't really much matter, because every user of
"--parents" will probably generally _also_ use "--topo-order", which will
cause the old non-incremental behaviour anyway. However, I'm hopeful that
we could make even the topological sort incremental, or at least
_partially_ so (for example, make it incremental up to the first merge).
In the meantime, this at least moves things in the right direction, and
removes a strange special case.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-04-09 02:05:58 +02:00
|
|
|
static void rewrite_parents(struct rev_info *revs, struct commit *commit)
|
2006-03-01 00:07:20 +01:00
|
|
|
{
|
|
|
|
struct commit_list **pp = &commit->parents;
|
|
|
|
while (*pp) {
|
|
|
|
struct commit_list *parent = *pp;
|
Make "--parents" logs also be incremental
The parent rewriting feature caused us to create the whole history in one
go, and then simplify it later, because of how rewrite_parents() had been
written. However, with a little tweaking, it's perfectly possible to do
even that one incrementally.
Right now, this doesn't really much matter, because every user of
"--parents" will probably generally _also_ use "--topo-order", which will
cause the old non-incremental behaviour anyway. However, I'm hopeful that
we could make even the topological sort incremental, or at least
_partially_ so (for example, make it incremental up to the first merge).
In the meantime, this at least moves things in the right direction, and
removes a strange special case.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-04-09 02:05:58 +02:00
|
|
|
if (rewrite_one(revs, &parent->item) < 0) {
|
2006-03-01 00:07:20 +01:00
|
|
|
*pp = parent->next;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
pp = &parent->next;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-04-17 03:12:49 +02:00
|
|
|
static void mark_boundary_to_show(struct commit *commit)
|
|
|
|
{
|
|
|
|
struct commit_list *p = commit->parents;
|
|
|
|
while (p) {
|
|
|
|
commit = p->item;
|
|
|
|
p = p->next;
|
|
|
|
if (commit->object.flags & BOUNDARY)
|
|
|
|
commit->object.flags |= BOUNDARY_SHOW;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-02-28 20:24:00 +01:00
|
|
|
struct commit *get_revision(struct rev_info *revs)
|
|
|
|
{
|
2006-03-01 00:07:20 +01:00
|
|
|
struct commit_list *list = revs->commits;
|
|
|
|
|
|
|
|
if (!list)
|
2006-02-28 20:24:00 +01:00
|
|
|
return NULL;
|
|
|
|
|
2006-03-01 00:07:20 +01:00
|
|
|
/* Check the max_count ... */
|
|
|
|
switch (revs->max_count) {
|
|
|
|
case -1:
|
|
|
|
break;
|
|
|
|
case 0:
|
|
|
|
return NULL;
|
|
|
|
default:
|
|
|
|
revs->max_count--;
|
|
|
|
}
|
2006-02-28 20:24:00 +01:00
|
|
|
|
2006-03-01 00:07:20 +01:00
|
|
|
do {
|
2006-03-05 18:53:52 +01:00
|
|
|
struct commit *commit = revs->commits->item;
|
|
|
|
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
revs->commits = revs->commits->next;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If we haven't done the list limiting, we need to look at
|
2006-04-02 02:35:06 +02:00
|
|
|
* the parents here. We also need to do the date-based limiting
|
|
|
|
* that we'd otherwise have done in limit_list().
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
*/
|
2006-04-02 02:35:06 +02:00
|
|
|
if (!revs->limited) {
|
2006-04-02 04:52:58 +02:00
|
|
|
if ((revs->unpacked &&
|
|
|
|
has_sha1_pack(commit->object.sha1)) ||
|
|
|
|
(revs->max_age != -1 &&
|
|
|
|
(commit->date < revs->max_age)))
|
2006-04-02 02:35:06 +02:00
|
|
|
continue;
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
add_parents_to_list(revs, commit, &revs->commits);
|
2006-04-02 02:35:06 +02:00
|
|
|
}
|
2006-03-28 09:58:34 +02:00
|
|
|
if (commit->object.flags & SHOWN)
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
continue;
|
2006-04-17 03:12:49 +02:00
|
|
|
|
|
|
|
/* We want to show boundary commits only when their
|
|
|
|
* children are shown. When path-limiter is in effect,
|
|
|
|
* rewrite_parents() drops some commits from getting shown,
|
|
|
|
* and there is no point showing boundary parents that
|
|
|
|
* are not shown. After rewrite_parents() rewrites the
|
|
|
|
* parents of a commit that is shown, we mark the boundary
|
|
|
|
* parents with BOUNDARY_SHOW.
|
|
|
|
*/
|
|
|
|
if (commit->object.flags & BOUNDARY_SHOW) {
|
|
|
|
commit->object.flags |= SHOWN;
|
|
|
|
return commit;
|
|
|
|
}
|
|
|
|
if (commit->object.flags & UNINTERESTING)
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
continue;
|
2006-03-01 00:07:20 +01:00
|
|
|
if (revs->min_age != -1 && (commit->date > revs->min_age))
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
continue;
|
2006-03-28 09:58:34 +02:00
|
|
|
if (revs->no_merges &&
|
|
|
|
commit->parents && commit->parents->next)
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
continue;
|
2006-03-10 10:21:39 +01:00
|
|
|
if (revs->prune_fn && revs->dense) {
|
2006-03-01 00:07:20 +01:00
|
|
|
if (!(commit->object.flags & TREECHANGE))
|
Make path-limiting be incremental when possible.
This makes git-rev-list able to do path-limiting without having to parse
all of history before it starts showing the results.
This makes things like "git log -- pathname" much more pleasant to use.
This is actually a pretty small patch, and the biggest part of it is
purely cleanups (turning the "goto next" statements into "continue"), but
it's conceptually a lot bigger than it looks.
What it does is that if you do a path-limited revision list, and you do
_not_ ask for pseudo-parenthood information, it won't do all the
path-limiting up-front, but instead do it incrementally in
"get_revision()".
This is an absolutely huge deal for anything like "git log -- <pathname>",
but also for some things that we don't do yet - like the "find where
things changed" logic I've described elsewhere, where we want to find the
previous revision that changed a file.
The reason I put "RFC" in the subject line is that while I've validated it
various ways, like doing
git-rev-list HEAD -- drivers/char/ | md5sum
before-and-after on the kernel archive, it's "git-rev-list" after all. In
other words, it's that really really subtle and complex central piece of
software. So while I think this is important and should go in asap, I also
think it should get lots of testing and eyeballs looking at the code.
Btw, don't even bother testing this with the git archive. git itself is so
small that parsing the whole revision history for it takes about a second
even with path limiting. The thing that _really_ shows this off is doing
git log drivers/
on the kernel archive, or even better, on the _historic_ kernel archive.
With this change, the response is instantaneous (although seeking to the
end of the result will obviously take as long as it ever did). Before this
change, the command would think about the result for tens of seconds - or
even minutes, in the case of the bigger old kernel archive - before
starting to output the results.
NOTE NOTE NOTE! Using path limiting with things like "gitk", which uses
the "--parents" flag to actually generate a pseudo-history of the
resulting commits won't actually see the improvement in interactivity,
since that forces git-rev-list to do the whole-history thing after all.
MAYBE we can fix that too at some point, but I won't promise anything.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-03-31 03:05:25 +02:00
|
|
|
continue;
|
|
|
|
if (revs->parents)
|
Make "--parents" logs also be incremental
The parent rewriting feature caused us to create the whole history in one
go, and then simplify it later, because of how rewrite_parents() had been
written. However, with a little tweaking, it's perfectly possible to do
even that one incrementally.
Right now, this doesn't really much matter, because every user of
"--parents" will probably generally _also_ use "--topo-order", which will
cause the old non-incremental behaviour anyway. However, I'm hopeful that
we could make even the topological sort incremental, or at least
_partially_ so (for example, make it incremental up to the first merge).
In the meantime, this at least moves things in the right direction, and
removes a strange special case.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-04-09 02:05:58 +02:00
|
|
|
rewrite_parents(revs, commit);
|
2006-03-28 09:58:34 +02:00
|
|
|
}
|
2006-04-17 03:12:49 +02:00
|
|
|
if (revs->boundary)
|
|
|
|
mark_boundary_to_show(commit);
|
2006-03-01 00:07:20 +01:00
|
|
|
commit->object.flags |= SHOWN;
|
|
|
|
return commit;
|
|
|
|
} while (revs->commits);
|
|
|
|
return NULL;
|
|
|
|
}
|