2008-07-09 14:58:57 +02:00
|
|
|
#include "cache.h"
|
2014-10-01 12:28:42 +02:00
|
|
|
#include "lockfile.h"
|
2008-07-21 20:03:49 +02:00
|
|
|
#include "string-list.h"
|
2008-07-09 14:58:57 +02:00
|
|
|
#include "rerere.h"
|
|
|
|
#include "xdiff-interface.h"
|
2009-12-26 00:51:32 +01:00
|
|
|
#include "dir.h"
|
|
|
|
#include "resolve-undo.h"
|
|
|
|
#include "ll-merge.h"
|
2010-01-17 08:28:46 +01:00
|
|
|
#include "attr.h"
|
2013-07-14 10:35:40 +02:00
|
|
|
#include "pathspec.h"
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
#include "sha1-lookup.h"
|
2008-07-09 14:58:57 +02:00
|
|
|
|
2011-02-16 11:47:44 +01:00
|
|
|
#define RESOLVED 0
|
|
|
|
#define PUNTED 1
|
|
|
|
#define THREE_STAGED 2
|
|
|
|
void *RERERE_RESOLVED = &RERERE_RESOLVED;
|
|
|
|
|
2008-07-09 14:58:57 +02:00
|
|
|
/* if rerere_enabled == -1, fall back to detection of .git/rr-cache */
|
|
|
|
static int rerere_enabled = -1;
|
|
|
|
|
|
|
|
/* automatically update cleanly resolved paths to the index */
|
|
|
|
static int rerere_autoupdate;
|
|
|
|
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
static int rerere_dir_nr;
|
|
|
|
static int rerere_dir_alloc;
|
|
|
|
|
2015-07-17 00:47:13 +02:00
|
|
|
#define RR_HAS_POSTIMAGE 1
|
|
|
|
#define RR_HAS_PREIMAGE 2
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
static struct rerere_dir {
|
|
|
|
unsigned char sha1[20];
|
2015-07-23 23:23:24 +02:00
|
|
|
int status_alloc, status_nr;
|
|
|
|
unsigned char *status;
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
} **rerere_dir;
|
|
|
|
|
|
|
|
static void free_rerere_dirs(void)
|
|
|
|
{
|
|
|
|
int i;
|
2015-07-23 23:23:24 +02:00
|
|
|
for (i = 0; i < rerere_dir_nr; i++) {
|
|
|
|
free(rerere_dir[i]->status);
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
free(rerere_dir[i]);
|
2015-07-23 23:23:24 +02:00
|
|
|
}
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
free(rerere_dir);
|
|
|
|
rerere_dir_nr = rerere_dir_alloc = 0;
|
|
|
|
rerere_dir = NULL;
|
|
|
|
}
|
|
|
|
|
2015-07-05 02:38:34 +02:00
|
|
|
static void free_rerere_id(struct string_list_item *item)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2015-07-05 02:38:34 +02:00
|
|
|
free(item->util);
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
2015-07-05 02:38:34 +02:00
|
|
|
static const char *rerere_id_hex(const struct rerere_id *id)
|
|
|
|
{
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
return sha1_to_hex(id->collection->sha1);
|
2015-07-05 02:38:34 +02:00
|
|
|
}
|
|
|
|
|
2015-07-23 23:23:24 +02:00
|
|
|
static void fit_variant(struct rerere_dir *rr_dir, int variant)
|
|
|
|
{
|
|
|
|
variant++;
|
|
|
|
ALLOC_GROW(rr_dir->status, variant, rr_dir->status_alloc);
|
|
|
|
if (rr_dir->status_nr < variant) {
|
|
|
|
memset(rr_dir->status + rr_dir->status_nr,
|
|
|
|
'\0', variant - rr_dir->status_nr);
|
|
|
|
rr_dir->status_nr = variant;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void assign_variant(struct rerere_id *id)
|
|
|
|
{
|
|
|
|
int variant;
|
|
|
|
struct rerere_dir *rr_dir = id->collection;
|
|
|
|
|
|
|
|
variant = id->variant;
|
|
|
|
if (variant < 0) {
|
2015-07-31 00:49:18 +02:00
|
|
|
for (variant = 0; variant < rr_dir->status_nr; variant++)
|
|
|
|
if (!rr_dir->status[variant])
|
|
|
|
break;
|
2015-07-23 23:23:24 +02:00
|
|
|
}
|
|
|
|
fit_variant(rr_dir, variant);
|
|
|
|
id->variant = variant;
|
2015-07-05 02:38:34 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
const char *rerere_path(const struct rerere_id *id, const char *file)
|
|
|
|
{
|
|
|
|
if (!file)
|
|
|
|
return git_path("rr-cache/%s", rerere_id_hex(id));
|
|
|
|
|
2015-07-23 23:23:24 +02:00
|
|
|
if (id->variant <= 0)
|
|
|
|
return git_path("rr-cache/%s/%s", rerere_id_hex(id), file);
|
|
|
|
|
|
|
|
return git_path("rr-cache/%s/%s.%d",
|
|
|
|
rerere_id_hex(id), file, id->variant);
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
2015-07-23 23:23:24 +02:00
|
|
|
static int is_rr_file(const char *name, const char *filename, int *variant)
|
2015-07-17 00:47:13 +02:00
|
|
|
{
|
2015-07-23 23:23:24 +02:00
|
|
|
const char *suffix;
|
|
|
|
char *ep;
|
|
|
|
|
|
|
|
if (!strcmp(name, filename)) {
|
|
|
|
*variant = 0;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
if (!skip_prefix(name, filename, &suffix) || *suffix != '.')
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
errno = 0;
|
|
|
|
*variant = strtol(suffix + 1, &ep, 10);
|
|
|
|
if (errno || *ep)
|
|
|
|
return 0;
|
|
|
|
return 1;
|
2015-07-17 00:47:13 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static void scan_rerere_dir(struct rerere_dir *rr_dir)
|
|
|
|
{
|
|
|
|
struct dirent *de;
|
|
|
|
DIR *dir = opendir(git_path("rr-cache/%s", sha1_to_hex(rr_dir->sha1)));
|
|
|
|
|
|
|
|
if (!dir)
|
|
|
|
return;
|
|
|
|
while ((de = readdir(dir)) != NULL) {
|
2015-07-23 23:23:24 +02:00
|
|
|
int variant;
|
|
|
|
|
|
|
|
if (is_rr_file(de->d_name, "postimage", &variant)) {
|
|
|
|
fit_variant(rr_dir, variant);
|
|
|
|
rr_dir->status[variant] |= RR_HAS_POSTIMAGE;
|
|
|
|
} else if (is_rr_file(de->d_name, "preimage", &variant)) {
|
|
|
|
fit_variant(rr_dir, variant);
|
|
|
|
rr_dir->status[variant] |= RR_HAS_PREIMAGE;
|
|
|
|
}
|
2015-07-17 00:47:13 +02:00
|
|
|
}
|
|
|
|
closedir(dir);
|
|
|
|
}
|
|
|
|
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
static const unsigned char *rerere_dir_sha1(size_t i, void *table)
|
|
|
|
{
|
|
|
|
struct rerere_dir **rr_dir = table;
|
|
|
|
return rr_dir[i]->sha1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct rerere_dir *find_rerere_dir(const char *hex)
|
|
|
|
{
|
|
|
|
unsigned char sha1[20];
|
|
|
|
struct rerere_dir *rr_dir;
|
|
|
|
int pos;
|
|
|
|
|
|
|
|
if (get_sha1_hex(hex, sha1))
|
|
|
|
return NULL; /* BUG */
|
|
|
|
pos = sha1_pos(sha1, rerere_dir, rerere_dir_nr, rerere_dir_sha1);
|
|
|
|
if (pos < 0) {
|
|
|
|
rr_dir = xmalloc(sizeof(*rr_dir));
|
|
|
|
hashcpy(rr_dir->sha1, sha1);
|
2015-07-23 23:23:24 +02:00
|
|
|
rr_dir->status = NULL;
|
|
|
|
rr_dir->status_nr = 0;
|
|
|
|
rr_dir->status_alloc = 0;
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
pos = -1 - pos;
|
|
|
|
|
|
|
|
/* Make sure the array is big enough ... */
|
|
|
|
ALLOC_GROW(rerere_dir, rerere_dir_nr + 1, rerere_dir_alloc);
|
|
|
|
/* ... and add it in. */
|
|
|
|
rerere_dir_nr++;
|
|
|
|
memmove(rerere_dir + pos + 1, rerere_dir + pos,
|
|
|
|
(rerere_dir_nr - pos - 1) * sizeof(*rerere_dir));
|
|
|
|
rerere_dir[pos] = rr_dir;
|
2015-07-17 00:47:13 +02:00
|
|
|
scan_rerere_dir(rr_dir);
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
}
|
|
|
|
return rerere_dir[pos];
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
2015-07-05 02:38:34 +02:00
|
|
|
static int has_rerere_resolution(const struct rerere_id *id)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2015-07-17 22:28:31 +02:00
|
|
|
const int both = RR_HAS_POSTIMAGE|RR_HAS_PREIMAGE;
|
2015-07-23 23:23:24 +02:00
|
|
|
int variant = id->variant;
|
2015-07-05 02:38:34 +02:00
|
|
|
|
2015-07-23 23:23:24 +02:00
|
|
|
if (variant < 0)
|
|
|
|
return 0;
|
|
|
|
return ((id->collection->status[variant] & both) == both);
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
2015-07-05 02:38:34 +02:00
|
|
|
static struct rerere_id *new_rerere_id_hex(char *hex)
|
|
|
|
{
|
|
|
|
struct rerere_id *id = xmalloc(sizeof(*id));
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
id->collection = find_rerere_dir(hex);
|
2015-07-23 23:23:24 +02:00
|
|
|
id->variant = -1; /* not known yet */
|
2015-07-05 02:38:34 +02:00
|
|
|
return id;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct rerere_id *new_rerere_id(unsigned char *sha1)
|
|
|
|
{
|
|
|
|
return new_rerere_id_hex(sha1_to_hex(sha1));
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
2015-07-01 07:36:35 +02:00
|
|
|
/*
|
|
|
|
* $GIT_DIR/MERGE_RR file is a collection of records, each of which is
|
|
|
|
* "conflict ID", a HT and pathname, terminated with a NUL, and is
|
|
|
|
* used to keep track of the set of paths that "rerere" may need to
|
|
|
|
* work on (i.e. what is left by the previous invocation of "git
|
|
|
|
* rerere" during the current conflict resolution session).
|
|
|
|
*/
|
2008-07-21 20:03:49 +02:00
|
|
|
static void read_rr(struct string_list *rr)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2015-06-29 00:51:59 +02:00
|
|
|
struct strbuf buf = STRBUF_INIT;
|
memoize common git-path "constant" files
One of the most common uses of git_path() is to pass a
constant, like git_path("MERGE_MSG"). This has two
drawbacks:
1. The return value is a static buffer, and the lifetime
is dependent on other calls to git_path, etc.
2. There's no compile-time checking of the pathname. This
is OK for a one-off (after all, we have to spell it
correctly at least once), but many of these constant
strings appear throughout the code.
This patch introduces a series of functions to "memoize"
these strings, which are essentially globals for the
lifetime of the program. We compute the value once, take
ownership of the buffer, and return the cached value for
subsequent calls. cache.h provides a helper macro for
defining these functions as one-liners, and defines a few
common ones for global use.
Using a macro is a little bit gross, but it does nicely
document the purpose of the functions. If we need to touch
them all later (e.g., because we learned how to change the
git_dir variable at runtime, and need to invalidate all of
the stored values), it will be much easier to have the
complete list.
Note that the shared-global functions have separate, manual
declarations. We could do something clever with the macros
(e.g., expand it to a declaration in some places, and a
declaration _and_ a definition in path.c). But there aren't
that many, and it's probably better to stay away from
too-magical macros.
Likewise, if we abandon the C preprocessor in favor of
generating these with a script, we could get much fancier.
E.g., normalizing "FOO/BAR-BAZ" into "git_path_foo_bar_baz".
But the small amount of saved typing is probably not worth
the resulting confusion to readers who want to grep for the
function's definition.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-10 11:38:57 +02:00
|
|
|
FILE *in = fopen(git_path_merge_rr(), "r");
|
2015-06-29 00:51:59 +02:00
|
|
|
|
2008-07-09 14:58:57 +02:00
|
|
|
if (!in)
|
|
|
|
return;
|
2015-06-29 00:51:59 +02:00
|
|
|
while (!strbuf_getwholeline(&buf, in, '\0')) {
|
|
|
|
char *path;
|
|
|
|
unsigned char sha1[20];
|
2015-07-05 02:38:34 +02:00
|
|
|
struct rerere_id *id;
|
2015-07-23 23:23:24 +02:00
|
|
|
int variant;
|
2015-06-29 00:51:59 +02:00
|
|
|
|
|
|
|
/* There has to be the hash, tab, path and then NUL */
|
|
|
|
if (buf.len < 42 || get_sha1_hex(buf.buf, sha1))
|
2008-07-09 14:58:57 +02:00
|
|
|
die("corrupt MERGE_RR");
|
2015-06-29 00:51:59 +02:00
|
|
|
|
2015-07-23 23:23:24 +02:00
|
|
|
if (buf.buf[40] != '.') {
|
|
|
|
variant = 0;
|
|
|
|
path = buf.buf + 40;
|
|
|
|
} else {
|
|
|
|
errno = 0;
|
|
|
|
variant = strtol(buf.buf + 41, &path, 10);
|
|
|
|
if (errno)
|
|
|
|
die("corrupt MERGE_RR");
|
|
|
|
}
|
|
|
|
if (*(path++) != '\t')
|
2008-07-09 14:58:57 +02:00
|
|
|
die("corrupt MERGE_RR");
|
2015-06-29 00:51:59 +02:00
|
|
|
buf.buf[40] = '\0';
|
2015-07-05 02:38:34 +02:00
|
|
|
id = new_rerere_id_hex(buf.buf);
|
2015-07-23 23:23:24 +02:00
|
|
|
id->variant = variant;
|
2015-07-05 02:38:34 +02:00
|
|
|
string_list_insert(rr, path)->util = id;
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
2015-06-29 00:51:59 +02:00
|
|
|
strbuf_release(&buf);
|
2008-07-09 14:58:57 +02:00
|
|
|
fclose(in);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct lock_file write_lock;
|
|
|
|
|
2008-07-21 20:03:49 +02:00
|
|
|
static int write_rr(struct string_list *rr, int out_fd)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0; i < rr->nr; i++) {
|
2015-06-29 01:28:00 +02:00
|
|
|
struct strbuf buf = STRBUF_INIT;
|
2015-07-05 02:38:34 +02:00
|
|
|
struct rerere_id *id;
|
2015-06-29 01:28:00 +02:00
|
|
|
|
|
|
|
assert(rr->items[i].util != RERERE_RESOLVED);
|
2015-07-05 02:38:34 +02:00
|
|
|
|
|
|
|
id = rr->items[i].util;
|
|
|
|
if (!id)
|
2008-07-09 14:58:57 +02:00
|
|
|
continue;
|
2015-07-23 23:23:24 +02:00
|
|
|
assert(id->variant >= 0);
|
|
|
|
if (0 < id->variant)
|
|
|
|
strbuf_addf(&buf, "%s.%d\t%s%c",
|
|
|
|
rerere_id_hex(id), id->variant,
|
|
|
|
rr->items[i].string, 0);
|
|
|
|
else
|
|
|
|
strbuf_addf(&buf, "%s\t%s%c",
|
|
|
|
rerere_id_hex(id),
|
|
|
|
rr->items[i].string, 0);
|
|
|
|
|
2015-06-29 01:28:00 +02:00
|
|
|
if (write_in_full(out_fd, buf.buf, buf.len) != buf.len)
|
2008-07-09 14:58:57 +02:00
|
|
|
die("unable to write rerere record");
|
2015-06-29 01:28:00 +02:00
|
|
|
|
|
|
|
strbuf_release(&buf);
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
if (commit_lock_file(&write_lock) != 0)
|
|
|
|
die("unable to write rerere record");
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:33:19 +02:00
|
|
|
/*
|
|
|
|
* "rerere" interacts with conflicted file contents using this I/O
|
|
|
|
* abstraction. It reads a conflicted contents from one place via
|
|
|
|
* "getline()" method, and optionally can write it out after
|
|
|
|
* normalizing the conflicted hunks to the "output". Subclasses of
|
|
|
|
* rerere_io embed this structure at the beginning of their own
|
|
|
|
* rerere_io object.
|
|
|
|
*/
|
|
|
|
struct rerere_io {
|
|
|
|
int (*getline)(struct strbuf *, struct rerere_io *);
|
|
|
|
FILE *output;
|
|
|
|
int wrerror;
|
|
|
|
/* some more stuff */
|
|
|
|
};
|
|
|
|
|
2008-12-05 01:35:48 +01:00
|
|
|
static void ferr_write(const void *p, size_t count, FILE *fp, int *err)
|
|
|
|
{
|
|
|
|
if (!count || *err)
|
|
|
|
return;
|
|
|
|
if (fwrite(p, count, 1, fp) != 1)
|
|
|
|
*err = errno;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void ferr_puts(const char *s, FILE *fp, int *err)
|
|
|
|
{
|
|
|
|
ferr_write(s, strlen(s), fp, err);
|
|
|
|
}
|
|
|
|
|
2009-12-25 23:34:53 +01:00
|
|
|
static void rerere_io_putstr(const char *str, struct rerere_io *io)
|
|
|
|
{
|
|
|
|
if (io->output)
|
|
|
|
ferr_puts(str, io->output, &io->wrerror);
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:33:19 +02:00
|
|
|
/*
|
|
|
|
* Write a conflict marker to io->output (if defined).
|
|
|
|
*/
|
2010-01-17 08:06:45 +01:00
|
|
|
static void rerere_io_putconflict(int ch, int size, struct rerere_io *io)
|
|
|
|
{
|
|
|
|
char buf[64];
|
|
|
|
|
|
|
|
while (size) {
|
2015-07-25 01:01:48 +02:00
|
|
|
if (size <= sizeof(buf) - 2) {
|
2010-01-17 08:06:45 +01:00
|
|
|
memset(buf, ch, size);
|
|
|
|
buf[size] = '\n';
|
|
|
|
buf[size + 1] = '\0';
|
|
|
|
size = 0;
|
|
|
|
} else {
|
|
|
|
int sz = sizeof(buf) - 1;
|
2015-07-25 01:01:48 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Make sure we will not write everything out
|
|
|
|
* in this round by leaving at least 1 byte
|
|
|
|
* for the next round, giving the next round
|
|
|
|
* a chance to add the terminating LF. Yuck.
|
|
|
|
*/
|
2010-01-17 08:06:45 +01:00
|
|
|
if (size <= sz)
|
|
|
|
sz -= (sz - size) + 1;
|
|
|
|
memset(buf, ch, sz);
|
|
|
|
buf[sz] = '\0';
|
|
|
|
size -= sz;
|
|
|
|
}
|
|
|
|
rerere_io_putstr(buf, io);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-12-25 23:34:53 +01:00
|
|
|
static void rerere_io_putmem(const char *mem, size_t sz, struct rerere_io *io)
|
|
|
|
{
|
|
|
|
if (io->output)
|
|
|
|
ferr_write(mem, sz, io->output, &io->wrerror);
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:33:19 +02:00
|
|
|
/*
|
|
|
|
* Subclass of rerere_io that reads from an on-disk file
|
|
|
|
*/
|
2009-12-25 23:34:53 +01:00
|
|
|
struct rerere_io_file {
|
|
|
|
struct rerere_io io;
|
|
|
|
FILE *input;
|
|
|
|
};
|
|
|
|
|
2015-07-01 07:33:19 +02:00
|
|
|
/*
|
|
|
|
* ... and its getline() method implementation
|
|
|
|
*/
|
2009-12-25 23:34:53 +01:00
|
|
|
static int rerere_file_getline(struct strbuf *sb, struct rerere_io *io_)
|
|
|
|
{
|
|
|
|
struct rerere_io_file *io = (struct rerere_io_file *)io_;
|
|
|
|
return strbuf_getwholeline(sb, io->input, '\n');
|
|
|
|
}
|
|
|
|
|
2015-06-30 00:05:24 +02:00
|
|
|
/*
|
|
|
|
* Require the exact number of conflict marker letters, no more, no
|
|
|
|
* less, followed by SP or any whitespace
|
|
|
|
* (including LF).
|
|
|
|
*/
|
|
|
|
static int is_cmarker(char *buf, int marker_char, int marker_size)
|
2010-01-17 08:06:45 +01:00
|
|
|
{
|
2015-06-30 00:05:24 +02:00
|
|
|
int want_sp;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The beginning of our version and the end of their version
|
|
|
|
* always are labeled like "<<<<< ours" or ">>>>> theirs",
|
|
|
|
* hence we set want_sp for them. Note that the version from
|
|
|
|
* the common ancestor in diff3-style output is not always
|
|
|
|
* labelled (e.g. "||||| common" is often seen but "|||||"
|
|
|
|
* alone is also valid), so we do not set want_sp.
|
|
|
|
*/
|
|
|
|
want_sp = (marker_char == '<') || (marker_char == '>');
|
|
|
|
|
2010-01-17 08:06:45 +01:00
|
|
|
while (marker_size--)
|
|
|
|
if (*buf++ != marker_char)
|
|
|
|
return 0;
|
|
|
|
if (want_sp && *buf != ' ')
|
|
|
|
return 0;
|
|
|
|
return isspace(*buf);
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:40:35 +02:00
|
|
|
/*
|
|
|
|
* Read contents a file with conflicts, normalize the conflicts
|
|
|
|
* by (1) discarding the common ancestor version in diff3-style,
|
|
|
|
* (2) reordering our side and their side so that whichever sorts
|
|
|
|
* alphabetically earlier comes before the other one, while
|
|
|
|
* computing the "conflict ID", which is just an SHA-1 hash of
|
|
|
|
* one side of the conflict, NUL, the other side of the conflict,
|
|
|
|
* and NUL concatenated together.
|
|
|
|
*
|
|
|
|
* Return the number of conflict hunks found.
|
|
|
|
*
|
|
|
|
* NEEDSWORK: the logic and theory of operation behind this conflict
|
|
|
|
* normalization may deserve to be documented somewhere, perhaps in
|
|
|
|
* Documentation/technical/rerere.txt.
|
|
|
|
*/
|
2010-01-17 08:06:45 +01:00
|
|
|
static int handle_path(unsigned char *sha1, struct rerere_io *io, int marker_size)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2008-10-01 20:05:20 +02:00
|
|
|
git_SHA_CTX ctx;
|
2008-08-29 19:12:23 +02:00
|
|
|
int hunk_no = 0;
|
|
|
|
enum {
|
2010-05-14 11:31:35 +02:00
|
|
|
RR_CONTEXT = 0, RR_SIDE_1, RR_SIDE_2, RR_ORIGINAL
|
2008-08-29 19:12:23 +02:00
|
|
|
} hunk = RR_CONTEXT;
|
2008-10-09 21:12:12 +02:00
|
|
|
struct strbuf one = STRBUF_INIT, two = STRBUF_INIT;
|
2009-12-25 22:55:29 +01:00
|
|
|
struct strbuf buf = STRBUF_INIT;
|
2008-07-09 14:58:57 +02:00
|
|
|
|
|
|
|
if (sha1)
|
2008-10-01 20:05:20 +02:00
|
|
|
git_SHA1_Init(&ctx);
|
2008-07-09 14:58:57 +02:00
|
|
|
|
2009-12-25 23:34:53 +01:00
|
|
|
while (!io->getline(&buf, io)) {
|
2015-06-30 00:05:24 +02:00
|
|
|
if (is_cmarker(buf.buf, '<', marker_size)) {
|
2008-08-29 19:12:23 +02:00
|
|
|
if (hunk != RR_CONTEXT)
|
2008-07-09 14:58:57 +02:00
|
|
|
goto bad;
|
2008-08-29 19:12:23 +02:00
|
|
|
hunk = RR_SIDE_1;
|
2015-06-30 00:05:24 +02:00
|
|
|
} else if (is_cmarker(buf.buf, '|', marker_size)) {
|
2008-08-29 19:12:23 +02:00
|
|
|
if (hunk != RR_SIDE_1)
|
2008-07-09 14:58:57 +02:00
|
|
|
goto bad;
|
2008-08-29 19:24:45 +02:00
|
|
|
hunk = RR_ORIGINAL;
|
2015-06-30 00:05:24 +02:00
|
|
|
} else if (is_cmarker(buf.buf, '=', marker_size)) {
|
2008-08-29 19:24:45 +02:00
|
|
|
if (hunk != RR_SIDE_1 && hunk != RR_ORIGINAL)
|
2008-07-09 14:58:57 +02:00
|
|
|
goto bad;
|
2008-08-29 19:12:23 +02:00
|
|
|
hunk = RR_SIDE_2;
|
2015-06-30 00:05:24 +02:00
|
|
|
} else if (is_cmarker(buf.buf, '>', marker_size)) {
|
2008-08-29 19:12:23 +02:00
|
|
|
if (hunk != RR_SIDE_2)
|
2008-07-09 14:58:57 +02:00
|
|
|
goto bad;
|
|
|
|
if (strbuf_cmp(&one, &two) > 0)
|
|
|
|
strbuf_swap(&one, &two);
|
|
|
|
hunk_no++;
|
2008-08-29 19:12:23 +02:00
|
|
|
hunk = RR_CONTEXT;
|
2010-01-17 08:06:45 +01:00
|
|
|
rerere_io_putconflict('<', marker_size, io);
|
2009-12-25 23:34:53 +01:00
|
|
|
rerere_io_putmem(one.buf, one.len, io);
|
2010-01-17 08:06:45 +01:00
|
|
|
rerere_io_putconflict('=', marker_size, io);
|
2009-12-25 23:34:53 +01:00
|
|
|
rerere_io_putmem(two.buf, two.len, io);
|
2010-01-17 08:06:45 +01:00
|
|
|
rerere_io_putconflict('>', marker_size, io);
|
2008-07-09 14:58:57 +02:00
|
|
|
if (sha1) {
|
2008-10-01 20:05:20 +02:00
|
|
|
git_SHA1_Update(&ctx, one.buf ? one.buf : "",
|
2008-07-09 14:58:57 +02:00
|
|
|
one.len + 1);
|
2008-10-01 20:05:20 +02:00
|
|
|
git_SHA1_Update(&ctx, two.buf ? two.buf : "",
|
2008-07-09 14:58:57 +02:00
|
|
|
two.len + 1);
|
|
|
|
}
|
|
|
|
strbuf_reset(&one);
|
|
|
|
strbuf_reset(&two);
|
2008-08-29 19:12:23 +02:00
|
|
|
} else if (hunk == RR_SIDE_1)
|
2014-07-10 10:52:21 +02:00
|
|
|
strbuf_addbuf(&one, &buf);
|
2008-08-29 19:24:45 +02:00
|
|
|
else if (hunk == RR_ORIGINAL)
|
|
|
|
; /* discard */
|
2008-08-29 19:12:23 +02:00
|
|
|
else if (hunk == RR_SIDE_2)
|
2014-07-10 10:52:21 +02:00
|
|
|
strbuf_addbuf(&two, &buf);
|
2009-12-25 23:34:53 +01:00
|
|
|
else
|
|
|
|
rerere_io_putstr(buf.buf, io);
|
2008-07-09 14:58:57 +02:00
|
|
|
continue;
|
|
|
|
bad:
|
|
|
|
hunk = 99; /* force error exit */
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
strbuf_release(&one);
|
|
|
|
strbuf_release(&two);
|
2009-12-25 22:55:29 +01:00
|
|
|
strbuf_release(&buf);
|
2008-07-09 14:58:57 +02:00
|
|
|
|
|
|
|
if (sha1)
|
2008-10-01 20:05:20 +02:00
|
|
|
git_SHA1_Final(sha1, &ctx);
|
2009-12-25 23:34:53 +01:00
|
|
|
if (hunk != RR_CONTEXT)
|
|
|
|
return -1;
|
|
|
|
return hunk_no;
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:40:35 +02:00
|
|
|
/*
|
|
|
|
* Scan the path for conflicts, do the "handle_path()" thing above, and
|
|
|
|
* return the number of conflict hunks found.
|
|
|
|
*/
|
2009-12-25 23:34:53 +01:00
|
|
|
static int handle_file(const char *path, unsigned char *sha1, const char *output)
|
|
|
|
{
|
|
|
|
int hunk_no = 0;
|
|
|
|
struct rerere_io_file io;
|
2010-01-17 08:28:46 +01:00
|
|
|
int marker_size = ll_merge_marker_size(path);
|
2009-12-25 23:34:53 +01:00
|
|
|
|
|
|
|
memset(&io, 0, sizeof(io));
|
|
|
|
io.io.getline = rerere_file_getline;
|
|
|
|
io.input = fopen(path, "r");
|
|
|
|
io.io.wrerror = 0;
|
|
|
|
if (!io.input)
|
|
|
|
return error("Could not open %s", path);
|
|
|
|
|
|
|
|
if (output) {
|
|
|
|
io.io.output = fopen(output, "w");
|
|
|
|
if (!io.io.output) {
|
|
|
|
fclose(io.input);
|
|
|
|
return error("Could not write %s", output);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-01-17 08:06:45 +01:00
|
|
|
hunk_no = handle_path(sha1, (struct rerere_io *)&io, marker_size);
|
2009-12-25 23:34:53 +01:00
|
|
|
|
|
|
|
fclose(io.input);
|
|
|
|
if (io.io.wrerror)
|
|
|
|
error("There were errors while writing %s (%s)",
|
|
|
|
path, strerror(io.io.wrerror));
|
|
|
|
if (io.io.output && fclose(io.io.output))
|
2016-05-08 11:47:52 +02:00
|
|
|
io.io.wrerror = error_errno("Failed to flush %s", path);
|
2009-12-25 23:34:53 +01:00
|
|
|
|
|
|
|
if (hunk_no < 0) {
|
2008-07-09 14:58:57 +02:00
|
|
|
if (output)
|
2009-04-29 23:22:56 +02:00
|
|
|
unlink_or_warn(output);
|
2008-07-09 14:58:57 +02:00
|
|
|
return error("Could not parse conflict hunks in %s", path);
|
|
|
|
}
|
2009-12-25 23:34:53 +01:00
|
|
|
if (io.io.wrerror)
|
2008-12-05 01:35:48 +01:00
|
|
|
return -1;
|
2008-07-09 14:58:57 +02:00
|
|
|
return hunk_no;
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:36:35 +02:00
|
|
|
/*
|
|
|
|
* Look at a cache entry at "i" and see if it is not conflicting,
|
|
|
|
* conflicting and we are willing to handle, or conflicting and
|
|
|
|
* we are unable to handle, and return the determination in *type.
|
|
|
|
* Return the cache index to be looked at next, by skipping the
|
|
|
|
* stages we have already looked at in this invocation of this
|
|
|
|
* function.
|
|
|
|
*/
|
2011-02-16 11:47:44 +01:00
|
|
|
static int check_one_conflict(int i, int *type)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
Convert "struct cache_entry *" to "const ..." wherever possible
I attempted to make index_state->cache[] a "const struct cache_entry **"
to find out how existing entries in index are modified and where. The
question I have is what do we do if we really need to keep track of on-disk
changes in the index. The result is
- diff-lib.c: setting CE_UPTODATE
- name-hash.c: setting CE_HASHED
- preload-index.c, read-cache.c, unpack-trees.c and
builtin/update-index: obvious
- entry.c: write_entry() may refresh the checked out entry via
fill_stat_cache_info(). This causes "non-const struct cache_entry
*" in builtin/apply.c, builtin/checkout-index.c and
builtin/checkout.c
- builtin/ls-files.c: --with-tree changes stagemask and may set
CE_UPDATE
Of these, write_entry() and its call sites are probably most
interesting because it modifies on-disk info. But this is stat info
and can be retrieved via refresh, at least for porcelain
commands. Other just uses ce_flags for local purposes.
So, keeping track of "dirty" entries is just a matter of setting a
flag in index modification functions exposed by read-cache.c. Except
unpack-trees, the rest of the code base does not do anything funny
behind read-cache's back.
The actual patch is less valueable than the summary above. But if
anyone wants to re-identify the above sites. Applying this patch, then
this:
diff --git a/cache.h b/cache.h
index 430d021..1692891 100644
--- a/cache.h
+++ b/cache.h
@@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode)
#define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1)
struct index_state {
- struct cache_entry **cache;
+ const struct cache_entry **cache;
unsigned int version;
unsigned int cache_nr, cache_alloc, cache_changed;
struct string_list *resolve_undo;
will help quickly identify them without bogus warnings.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
|
|
|
const struct cache_entry *e = active_cache[i];
|
2011-02-16 11:47:44 +01:00
|
|
|
|
|
|
|
if (!ce_stage(e)) {
|
|
|
|
*type = RESOLVED;
|
|
|
|
return i + 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
*type = PUNTED;
|
2015-07-25 00:08:03 +02:00
|
|
|
while (ce_stage(active_cache[i]) == 1)
|
2015-06-28 23:35:13 +02:00
|
|
|
i++;
|
2011-02-16 11:47:44 +01:00
|
|
|
|
|
|
|
/* Only handle regular files with both stages #2 and #3 */
|
|
|
|
if (i + 1 < active_nr) {
|
Convert "struct cache_entry *" to "const ..." wherever possible
I attempted to make index_state->cache[] a "const struct cache_entry **"
to find out how existing entries in index are modified and where. The
question I have is what do we do if we really need to keep track of on-disk
changes in the index. The result is
- diff-lib.c: setting CE_UPTODATE
- name-hash.c: setting CE_HASHED
- preload-index.c, read-cache.c, unpack-trees.c and
builtin/update-index: obvious
- entry.c: write_entry() may refresh the checked out entry via
fill_stat_cache_info(). This causes "non-const struct cache_entry
*" in builtin/apply.c, builtin/checkout-index.c and
builtin/checkout.c
- builtin/ls-files.c: --with-tree changes stagemask and may set
CE_UPDATE
Of these, write_entry() and its call sites are probably most
interesting because it modifies on-disk info. But this is stat info
and can be retrieved via refresh, at least for porcelain
commands. Other just uses ce_flags for local purposes.
So, keeping track of "dirty" entries is just a matter of setting a
flag in index modification functions exposed by read-cache.c. Except
unpack-trees, the rest of the code base does not do anything funny
behind read-cache's back.
The actual patch is less valueable than the summary above. But if
anyone wants to re-identify the above sites. Applying this patch, then
this:
diff --git a/cache.h b/cache.h
index 430d021..1692891 100644
--- a/cache.h
+++ b/cache.h
@@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode)
#define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1)
struct index_state {
- struct cache_entry **cache;
+ const struct cache_entry **cache;
unsigned int version;
unsigned int cache_nr, cache_alloc, cache_changed;
struct string_list *resolve_undo;
will help quickly identify them without bogus warnings.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
|
|
|
const struct cache_entry *e2 = active_cache[i];
|
|
|
|
const struct cache_entry *e3 = active_cache[i + 1];
|
2008-07-09 14:58:57 +02:00
|
|
|
if (ce_stage(e2) == 2 &&
|
|
|
|
ce_stage(e3) == 3 &&
|
2011-02-16 11:47:44 +01:00
|
|
|
ce_same_name(e, e3) &&
|
2008-07-09 14:58:57 +02:00
|
|
|
S_ISREG(e2->ce_mode) &&
|
2011-02-16 11:47:44 +01:00
|
|
|
S_ISREG(e3->ce_mode))
|
|
|
|
*type = THREE_STAGED;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Skip the entries with the same name */
|
|
|
|
while (i < active_nr && ce_same_name(e, active_cache[i]))
|
|
|
|
i++;
|
|
|
|
return i;
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:36:35 +02:00
|
|
|
/*
|
|
|
|
* Scan the index and find paths that have conflicts that rerere can
|
|
|
|
* handle, i.e. the ones that has both stages #2 and #3.
|
|
|
|
*
|
|
|
|
* NEEDSWORK: we do not record or replay a previous "resolve by
|
|
|
|
* deletion" for a delete-modify conflict, as that is inherently risky
|
|
|
|
* without knowing what modification is being discarded. The only
|
|
|
|
* safe case, i.e. both side doing the deletion and modification that
|
|
|
|
* are identical to the previous round, might want to be handled,
|
|
|
|
* though.
|
|
|
|
*/
|
2011-02-16 11:47:44 +01:00
|
|
|
static int find_conflict(struct string_list *conflict)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
if (read_cache() < 0)
|
|
|
|
return error("Could not read index");
|
|
|
|
|
|
|
|
for (i = 0; i < active_nr;) {
|
|
|
|
int conflict_type;
|
Convert "struct cache_entry *" to "const ..." wherever possible
I attempted to make index_state->cache[] a "const struct cache_entry **"
to find out how existing entries in index are modified and where. The
question I have is what do we do if we really need to keep track of on-disk
changes in the index. The result is
- diff-lib.c: setting CE_UPTODATE
- name-hash.c: setting CE_HASHED
- preload-index.c, read-cache.c, unpack-trees.c and
builtin/update-index: obvious
- entry.c: write_entry() may refresh the checked out entry via
fill_stat_cache_info(). This causes "non-const struct cache_entry
*" in builtin/apply.c, builtin/checkout-index.c and
builtin/checkout.c
- builtin/ls-files.c: --with-tree changes stagemask and may set
CE_UPDATE
Of these, write_entry() and its call sites are probably most
interesting because it modifies on-disk info. But this is stat info
and can be retrieved via refresh, at least for porcelain
commands. Other just uses ce_flags for local purposes.
So, keeping track of "dirty" entries is just a matter of setting a
flag in index modification functions exposed by read-cache.c. Except
unpack-trees, the rest of the code base does not do anything funny
behind read-cache's back.
The actual patch is less valueable than the summary above. But if
anyone wants to re-identify the above sites. Applying this patch, then
this:
diff --git a/cache.h b/cache.h
index 430d021..1692891 100644
--- a/cache.h
+++ b/cache.h
@@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode)
#define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1)
struct index_state {
- struct cache_entry **cache;
+ const struct cache_entry **cache;
unsigned int version;
unsigned int cache_nr, cache_alloc, cache_changed;
struct string_list *resolve_undo;
will help quickly identify them without bogus warnings.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
|
|
|
const struct cache_entry *e = active_cache[i];
|
2011-02-16 11:47:44 +01:00
|
|
|
i = check_one_conflict(i, &conflict_type);
|
|
|
|
if (conflict_type == THREE_STAGED)
|
|
|
|
string_list_insert(conflict, (const char *)e->name);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:36:35 +02:00
|
|
|
/*
|
|
|
|
* The merge_rr list is meant to hold outstanding conflicted paths
|
|
|
|
* that rerere could handle. Abuse the list by adding other types of
|
|
|
|
* entries to allow the caller to show "rerere remaining".
|
|
|
|
*
|
|
|
|
* - Conflicted paths that rerere does not handle are added
|
|
|
|
* - Conflicted paths that have been resolved are marked as such
|
|
|
|
* by storing RERERE_RESOLVED to .util field (where conflict ID
|
|
|
|
* is expected to be stored).
|
|
|
|
*
|
|
|
|
* Do *not* write MERGE_RR file out after calling this function.
|
|
|
|
*
|
|
|
|
* NEEDSWORK: we may want to fix the caller that implements "rerere
|
|
|
|
* remaining" to do this without abusing merge_rr.
|
|
|
|
*/
|
2011-02-16 11:47:44 +01:00
|
|
|
int rerere_remaining(struct string_list *merge_rr)
|
|
|
|
{
|
|
|
|
int i;
|
rerere: release lockfile in non-writing functions
There's a bug in builtin/am.c in which we take a lock on
MERGE_RR recursively. But rather than fix am.c, this patch
fixes the confusing interface from rerere.c that caused the
bug. Read on for the gory details.
The setup_rerere() function both reads the existing MERGE_RR
file, and takes MERGE_RR.lock. In the rerere() and
rerere_forget() functions, we end up in write_rr(), which
will then commit the lock file.
But for functions like rerere_clear() that do not write to
MERGE_RR, we expect the caller to have handled
setup_rerere(). That caller would then need to release the
lockfile, but it can't; the lock struct is local to
rerere.c.
For builtin/rerere.c, this is OK. We run a single rerere
operation and then exit immediately, which has the side
effect of rolling back the lockfile.
But in builtin/am.c, this is actively wrong. If we run "git
am -3 --skip", we call setup-rerere twice without releasing
the lock:
1. The "--skip" causes us to call am_rerere_clear(), which
calls setup_rerere(), but never drops the lock.
2. We then proceed to the next patch.
3. The "--3way" may cause us to call rerere() to handle
conflicts in that patch, but we are already holding the
lock. The lockfile code dies with:
BUG: prepare_tempfile_object called for active object
We could fix this by having rerere_clear() call
rollback_lock_file(). But it feels a bit odd for it to roll
back a lockfile that it did not itself take. So let's
simplify the interface further, and handle setup_rerere in
the function itself, taking away the question from the
caller over whether they need to do so.
We can give rerere_gc() the same treatment, as well (even
though it doesn't have any callers besides builtin/rerere.c
at this point). Note that these functions don't take flags
from their callers to pass along to setup_rerere; that's OK,
because the flags would not be meaningful for what they are
doing.
Both of those functions need to hold the lock because even
though they do not write to MERGE_RR, they are still writing
and should be protected from a simultaneous "rerere" run.
But rerere_remaining(), "rerere diff", and "rerere status"
are all read-only operations. They want to setup_rerere(),
but do not care about taking the lock in the first place.
Since our update of MERGE_RR is the usual atomic rename done
by commit_lock_file, they can just do a lockless read. For
that, we teach setup_rerere a READONLY flag to avoid the
lock.
As a bonus, this pushes builtin/rerere.c's setup_rerere call
closer to the functions that use it. Which means that "git
rerere totally-bogus-command" will no longer silently
exit(0) in a repository without rerere enabled.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-02 00:14:09 +02:00
|
|
|
if (setup_rerere(merge_rr, RERERE_READONLY))
|
|
|
|
return 0;
|
2011-02-16 11:47:44 +01:00
|
|
|
if (read_cache() < 0)
|
|
|
|
return error("Could not read index");
|
|
|
|
|
|
|
|
for (i = 0; i < active_nr;) {
|
|
|
|
int conflict_type;
|
Convert "struct cache_entry *" to "const ..." wherever possible
I attempted to make index_state->cache[] a "const struct cache_entry **"
to find out how existing entries in index are modified and where. The
question I have is what do we do if we really need to keep track of on-disk
changes in the index. The result is
- diff-lib.c: setting CE_UPTODATE
- name-hash.c: setting CE_HASHED
- preload-index.c, read-cache.c, unpack-trees.c and
builtin/update-index: obvious
- entry.c: write_entry() may refresh the checked out entry via
fill_stat_cache_info(). This causes "non-const struct cache_entry
*" in builtin/apply.c, builtin/checkout-index.c and
builtin/checkout.c
- builtin/ls-files.c: --with-tree changes stagemask and may set
CE_UPDATE
Of these, write_entry() and its call sites are probably most
interesting because it modifies on-disk info. But this is stat info
and can be retrieved via refresh, at least for porcelain
commands. Other just uses ce_flags for local purposes.
So, keeping track of "dirty" entries is just a matter of setting a
flag in index modification functions exposed by read-cache.c. Except
unpack-trees, the rest of the code base does not do anything funny
behind read-cache's back.
The actual patch is less valueable than the summary above. But if
anyone wants to re-identify the above sites. Applying this patch, then
this:
diff --git a/cache.h b/cache.h
index 430d021..1692891 100644
--- a/cache.h
+++ b/cache.h
@@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode)
#define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1)
struct index_state {
- struct cache_entry **cache;
+ const struct cache_entry **cache;
unsigned int version;
unsigned int cache_nr, cache_alloc, cache_changed;
struct string_list *resolve_undo;
will help quickly identify them without bogus warnings.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
|
|
|
const struct cache_entry *e = active_cache[i];
|
2011-02-16 11:47:44 +01:00
|
|
|
i = check_one_conflict(i, &conflict_type);
|
|
|
|
if (conflict_type == PUNTED)
|
|
|
|
string_list_insert(merge_rr, (const char *)e->name);
|
|
|
|
else if (conflict_type == RESOLVED) {
|
|
|
|
struct string_list_item *it;
|
|
|
|
it = string_list_lookup(merge_rr, (const char *)e->name);
|
|
|
|
if (it != NULL) {
|
2015-07-05 02:38:34 +02:00
|
|
|
free_rerere_id(it);
|
2011-02-16 11:47:44 +01:00
|
|
|
it->util = RERERE_RESOLVED;
|
|
|
|
}
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2016-03-11 23:53:05 +01:00
|
|
|
/*
|
|
|
|
* Try using the given conflict resolution "ID" to see
|
|
|
|
* if that recorded conflict resolves cleanly what we
|
|
|
|
* got in the "cur".
|
|
|
|
*/
|
|
|
|
static int try_merge(const struct rerere_id *id, const char *path,
|
|
|
|
mmfile_t *cur, mmbuffer_t *result)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
mmfile_t base = {NULL, 0}, other = {NULL, 0};
|
|
|
|
|
|
|
|
if (read_mmfile(&base, rerere_path(id, "preimage")) ||
|
|
|
|
read_mmfile(&other, rerere_path(id, "postimage")))
|
|
|
|
ret = 1;
|
|
|
|
else
|
|
|
|
/*
|
|
|
|
* A three-way merge. Note that this honors user-customizable
|
|
|
|
* low-level merge driver settings.
|
|
|
|
*/
|
|
|
|
ret = ll_merge(result, path, &base, NULL, cur, "", &other, "", NULL);
|
|
|
|
|
|
|
|
free(base.ptr);
|
|
|
|
free(other.ptr);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:40:35 +02:00
|
|
|
/*
|
2015-07-06 23:45:55 +02:00
|
|
|
* Find the conflict identified by "id"; the change between its
|
2015-07-01 07:40:35 +02:00
|
|
|
* "preimage" (i.e. a previous contents with conflict markers) and its
|
|
|
|
* "postimage" (i.e. the corresponding contents with conflicts
|
|
|
|
* resolved) may apply cleanly to the contents stored in "path", i.e.
|
|
|
|
* the conflict this time around.
|
|
|
|
*
|
|
|
|
* Returns 0 for successful replay of recorded resolution, or non-zero
|
|
|
|
* for failure.
|
|
|
|
*/
|
2015-07-05 02:38:34 +02:00
|
|
|
static int merge(const struct rerere_id *id, const char *path)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2015-07-07 00:32:53 +02:00
|
|
|
FILE *f;
|
2008-07-09 14:58:57 +02:00
|
|
|
int ret;
|
2016-03-11 23:53:05 +01:00
|
|
|
mmfile_t cur = {NULL, 0};
|
2008-07-09 14:58:57 +02:00
|
|
|
mmbuffer_t result = {NULL, 0};
|
|
|
|
|
2015-07-01 07:40:35 +02:00
|
|
|
/*
|
|
|
|
* Normalize the conflicts in path and write it out to
|
|
|
|
* "thisimage" temporary file.
|
|
|
|
*/
|
2016-03-11 23:53:05 +01:00
|
|
|
if ((handle_file(path, NULL, rerere_path(id, "thisimage")) < 0) ||
|
|
|
|
read_mmfile(&cur, rerere_path(id, "thisimage"))) {
|
2010-02-23 21:11:53 +01:00
|
|
|
ret = 1;
|
|
|
|
goto out;
|
|
|
|
}
|
2015-07-01 07:40:35 +02:00
|
|
|
|
2016-03-11 23:53:05 +01:00
|
|
|
ret = try_merge(id, path, &cur, &result);
|
2015-07-07 00:32:53 +02:00
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* A successful replay of recorded resolution.
|
|
|
|
* Mark that "postimage" was used to help gc.
|
|
|
|
*/
|
|
|
|
if (utime(rerere_path(id, "postimage"), NULL) < 0)
|
2016-05-08 11:47:52 +02:00
|
|
|
warning_errno("failed utime() on %s",
|
|
|
|
rerere_path(id, "postimage"));
|
2015-07-07 00:32:53 +02:00
|
|
|
|
|
|
|
/* Update "path" with the resolution */
|
|
|
|
f = fopen(path, "w");
|
|
|
|
if (!f)
|
2016-05-08 11:47:52 +02:00
|
|
|
return error_errno("Could not open %s", path);
|
2015-07-07 00:32:53 +02:00
|
|
|
if (fwrite(result.ptr, result.size, 1, f) != 1)
|
2016-05-08 11:47:52 +02:00
|
|
|
error_errno("Could not write %s", path);
|
2015-07-07 00:32:53 +02:00
|
|
|
if (fclose(f))
|
2016-05-08 11:47:52 +02:00
|
|
|
return error_errno("Writing %s failed", path);
|
2008-07-09 14:58:57 +02:00
|
|
|
|
2010-02-23 21:11:53 +01:00
|
|
|
out:
|
2008-07-09 14:58:57 +02:00
|
|
|
free(cur.ptr);
|
|
|
|
free(result.ptr);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct lock_file index_lock;
|
|
|
|
|
2014-12-03 05:20:49 +01:00
|
|
|
static void update_paths(struct string_list *update)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2014-12-03 05:20:49 +01:00
|
|
|
hold_locked_index(&index_lock, 1);
|
2008-07-09 14:58:57 +02:00
|
|
|
|
|
|
|
for (i = 0; i < update->nr; i++) {
|
2008-07-21 20:03:49 +02:00
|
|
|
struct string_list_item *item = &update->items[i];
|
2014-12-03 05:20:49 +01:00
|
|
|
if (add_file_to_cache(item->string, 0))
|
|
|
|
exit(128);
|
2015-06-29 06:13:24 +02:00
|
|
|
fprintf(stderr, "Staged '%s' using previous resolution.\n",
|
|
|
|
item->string);
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
2014-12-03 05:20:49 +01:00
|
|
|
if (active_cache_changed) {
|
2014-06-13 14:19:23 +02:00
|
|
|
if (write_locked_index(&the_index, &index_lock, COMMIT_LOCK))
|
2008-07-09 14:58:57 +02:00
|
|
|
die("Unable to write new index file");
|
2014-12-03 05:20:49 +01:00
|
|
|
} else
|
2008-07-09 14:58:57 +02:00
|
|
|
rollback_lock_file(&index_lock);
|
|
|
|
}
|
|
|
|
|
2015-07-31 00:49:18 +02:00
|
|
|
static void remove_variant(struct rerere_id *id)
|
|
|
|
{
|
|
|
|
unlink_or_warn(rerere_path(id, "postimage"));
|
|
|
|
unlink_or_warn(rerere_path(id, "preimage"));
|
|
|
|
id->collection->status[id->variant] = 0;
|
|
|
|
}
|
|
|
|
|
2015-07-01 04:36:24 +02:00
|
|
|
/*
|
|
|
|
* The path indicated by rr_item may still have conflict for which we
|
|
|
|
* have a recorded resolution, in which case replay it and optionally
|
|
|
|
* update it. Or it may have been resolved by the user and we may
|
|
|
|
* only have the preimage for that conflict, in which case the result
|
|
|
|
* needs to be recorded as a resolution in a postimage file.
|
|
|
|
*/
|
|
|
|
static void do_rerere_one_path(struct string_list_item *rr_item,
|
|
|
|
struct string_list *update)
|
|
|
|
{
|
|
|
|
const char *path = rr_item->string;
|
2015-07-23 23:23:24 +02:00
|
|
|
struct rerere_id *id = rr_item->util;
|
2015-07-31 00:49:18 +02:00
|
|
|
struct rerere_dir *rr_dir = id->collection;
|
2015-07-23 23:23:24 +02:00
|
|
|
int variant;
|
|
|
|
|
|
|
|
variant = id->variant;
|
2015-07-01 04:36:24 +02:00
|
|
|
|
2015-07-31 00:49:18 +02:00
|
|
|
/* Has the user resolved it already? */
|
|
|
|
if (variant >= 0) {
|
|
|
|
if (!handle_file(path, NULL, NULL)) {
|
|
|
|
copy_file(rerere_path(id, "postimage"), path, 0666);
|
|
|
|
id->collection->status[variant] |= RR_HAS_POSTIMAGE;
|
|
|
|
fprintf(stderr, "Recorded resolution for '%s'.\n", path);
|
|
|
|
free_rerere_id(rr_item);
|
|
|
|
rr_item->util = NULL;
|
|
|
|
return;
|
|
|
|
}
|
rerere: delay the recording of preimage
We record the preimage only when there is no directory to record the
conflict we encountered, i.e. when $GIT_DIR/rr-cache/$ID does not
exist. As the plan is to allow multiple <preimage,postimage> pairs
as variants for the same conflict ID eventually, this logic needs to
go.
As the first step in that direction, stop the "did we create the
directory? Then we record the preimage" logic. Instead, we record
if a preimage does not exist when we saw a conflict in a path. Also
make sure that we remove a stale postimage, which most likely is
totally unrelated to the resolution of this new conflict, when we
create a new preimage under $ID when $GIT_DIR/rr-cache/$ID already
exists.
In later patches, we will further update this logic to be "do we
have <preimage,postimage> pair that cleanly resolve the current
conflicts? If not, record a new preimage as a new variant", but
that does not happen at this stage yet.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-21 00:19:44 +02:00
|
|
|
/*
|
2015-07-31 00:49:18 +02:00
|
|
|
* There may be other variants that can cleanly
|
|
|
|
* replay. Try them and update the variant number for
|
|
|
|
* this one.
|
rerere: delay the recording of preimage
We record the preimage only when there is no directory to record the
conflict we encountered, i.e. when $GIT_DIR/rr-cache/$ID does not
exist. As the plan is to allow multiple <preimage,postimage> pairs
as variants for the same conflict ID eventually, this logic needs to
go.
As the first step in that direction, stop the "did we create the
directory? Then we record the preimage" logic. Instead, we record
if a preimage does not exist when we saw a conflict in a path. Also
make sure that we remove a stale postimage, which most likely is
totally unrelated to the resolution of this new conflict, when we
create a new preimage under $ID when $GIT_DIR/rr-cache/$ID already
exists.
In later patches, we will further update this logic to be "do we
have <preimage,postimage> pair that cleanly resolve the current
conflicts? If not, record a new preimage as a new variant", but
that does not happen at this stage yet.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-21 00:19:44 +02:00
|
|
|
*/
|
2015-07-31 00:49:18 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Does any existing resolution apply cleanly? */
|
|
|
|
for (variant = 0; variant < rr_dir->status_nr; variant++) {
|
|
|
|
const int both = RR_HAS_PREIMAGE | RR_HAS_POSTIMAGE;
|
|
|
|
struct rerere_id vid = *id;
|
|
|
|
|
|
|
|
if ((rr_dir->status[variant] & both) != both)
|
|
|
|
continue;
|
2015-07-01 04:36:24 +02:00
|
|
|
|
2015-07-31 00:49:18 +02:00
|
|
|
vid.variant = variant;
|
|
|
|
if (merge(&vid, path))
|
|
|
|
continue; /* failed to replay */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If there already is a different variant that applies
|
|
|
|
* cleanly, there is no point maintaining our own variant.
|
|
|
|
*/
|
|
|
|
if (0 <= id->variant && id->variant != variant)
|
|
|
|
remove_variant(id);
|
2015-07-01 04:36:24 +02:00
|
|
|
|
|
|
|
if (rerere_autoupdate)
|
|
|
|
string_list_insert(update, path);
|
|
|
|
else
|
|
|
|
fprintf(stderr,
|
|
|
|
"Resolved '%s' using previous resolution.\n",
|
|
|
|
path);
|
2015-07-31 00:49:18 +02:00
|
|
|
free_rerere_id(rr_item);
|
|
|
|
rr_item->util = NULL;
|
2015-07-06 23:18:09 +02:00
|
|
|
return;
|
2015-07-01 04:36:24 +02:00
|
|
|
}
|
2015-07-31 00:49:18 +02:00
|
|
|
|
|
|
|
/* None of the existing one applies; we need a new variant */
|
|
|
|
assign_variant(id);
|
|
|
|
|
|
|
|
variant = id->variant;
|
|
|
|
handle_file(path, NULL, rerere_path(id, "preimage"));
|
|
|
|
if (id->collection->status[variant] & RR_HAS_POSTIMAGE) {
|
|
|
|
const char *path = rerere_path(id, "postimage");
|
|
|
|
if (unlink(path))
|
|
|
|
die_errno("cannot unlink stray '%s'", path);
|
|
|
|
id->collection->status[variant] &= ~RR_HAS_POSTIMAGE;
|
|
|
|
}
|
|
|
|
id->collection->status[variant] |= RR_HAS_PREIMAGE;
|
|
|
|
fprintf(stderr, "Recorded preimage for '%s'\n", path);
|
2015-07-01 04:36:24 +02:00
|
|
|
}
|
|
|
|
|
2008-07-21 20:03:49 +02:00
|
|
|
static int do_plain_rerere(struct string_list *rr, int fd)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2010-07-04 21:46:19 +02:00
|
|
|
struct string_list conflict = STRING_LIST_INIT_DUP;
|
|
|
|
struct string_list update = STRING_LIST_INIT_DUP;
|
2008-07-09 14:58:57 +02:00
|
|
|
int i;
|
|
|
|
|
|
|
|
find_conflict(&conflict);
|
|
|
|
|
|
|
|
/*
|
2015-07-01 07:40:35 +02:00
|
|
|
* MERGE_RR records paths with conflicts immediately after
|
|
|
|
* merge failed. Some of the conflicted paths might have been
|
|
|
|
* hand resolved in the working tree since then, but the
|
|
|
|
* initial run would catch all and register their preimages.
|
2008-07-09 14:58:57 +02:00
|
|
|
*/
|
|
|
|
for (i = 0; i < conflict.nr; i++) {
|
2015-07-05 02:38:34 +02:00
|
|
|
struct rerere_id *id;
|
2015-07-05 02:17:38 +02:00
|
|
|
unsigned char sha1[20];
|
2008-07-21 20:03:49 +02:00
|
|
|
const char *path = conflict.items[i].string;
|
2015-07-05 02:38:34 +02:00
|
|
|
int ret;
|
2008-07-09 14:58:57 +02:00
|
|
|
|
2015-07-05 02:17:38 +02:00
|
|
|
if (string_list_has_string(rr, path))
|
|
|
|
continue;
|
2008-07-09 14:58:57 +02:00
|
|
|
|
2015-07-05 02:17:38 +02:00
|
|
|
/*
|
|
|
|
* Ask handle_file() to scan and assign a
|
|
|
|
* conflict ID. No need to write anything out
|
|
|
|
* yet.
|
|
|
|
*/
|
|
|
|
ret = handle_file(path, sha1, NULL);
|
|
|
|
if (ret < 1)
|
|
|
|
continue;
|
2008-07-09 14:58:57 +02:00
|
|
|
|
2015-07-05 02:38:34 +02:00
|
|
|
id = new_rerere_id(sha1);
|
2015-07-06 23:45:55 +02:00
|
|
|
string_list_insert(rr, path)->util = id;
|
2015-07-01 07:40:35 +02:00
|
|
|
|
rerere: delay the recording of preimage
We record the preimage only when there is no directory to record the
conflict we encountered, i.e. when $GIT_DIR/rr-cache/$ID does not
exist. As the plan is to allow multiple <preimage,postimage> pairs
as variants for the same conflict ID eventually, this logic needs to
go.
As the first step in that direction, stop the "did we create the
directory? Then we record the preimage" logic. Instead, we record
if a preimage does not exist when we saw a conflict in a path. Also
make sure that we remove a stale postimage, which most likely is
totally unrelated to the resolution of this new conflict, when we
create a new preimage under $ID when $GIT_DIR/rr-cache/$ID already
exists.
In later patches, we will further update this logic to be "do we
have <preimage,postimage> pair that cleanly resolve the current
conflicts? If not, record a new preimage as a new variant", but
that does not happen at this stage yet.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-21 00:19:44 +02:00
|
|
|
/* Ensure that the directory exists. */
|
|
|
|
mkdir_in_gitdir(rerere_path(id, NULL));
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
2015-07-01 04:36:24 +02:00
|
|
|
for (i = 0; i < rr->nr; i++)
|
|
|
|
do_rerere_one_path(&rr->items[i], &update);
|
2008-07-09 14:58:57 +02:00
|
|
|
|
|
|
|
if (update.nr)
|
|
|
|
update_paths(&update);
|
|
|
|
|
|
|
|
return write_rr(rr, fd);
|
|
|
|
}
|
|
|
|
|
2014-08-07 18:21:21 +02:00
|
|
|
static void git_rerere_config(void)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2014-08-07 18:21:21 +02:00
|
|
|
git_config_get_bool("rerere.enabled", &rerere_enabled);
|
|
|
|
git_config_get_bool("rerere.autoupdate", &rerere_autoupdate);
|
|
|
|
git_config(git_default_config, NULL);
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
|
|
|
|
memoize common git-path "constant" files
One of the most common uses of git_path() is to pass a
constant, like git_path("MERGE_MSG"). This has two
drawbacks:
1. The return value is a static buffer, and the lifetime
is dependent on other calls to git_path, etc.
2. There's no compile-time checking of the pathname. This
is OK for a one-off (after all, we have to spell it
correctly at least once), but many of these constant
strings appear throughout the code.
This patch introduces a series of functions to "memoize"
these strings, which are essentially globals for the
lifetime of the program. We compute the value once, take
ownership of the buffer, and return the cached value for
subsequent calls. cache.h provides a helper macro for
defining these functions as one-liners, and defines a few
common ones for global use.
Using a macro is a little bit gross, but it does nicely
document the purpose of the functions. If we need to touch
them all later (e.g., because we learned how to change the
git_dir variable at runtime, and need to invalidate all of
the stored values), it will be much easier to have the
complete list.
Note that the shared-global functions have separate, manual
declarations. We could do something clever with the macros
(e.g., expand it to a declaration in some places, and a
declaration _and_ a definition in path.c). But there aren't
that many, and it's probably better to stay away from
too-magical macros.
Likewise, if we abandon the C preprocessor in favor of
generating these with a script, we could get much fancier.
E.g., normalizing "FOO/BAR-BAZ" into "git_path_foo_bar_baz".
But the small amount of saved typing is probably not worth
the resulting confusion to readers who want to grep for the
function's definition.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-10 11:38:57 +02:00
|
|
|
static GIT_PATH_FUNC(git_path_rr_cache, "rr-cache")
|
|
|
|
|
2008-07-09 14:58:57 +02:00
|
|
|
static int is_rerere_enabled(void)
|
|
|
|
{
|
|
|
|
int rr_cache_exists;
|
|
|
|
|
|
|
|
if (!rerere_enabled)
|
|
|
|
return 0;
|
|
|
|
|
memoize common git-path "constant" files
One of the most common uses of git_path() is to pass a
constant, like git_path("MERGE_MSG"). This has two
drawbacks:
1. The return value is a static buffer, and the lifetime
is dependent on other calls to git_path, etc.
2. There's no compile-time checking of the pathname. This
is OK for a one-off (after all, we have to spell it
correctly at least once), but many of these constant
strings appear throughout the code.
This patch introduces a series of functions to "memoize"
these strings, which are essentially globals for the
lifetime of the program. We compute the value once, take
ownership of the buffer, and return the cached value for
subsequent calls. cache.h provides a helper macro for
defining these functions as one-liners, and defines a few
common ones for global use.
Using a macro is a little bit gross, but it does nicely
document the purpose of the functions. If we need to touch
them all later (e.g., because we learned how to change the
git_dir variable at runtime, and need to invalidate all of
the stored values), it will be much easier to have the
complete list.
Note that the shared-global functions have separate, manual
declarations. We could do something clever with the macros
(e.g., expand it to a declaration in some places, and a
declaration _and_ a definition in path.c). But there aren't
that many, and it's probably better to stay away from
too-magical macros.
Likewise, if we abandon the C preprocessor in favor of
generating these with a script, we could get much fancier.
E.g., normalizing "FOO/BAR-BAZ" into "git_path_foo_bar_baz".
But the small amount of saved typing is probably not worth
the resulting confusion to readers who want to grep for the
function's definition.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-10 11:38:57 +02:00
|
|
|
rr_cache_exists = is_directory(git_path_rr_cache());
|
2008-07-09 14:58:57 +02:00
|
|
|
if (rerere_enabled < 0)
|
|
|
|
return rr_cache_exists;
|
|
|
|
|
memoize common git-path "constant" files
One of the most common uses of git_path() is to pass a
constant, like git_path("MERGE_MSG"). This has two
drawbacks:
1. The return value is a static buffer, and the lifetime
is dependent on other calls to git_path, etc.
2. There's no compile-time checking of the pathname. This
is OK for a one-off (after all, we have to spell it
correctly at least once), but many of these constant
strings appear throughout the code.
This patch introduces a series of functions to "memoize"
these strings, which are essentially globals for the
lifetime of the program. We compute the value once, take
ownership of the buffer, and return the cached value for
subsequent calls. cache.h provides a helper macro for
defining these functions as one-liners, and defines a few
common ones for global use.
Using a macro is a little bit gross, but it does nicely
document the purpose of the functions. If we need to touch
them all later (e.g., because we learned how to change the
git_dir variable at runtime, and need to invalidate all of
the stored values), it will be much easier to have the
complete list.
Note that the shared-global functions have separate, manual
declarations. We could do something clever with the macros
(e.g., expand it to a declaration in some places, and a
declaration _and_ a definition in path.c). But there aren't
that many, and it's probably better to stay away from
too-magical macros.
Likewise, if we abandon the C preprocessor in favor of
generating these with a script, we could get much fancier.
E.g., normalizing "FOO/BAR-BAZ" into "git_path_foo_bar_baz".
But the small amount of saved typing is probably not worth
the resulting confusion to readers who want to grep for the
function's definition.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-10 11:38:57 +02:00
|
|
|
if (!rr_cache_exists && mkdir_in_gitdir(git_path_rr_cache()))
|
|
|
|
die("Could not create directory %s", git_path_rr_cache());
|
2008-07-09 14:58:57 +02:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2009-12-04 09:20:48 +01:00
|
|
|
int setup_rerere(struct string_list *merge_rr, int flags)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
|
|
|
int fd;
|
|
|
|
|
2014-08-07 18:21:21 +02:00
|
|
|
git_rerere_config();
|
2008-07-09 14:58:57 +02:00
|
|
|
if (!is_rerere_enabled())
|
|
|
|
return -1;
|
|
|
|
|
2009-12-04 09:20:48 +01:00
|
|
|
if (flags & (RERERE_AUTOUPDATE|RERERE_NOAUTOUPDATE))
|
|
|
|
rerere_autoupdate = !!(flags & RERERE_AUTOUPDATE);
|
rerere: release lockfile in non-writing functions
There's a bug in builtin/am.c in which we take a lock on
MERGE_RR recursively. But rather than fix am.c, this patch
fixes the confusing interface from rerere.c that caused the
bug. Read on for the gory details.
The setup_rerere() function both reads the existing MERGE_RR
file, and takes MERGE_RR.lock. In the rerere() and
rerere_forget() functions, we end up in write_rr(), which
will then commit the lock file.
But for functions like rerere_clear() that do not write to
MERGE_RR, we expect the caller to have handled
setup_rerere(). That caller would then need to release the
lockfile, but it can't; the lock struct is local to
rerere.c.
For builtin/rerere.c, this is OK. We run a single rerere
operation and then exit immediately, which has the side
effect of rolling back the lockfile.
But in builtin/am.c, this is actively wrong. If we run "git
am -3 --skip", we call setup-rerere twice without releasing
the lock:
1. The "--skip" causes us to call am_rerere_clear(), which
calls setup_rerere(), but never drops the lock.
2. We then proceed to the next patch.
3. The "--3way" may cause us to call rerere() to handle
conflicts in that patch, but we are already holding the
lock. The lockfile code dies with:
BUG: prepare_tempfile_object called for active object
We could fix this by having rerere_clear() call
rollback_lock_file(). But it feels a bit odd for it to roll
back a lockfile that it did not itself take. So let's
simplify the interface further, and handle setup_rerere in
the function itself, taking away the question from the
caller over whether they need to do so.
We can give rerere_gc() the same treatment, as well (even
though it doesn't have any callers besides builtin/rerere.c
at this point). Note that these functions don't take flags
from their callers to pass along to setup_rerere; that's OK,
because the flags would not be meaningful for what they are
doing.
Both of those functions need to hold the lock because even
though they do not write to MERGE_RR, they are still writing
and should be protected from a simultaneous "rerere" run.
But rerere_remaining(), "rerere diff", and "rerere status"
are all read-only operations. They want to setup_rerere(),
but do not care about taking the lock in the first place.
Since our update of MERGE_RR is the usual atomic rename done
by commit_lock_file, they can just do a lockless read. For
that, we teach setup_rerere a READONLY flag to avoid the
lock.
As a bonus, this pushes builtin/rerere.c's setup_rerere call
closer to the functions that use it. Which means that "git
rerere totally-bogus-command" will no longer silently
exit(0) in a repository without rerere enabled.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-02 00:14:09 +02:00
|
|
|
if (flags & RERERE_READONLY)
|
|
|
|
fd = 0;
|
|
|
|
else
|
|
|
|
fd = hold_lock_file_for_update(&write_lock, git_path_merge_rr(),
|
|
|
|
LOCK_DIE_ON_ERROR);
|
2008-07-09 14:58:57 +02:00
|
|
|
read_rr(merge_rr);
|
|
|
|
return fd;
|
|
|
|
}
|
|
|
|
|
2015-07-01 07:40:35 +02:00
|
|
|
/*
|
|
|
|
* The main entry point that is called internally from codepaths that
|
|
|
|
* perform mergy operations, possibly leaving conflicted index entries
|
|
|
|
* and working tree files.
|
|
|
|
*/
|
2009-12-04 09:20:48 +01:00
|
|
|
int rerere(int flags)
|
2008-07-09 14:58:57 +02:00
|
|
|
{
|
2010-07-04 21:46:19 +02:00
|
|
|
struct string_list merge_rr = STRING_LIST_INIT_DUP;
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
int fd, status;
|
2008-07-09 14:58:57 +02:00
|
|
|
|
2009-12-04 09:20:48 +01:00
|
|
|
fd = setup_rerere(&merge_rr, flags);
|
2008-07-09 14:58:57 +02:00
|
|
|
if (fd < 0)
|
|
|
|
return 0;
|
rerere: split conflict ID further
The plan is to keep assigning the backward compatible conflict ID
based on the hash of the (normalized) text of conflicts, keep using
that conflict ID as the directory name under $GIT_DIR/rr-cache/, but
allow each conflicted path to use a separate "variant" to record
resolutions, i.e. having more than one <preimage,postimage> pairs
under $GIT_DIR/rr-cache/$ID/ directory. As the first step in that
direction, separate the shared "conflict ID" out of the rerere_id
structure.
The plan is to keep information per $ID in rerere_dir, that can be
shared among rerere_id that is per conflicted path.
When we are done with rerere(), which can be directly called from
other programs like "git apply", "git commit" and "git merge", the
shared rerere_dir structures can be freed entirely, so they are not
reference-counted and they are not freed when we release rerere_id's
that reference them.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-07-16 23:50:05 +02:00
|
|
|
status = do_plain_rerere(&merge_rr, fd);
|
|
|
|
free_rerere_dirs();
|
|
|
|
return status;
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
2009-12-26 00:51:32 +01:00
|
|
|
|
2016-03-14 23:10:39 +01:00
|
|
|
/*
|
|
|
|
* Subclass of rerere_io that reads from an in-core buffer that is a
|
|
|
|
* strbuf
|
|
|
|
*/
|
|
|
|
struct rerere_io_mem {
|
|
|
|
struct rerere_io io;
|
|
|
|
struct strbuf input;
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ... and its getline() method implementation
|
|
|
|
*/
|
|
|
|
static int rerere_mem_getline(struct strbuf *sb, struct rerere_io *io_)
|
|
|
|
{
|
|
|
|
struct rerere_io_mem *io = (struct rerere_io_mem *)io_;
|
|
|
|
char *ep;
|
|
|
|
size_t len;
|
|
|
|
|
|
|
|
strbuf_release(sb);
|
|
|
|
if (!io->input.len)
|
|
|
|
return -1;
|
|
|
|
ep = memchr(io->input.buf, '\n', io->input.len);
|
|
|
|
if (!ep)
|
|
|
|
ep = io->input.buf + io->input.len;
|
|
|
|
else if (*ep == '\n')
|
|
|
|
ep++;
|
|
|
|
len = ep - io->input.buf;
|
|
|
|
strbuf_add(sb, io->input.buf, len);
|
|
|
|
strbuf_remove(&io->input, 0, len);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int handle_cache(const char *path, unsigned char *sha1, const char *output)
|
|
|
|
{
|
|
|
|
mmfile_t mmfile[3] = {{NULL}};
|
|
|
|
mmbuffer_t result = {NULL, 0};
|
|
|
|
const struct cache_entry *ce;
|
|
|
|
int pos, len, i, hunk_no;
|
|
|
|
struct rerere_io_mem io;
|
|
|
|
int marker_size = ll_merge_marker_size(path);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Reproduce the conflicted merge in-core
|
|
|
|
*/
|
|
|
|
len = strlen(path);
|
|
|
|
pos = cache_name_pos(path, len);
|
|
|
|
if (0 <= pos)
|
|
|
|
return -1;
|
|
|
|
pos = -pos - 1;
|
|
|
|
|
|
|
|
while (pos < active_nr) {
|
|
|
|
enum object_type type;
|
|
|
|
unsigned long size;
|
|
|
|
|
|
|
|
ce = active_cache[pos++];
|
|
|
|
if (ce_namelen(ce) != len || memcmp(ce->name, path, len))
|
|
|
|
break;
|
|
|
|
i = ce_stage(ce) - 1;
|
|
|
|
if (!mmfile[i].ptr) {
|
|
|
|
mmfile[i].ptr = read_sha1_file(ce->sha1, &type, &size);
|
|
|
|
mmfile[i].size = size;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (i = 0; i < 3; i++)
|
|
|
|
if (!mmfile[i].ptr && !mmfile[i].size)
|
|
|
|
mmfile[i].ptr = xstrdup("");
|
|
|
|
|
|
|
|
/*
|
|
|
|
* NEEDSWORK: handle conflicts from merges with
|
|
|
|
* merge.renormalize set, too?
|
|
|
|
*/
|
|
|
|
ll_merge(&result, path, &mmfile[0], NULL,
|
|
|
|
&mmfile[1], "ours",
|
|
|
|
&mmfile[2], "theirs", NULL);
|
|
|
|
for (i = 0; i < 3; i++)
|
|
|
|
free(mmfile[i].ptr);
|
|
|
|
|
|
|
|
memset(&io, 0, sizeof(io));
|
|
|
|
io.io.getline = rerere_mem_getline;
|
|
|
|
if (output)
|
|
|
|
io.io.output = fopen(output, "w");
|
|
|
|
else
|
|
|
|
io.io.output = NULL;
|
|
|
|
strbuf_init(&io.input, 0);
|
|
|
|
strbuf_attach(&io.input, result.ptr, result.size, result.size);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Grab the conflict ID and optionally write the original
|
|
|
|
* contents with conflict markers out.
|
|
|
|
*/
|
|
|
|
hunk_no = handle_path(sha1, (struct rerere_io *)&io, marker_size);
|
|
|
|
strbuf_release(&io.input);
|
|
|
|
if (io.io.output)
|
|
|
|
fclose(io.io.output);
|
|
|
|
return hunk_no;
|
2008-07-09 14:58:57 +02:00
|
|
|
}
|
2009-12-26 00:51:32 +01:00
|
|
|
|
|
|
|
static int rerere_forget_one_path(const char *path, struct string_list *rr)
|
|
|
|
{
|
|
|
|
const char *filename;
|
2015-07-05 02:38:34 +02:00
|
|
|
struct rerere_id *id;
|
2009-12-26 00:51:32 +01:00
|
|
|
unsigned char sha1[20];
|
|
|
|
int ret;
|
2015-06-30 22:03:36 +02:00
|
|
|
struct string_list_item *item;
|
2009-12-26 00:51:32 +01:00
|
|
|
|
2015-07-01 07:42:34 +02:00
|
|
|
/*
|
|
|
|
* Recreate the original conflict from the stages in the
|
|
|
|
* index and compute the conflict ID
|
|
|
|
*/
|
2009-12-26 00:51:32 +01:00
|
|
|
ret = handle_cache(path, sha1, NULL);
|
|
|
|
if (ret < 1)
|
|
|
|
return error("Could not parse conflict hunks in '%s'", path);
|
2015-07-01 07:42:34 +02:00
|
|
|
|
|
|
|
/* Nuke the recorded resolution for the conflict */
|
2015-07-05 02:38:34 +02:00
|
|
|
id = new_rerere_id(sha1);
|
2016-03-28 23:48:13 +02:00
|
|
|
|
|
|
|
for (id->variant = 0;
|
|
|
|
id->variant < id->collection->status_nr;
|
|
|
|
id->variant++) {
|
|
|
|
mmfile_t cur = { NULL, 0 };
|
|
|
|
mmbuffer_t result = {NULL, 0};
|
|
|
|
int cleanly_resolved;
|
|
|
|
|
|
|
|
if (!has_rerere_resolution(id))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
handle_cache(path, sha1, rerere_path(id, "thisimage"));
|
|
|
|
if (read_mmfile(&cur, rerere_path(id, "thisimage"))) {
|
|
|
|
free(cur.ptr);
|
2016-05-12 01:19:17 +02:00
|
|
|
error("Failed to update conflicted state in '%s'", path);
|
|
|
|
goto fail_exit;
|
2016-03-28 23:48:13 +02:00
|
|
|
}
|
|
|
|
cleanly_resolved = !try_merge(id, path, &cur, &result);
|
|
|
|
free(result.ptr);
|
|
|
|
free(cur.ptr);
|
|
|
|
if (cleanly_resolved)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2016-05-12 01:19:17 +02:00
|
|
|
if (id->collection->status_nr <= id->variant) {
|
|
|
|
error("no remembered resolution for '%s'", path);
|
|
|
|
goto fail_exit;
|
|
|
|
}
|
2016-03-28 23:48:13 +02:00
|
|
|
|
2015-07-06 23:45:55 +02:00
|
|
|
filename = rerere_path(id, "postimage");
|
2016-05-12 01:19:17 +02:00
|
|
|
if (unlink(filename)) {
|
|
|
|
if (errno == ENOENT)
|
|
|
|
error("no remembered resolution for %s", path);
|
|
|
|
else
|
2016-05-23 23:54:37 +02:00
|
|
|
error_errno("cannot unlink %s", filename);
|
2016-05-12 01:19:17 +02:00
|
|
|
goto fail_exit;
|
2016-05-19 21:51:22 +02:00
|
|
|
}
|
2009-12-26 00:51:32 +01:00
|
|
|
|
2015-07-01 07:42:34 +02:00
|
|
|
/*
|
|
|
|
* Update the preimage so that the user can resolve the
|
|
|
|
* conflict in the working tree, run us again to record
|
|
|
|
* the postimage.
|
|
|
|
*/
|
2015-07-06 23:45:55 +02:00
|
|
|
handle_cache(path, sha1, rerere_path(id, "preimage"));
|
2009-12-26 00:51:32 +01:00
|
|
|
fprintf(stderr, "Updated preimage for '%s'\n", path);
|
|
|
|
|
2015-07-01 07:42:34 +02:00
|
|
|
/*
|
|
|
|
* And remember that we can record resolution for this
|
|
|
|
* conflict when the user is done.
|
|
|
|
*/
|
2015-06-30 22:03:36 +02:00
|
|
|
item = string_list_insert(rr, path);
|
2015-07-05 02:38:34 +02:00
|
|
|
free_rerere_id(item);
|
2015-07-06 23:45:55 +02:00
|
|
|
item->util = id;
|
2009-12-26 00:51:32 +01:00
|
|
|
fprintf(stderr, "Forgot resolution for %s\n", path);
|
|
|
|
return 0;
|
2016-05-12 01:19:17 +02:00
|
|
|
|
|
|
|
fail_exit:
|
|
|
|
free(id);
|
|
|
|
return -1;
|
2009-12-26 00:51:32 +01:00
|
|
|
}
|
|
|
|
|
2013-07-14 10:35:40 +02:00
|
|
|
int rerere_forget(struct pathspec *pathspec)
|
2009-12-26 00:51:32 +01:00
|
|
|
{
|
|
|
|
int i, fd;
|
2010-07-04 21:46:19 +02:00
|
|
|
struct string_list conflict = STRING_LIST_INIT_DUP;
|
|
|
|
struct string_list merge_rr = STRING_LIST_INIT_DUP;
|
2009-12-26 00:51:32 +01:00
|
|
|
|
|
|
|
if (read_cache() < 0)
|
|
|
|
return error("Could not read index");
|
|
|
|
|
2010-01-20 23:44:31 +01:00
|
|
|
fd = setup_rerere(&merge_rr, RERERE_NOAUTOUPDATE);
|
2015-05-14 21:20:52 +02:00
|
|
|
if (fd < 0)
|
|
|
|
return 0;
|
2009-12-26 00:51:32 +01:00
|
|
|
|
2015-07-01 07:42:34 +02:00
|
|
|
/*
|
|
|
|
* The paths may have been resolved (incorrectly);
|
|
|
|
* recover the original conflicted state and then
|
|
|
|
* find the conflicted paths.
|
|
|
|
*/
|
2009-12-26 00:51:32 +01:00
|
|
|
unmerge_cache(pathspec);
|
|
|
|
find_conflict(&conflict);
|
|
|
|
for (i = 0; i < conflict.nr; i++) {
|
|
|
|
struct string_list_item *it = &conflict.items[i];
|
2014-01-24 14:40:30 +01:00
|
|
|
if (!match_pathspec(pathspec, it->string,
|
2014-01-24 14:40:33 +01:00
|
|
|
strlen(it->string), 0, NULL, 0))
|
2009-12-26 00:51:32 +01:00
|
|
|
continue;
|
|
|
|
rerere_forget_one_path(it->string, &merge_rr);
|
|
|
|
}
|
|
|
|
return write_rr(&merge_rr, fd);
|
|
|
|
}
|
2011-05-08 21:55:34 +02:00
|
|
|
|
2015-07-01 07:43:37 +02:00
|
|
|
/*
|
|
|
|
* Garbage collection support
|
|
|
|
*/
|
2015-07-05 02:38:34 +02:00
|
|
|
|
2016-03-08 21:11:00 +01:00
|
|
|
static time_t rerere_created_at(struct rerere_id *id)
|
2011-05-08 21:55:34 +02:00
|
|
|
{
|
|
|
|
struct stat st;
|
2015-07-05 02:38:34 +02:00
|
|
|
|
2015-07-06 23:45:55 +02:00
|
|
|
return stat(rerere_path(id, "preimage"), &st) ? (time_t) 0 : st.st_mtime;
|
2011-05-08 21:55:34 +02:00
|
|
|
}
|
|
|
|
|
2016-03-08 21:11:00 +01:00
|
|
|
static time_t rerere_last_used_at(struct rerere_id *id)
|
2011-05-08 21:55:34 +02:00
|
|
|
{
|
|
|
|
struct stat st;
|
2015-07-05 02:38:34 +02:00
|
|
|
|
2015-07-06 23:45:55 +02:00
|
|
|
return stat(rerere_path(id, "postimage"), &st) ? (time_t) 0 : st.st_mtime;
|
2011-05-08 21:55:34 +02:00
|
|
|
}
|
|
|
|
|
2015-07-01 07:43:37 +02:00
|
|
|
/*
|
|
|
|
* Remove the recorded resolution for a given conflict ID
|
|
|
|
*/
|
2015-07-05 02:38:34 +02:00
|
|
|
static void unlink_rr_item(struct rerere_id *id)
|
2011-05-08 21:55:34 +02:00
|
|
|
{
|
2016-03-08 21:11:00 +01:00
|
|
|
unlink_or_warn(rerere_path(id, "thisimage"));
|
|
|
|
remove_variant(id);
|
|
|
|
id->collection->status[id->variant] = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void prune_one(struct rerere_id *id, time_t now,
|
|
|
|
int cutoff_resolve, int cutoff_noresolve)
|
|
|
|
{
|
|
|
|
time_t then;
|
|
|
|
int cutoff;
|
|
|
|
|
|
|
|
then = rerere_last_used_at(id);
|
|
|
|
if (then)
|
|
|
|
cutoff = cutoff_resolve;
|
|
|
|
else {
|
|
|
|
then = rerere_created_at(id);
|
|
|
|
if (!then)
|
|
|
|
return;
|
|
|
|
cutoff = cutoff_noresolve;
|
|
|
|
}
|
|
|
|
if (then < now - cutoff * 86400)
|
|
|
|
unlink_rr_item(id);
|
2011-05-08 21:55:34 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void rerere_gc(struct string_list *rr)
|
|
|
|
{
|
|
|
|
struct string_list to_remove = STRING_LIST_INIT_DUP;
|
|
|
|
DIR *dir;
|
|
|
|
struct dirent *e;
|
2016-03-08 21:11:00 +01:00
|
|
|
int i;
|
|
|
|
time_t now = time(NULL);
|
2014-08-07 18:21:21 +02:00
|
|
|
int cutoff_noresolve = 15;
|
|
|
|
int cutoff_resolve = 60;
|
2011-05-08 21:55:34 +02:00
|
|
|
|
rerere: release lockfile in non-writing functions
There's a bug in builtin/am.c in which we take a lock on
MERGE_RR recursively. But rather than fix am.c, this patch
fixes the confusing interface from rerere.c that caused the
bug. Read on for the gory details.
The setup_rerere() function both reads the existing MERGE_RR
file, and takes MERGE_RR.lock. In the rerere() and
rerere_forget() functions, we end up in write_rr(), which
will then commit the lock file.
But for functions like rerere_clear() that do not write to
MERGE_RR, we expect the caller to have handled
setup_rerere(). That caller would then need to release the
lockfile, but it can't; the lock struct is local to
rerere.c.
For builtin/rerere.c, this is OK. We run a single rerere
operation and then exit immediately, which has the side
effect of rolling back the lockfile.
But in builtin/am.c, this is actively wrong. If we run "git
am -3 --skip", we call setup-rerere twice without releasing
the lock:
1. The "--skip" causes us to call am_rerere_clear(), which
calls setup_rerere(), but never drops the lock.
2. We then proceed to the next patch.
3. The "--3way" may cause us to call rerere() to handle
conflicts in that patch, but we are already holding the
lock. The lockfile code dies with:
BUG: prepare_tempfile_object called for active object
We could fix this by having rerere_clear() call
rollback_lock_file(). But it feels a bit odd for it to roll
back a lockfile that it did not itself take. So let's
simplify the interface further, and handle setup_rerere in
the function itself, taking away the question from the
caller over whether they need to do so.
We can give rerere_gc() the same treatment, as well (even
though it doesn't have any callers besides builtin/rerere.c
at this point). Note that these functions don't take flags
from their callers to pass along to setup_rerere; that's OK,
because the flags would not be meaningful for what they are
doing.
Both of those functions need to hold the lock because even
though they do not write to MERGE_RR, they are still writing
and should be protected from a simultaneous "rerere" run.
But rerere_remaining(), "rerere diff", and "rerere status"
are all read-only operations. They want to setup_rerere(),
but do not care about taking the lock in the first place.
Since our update of MERGE_RR is the usual atomic rename done
by commit_lock_file, they can just do a lockless read. For
that, we teach setup_rerere a READONLY flag to avoid the
lock.
As a bonus, this pushes builtin/rerere.c's setup_rerere call
closer to the functions that use it. Which means that "git
rerere totally-bogus-command" will no longer silently
exit(0) in a repository without rerere enabled.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-02 00:14:09 +02:00
|
|
|
if (setup_rerere(rr, 0) < 0)
|
|
|
|
return;
|
|
|
|
|
2014-08-07 18:21:21 +02:00
|
|
|
git_config_get_int("gc.rerereresolved", &cutoff_resolve);
|
|
|
|
git_config_get_int("gc.rerereunresolved", &cutoff_noresolve);
|
|
|
|
git_config(git_default_config, NULL);
|
2011-05-08 21:55:34 +02:00
|
|
|
dir = opendir(git_path("rr-cache"));
|
|
|
|
if (!dir)
|
|
|
|
die_errno("unable to open rr-cache directory");
|
2015-07-01 07:43:37 +02:00
|
|
|
/* Collect stale conflict IDs ... */
|
2011-05-08 21:55:34 +02:00
|
|
|
while ((e = readdir(dir))) {
|
2016-03-08 21:11:00 +01:00
|
|
|
struct rerere_dir *rr_dir;
|
|
|
|
struct rerere_id id;
|
|
|
|
int now_empty;
|
|
|
|
|
2011-05-08 21:55:34 +02:00
|
|
|
if (is_dot_or_dotdot(e->d_name))
|
|
|
|
continue;
|
2016-03-08 21:11:00 +01:00
|
|
|
rr_dir = find_rerere_dir(e->d_name);
|
|
|
|
if (!rr_dir)
|
|
|
|
continue; /* or should we remove e->d_name? */
|
|
|
|
|
|
|
|
now_empty = 1;
|
|
|
|
for (id.variant = 0, id.collection = rr_dir;
|
|
|
|
id.variant < id.collection->status_nr;
|
|
|
|
id.variant++) {
|
|
|
|
prune_one(&id, now, cutoff_resolve, cutoff_noresolve);
|
|
|
|
if (id.collection->status[id.variant])
|
|
|
|
now_empty = 0;
|
2011-05-08 21:55:34 +02:00
|
|
|
}
|
2016-03-08 21:11:00 +01:00
|
|
|
if (now_empty)
|
2011-05-08 21:55:34 +02:00
|
|
|
string_list_append(&to_remove, e->d_name);
|
|
|
|
}
|
2011-05-26 15:55:50 +02:00
|
|
|
closedir(dir);
|
2016-03-08 21:11:00 +01:00
|
|
|
|
|
|
|
/* ... and then remove the empty directories */
|
2011-05-08 21:55:34 +02:00
|
|
|
for (i = 0; i < to_remove.nr; i++)
|
2016-03-08 21:11:00 +01:00
|
|
|
rmdir(git_path("rr-cache/%s", to_remove.items[i].string));
|
2011-05-08 21:55:34 +02:00
|
|
|
string_list_clear(&to_remove, 0);
|
rerere: release lockfile in non-writing functions
There's a bug in builtin/am.c in which we take a lock on
MERGE_RR recursively. But rather than fix am.c, this patch
fixes the confusing interface from rerere.c that caused the
bug. Read on for the gory details.
The setup_rerere() function both reads the existing MERGE_RR
file, and takes MERGE_RR.lock. In the rerere() and
rerere_forget() functions, we end up in write_rr(), which
will then commit the lock file.
But for functions like rerere_clear() that do not write to
MERGE_RR, we expect the caller to have handled
setup_rerere(). That caller would then need to release the
lockfile, but it can't; the lock struct is local to
rerere.c.
For builtin/rerere.c, this is OK. We run a single rerere
operation and then exit immediately, which has the side
effect of rolling back the lockfile.
But in builtin/am.c, this is actively wrong. If we run "git
am -3 --skip", we call setup-rerere twice without releasing
the lock:
1. The "--skip" causes us to call am_rerere_clear(), which
calls setup_rerere(), but never drops the lock.
2. We then proceed to the next patch.
3. The "--3way" may cause us to call rerere() to handle
conflicts in that patch, but we are already holding the
lock. The lockfile code dies with:
BUG: prepare_tempfile_object called for active object
We could fix this by having rerere_clear() call
rollback_lock_file(). But it feels a bit odd for it to roll
back a lockfile that it did not itself take. So let's
simplify the interface further, and handle setup_rerere in
the function itself, taking away the question from the
caller over whether they need to do so.
We can give rerere_gc() the same treatment, as well (even
though it doesn't have any callers besides builtin/rerere.c
at this point). Note that these functions don't take flags
from their callers to pass along to setup_rerere; that's OK,
because the flags would not be meaningful for what they are
doing.
Both of those functions need to hold the lock because even
though they do not write to MERGE_RR, they are still writing
and should be protected from a simultaneous "rerere" run.
But rerere_remaining(), "rerere diff", and "rerere status"
are all read-only operations. They want to setup_rerere(),
but do not care about taking the lock in the first place.
Since our update of MERGE_RR is the usual atomic rename done
by commit_lock_file, they can just do a lockless read. For
that, we teach setup_rerere a READONLY flag to avoid the
lock.
As a bonus, this pushes builtin/rerere.c's setup_rerere call
closer to the functions that use it. Which means that "git
rerere totally-bogus-command" will no longer silently
exit(0) in a repository without rerere enabled.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-02 00:14:09 +02:00
|
|
|
rollback_lock_file(&write_lock);
|
2011-05-08 21:55:34 +02:00
|
|
|
}
|
|
|
|
|
2015-07-01 07:43:37 +02:00
|
|
|
/*
|
|
|
|
* During a conflict resolution, after "rerere" recorded the
|
|
|
|
* preimages, abandon them if the user did not resolve them or
|
|
|
|
* record their resolutions. And drop $GIT_DIR/MERGE_RR.
|
|
|
|
*
|
|
|
|
* NEEDSWORK: shouldn't we be calling this from "reset --hard"?
|
|
|
|
*/
|
2011-05-08 21:55:34 +02:00
|
|
|
void rerere_clear(struct string_list *merge_rr)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
rerere: release lockfile in non-writing functions
There's a bug in builtin/am.c in which we take a lock on
MERGE_RR recursively. But rather than fix am.c, this patch
fixes the confusing interface from rerere.c that caused the
bug. Read on for the gory details.
The setup_rerere() function both reads the existing MERGE_RR
file, and takes MERGE_RR.lock. In the rerere() and
rerere_forget() functions, we end up in write_rr(), which
will then commit the lock file.
But for functions like rerere_clear() that do not write to
MERGE_RR, we expect the caller to have handled
setup_rerere(). That caller would then need to release the
lockfile, but it can't; the lock struct is local to
rerere.c.
For builtin/rerere.c, this is OK. We run a single rerere
operation and then exit immediately, which has the side
effect of rolling back the lockfile.
But in builtin/am.c, this is actively wrong. If we run "git
am -3 --skip", we call setup-rerere twice without releasing
the lock:
1. The "--skip" causes us to call am_rerere_clear(), which
calls setup_rerere(), but never drops the lock.
2. We then proceed to the next patch.
3. The "--3way" may cause us to call rerere() to handle
conflicts in that patch, but we are already holding the
lock. The lockfile code dies with:
BUG: prepare_tempfile_object called for active object
We could fix this by having rerere_clear() call
rollback_lock_file(). But it feels a bit odd for it to roll
back a lockfile that it did not itself take. So let's
simplify the interface further, and handle setup_rerere in
the function itself, taking away the question from the
caller over whether they need to do so.
We can give rerere_gc() the same treatment, as well (even
though it doesn't have any callers besides builtin/rerere.c
at this point). Note that these functions don't take flags
from their callers to pass along to setup_rerere; that's OK,
because the flags would not be meaningful for what they are
doing.
Both of those functions need to hold the lock because even
though they do not write to MERGE_RR, they are still writing
and should be protected from a simultaneous "rerere" run.
But rerere_remaining(), "rerere diff", and "rerere status"
are all read-only operations. They want to setup_rerere(),
but do not care about taking the lock in the first place.
Since our update of MERGE_RR is the usual atomic rename done
by commit_lock_file, they can just do a lockless read. For
that, we teach setup_rerere a READONLY flag to avoid the
lock.
As a bonus, this pushes builtin/rerere.c's setup_rerere call
closer to the functions that use it. Which means that "git
rerere totally-bogus-command" will no longer silently
exit(0) in a repository without rerere enabled.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-02 00:14:09 +02:00
|
|
|
if (setup_rerere(merge_rr, 0) < 0)
|
|
|
|
return;
|
|
|
|
|
2011-05-08 21:55:34 +02:00
|
|
|
for (i = 0; i < merge_rr->nr; i++) {
|
2015-07-05 02:38:34 +02:00
|
|
|
struct rerere_id *id = merge_rr->items[i].util;
|
2016-03-08 21:11:00 +01:00
|
|
|
if (!has_rerere_resolution(id)) {
|
2015-07-06 23:45:55 +02:00
|
|
|
unlink_rr_item(id);
|
2016-03-08 21:11:00 +01:00
|
|
|
rmdir(rerere_path(id, NULL));
|
|
|
|
}
|
2011-05-08 21:55:34 +02:00
|
|
|
}
|
memoize common git-path "constant" files
One of the most common uses of git_path() is to pass a
constant, like git_path("MERGE_MSG"). This has two
drawbacks:
1. The return value is a static buffer, and the lifetime
is dependent on other calls to git_path, etc.
2. There's no compile-time checking of the pathname. This
is OK for a one-off (after all, we have to spell it
correctly at least once), but many of these constant
strings appear throughout the code.
This patch introduces a series of functions to "memoize"
these strings, which are essentially globals for the
lifetime of the program. We compute the value once, take
ownership of the buffer, and return the cached value for
subsequent calls. cache.h provides a helper macro for
defining these functions as one-liners, and defines a few
common ones for global use.
Using a macro is a little bit gross, but it does nicely
document the purpose of the functions. If we need to touch
them all later (e.g., because we learned how to change the
git_dir variable at runtime, and need to invalidate all of
the stored values), it will be much easier to have the
complete list.
Note that the shared-global functions have separate, manual
declarations. We could do something clever with the macros
(e.g., expand it to a declaration in some places, and a
declaration _and_ a definition in path.c). But there aren't
that many, and it's probably better to stay away from
too-magical macros.
Likewise, if we abandon the C preprocessor in favor of
generating these with a script, we could get much fancier.
E.g., normalizing "FOO/BAR-BAZ" into "git_path_foo_bar_baz".
But the small amount of saved typing is probably not worth
the resulting confusion to readers who want to grep for the
function's definition.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-10 11:38:57 +02:00
|
|
|
unlink_or_warn(git_path_merge_rr());
|
rerere: release lockfile in non-writing functions
There's a bug in builtin/am.c in which we take a lock on
MERGE_RR recursively. But rather than fix am.c, this patch
fixes the confusing interface from rerere.c that caused the
bug. Read on for the gory details.
The setup_rerere() function both reads the existing MERGE_RR
file, and takes MERGE_RR.lock. In the rerere() and
rerere_forget() functions, we end up in write_rr(), which
will then commit the lock file.
But for functions like rerere_clear() that do not write to
MERGE_RR, we expect the caller to have handled
setup_rerere(). That caller would then need to release the
lockfile, but it can't; the lock struct is local to
rerere.c.
For builtin/rerere.c, this is OK. We run a single rerere
operation and then exit immediately, which has the side
effect of rolling back the lockfile.
But in builtin/am.c, this is actively wrong. If we run "git
am -3 --skip", we call setup-rerere twice without releasing
the lock:
1. The "--skip" causes us to call am_rerere_clear(), which
calls setup_rerere(), but never drops the lock.
2. We then proceed to the next patch.
3. The "--3way" may cause us to call rerere() to handle
conflicts in that patch, but we are already holding the
lock. The lockfile code dies with:
BUG: prepare_tempfile_object called for active object
We could fix this by having rerere_clear() call
rollback_lock_file(). But it feels a bit odd for it to roll
back a lockfile that it did not itself take. So let's
simplify the interface further, and handle setup_rerere in
the function itself, taking away the question from the
caller over whether they need to do so.
We can give rerere_gc() the same treatment, as well (even
though it doesn't have any callers besides builtin/rerere.c
at this point). Note that these functions don't take flags
from their callers to pass along to setup_rerere; that's OK,
because the flags would not be meaningful for what they are
doing.
Both of those functions need to hold the lock because even
though they do not write to MERGE_RR, they are still writing
and should be protected from a simultaneous "rerere" run.
But rerere_remaining(), "rerere diff", and "rerere status"
are all read-only operations. They want to setup_rerere(),
but do not care about taking the lock in the first place.
Since our update of MERGE_RR is the usual atomic rename done
by commit_lock_file, they can just do a lockless read. For
that, we teach setup_rerere a READONLY flag to avoid the
lock.
As a bonus, this pushes builtin/rerere.c's setup_rerere call
closer to the functions that use it. Which means that "git
rerere totally-bogus-command" will no longer silently
exit(0) in a repository without rerere enabled.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-02 00:14:09 +02:00
|
|
|
rollback_lock_file(&write_lock);
|
2011-05-08 21:55:34 +02:00
|
|
|
}
|