From: Jeff King <peff@peff.net>
To: Junio C Hamano <gitster@pobox.com>
Cc: git@vger.kernel.org, Christian Couder <chriscool@tuxfamily.org>,
Jakub Narebski <jnareb@gmail.com>,
Eric Sunshine <sunshine@sunshineco.com>
Subject: Re: [PATCH 08/17] provide a helper to free commit buffer
Date: Thu, 12 Jun 2014 18:05:37 -0400 [thread overview]
Message-ID: <20140612220537.GA14445@sigill.intra.peff.net> (raw)
In-Reply-To: <20140612200855.GB4468@sigill.intra.peff.net>
On Thu, Jun 12, 2014 at 04:08:55PM -0400, Jeff King wrote:
> So just putting in the safety check is probably the least-disruptive
> thing. It doesn't automatically adapt to a change in the underlying
> commit_buffer code, but it would at least notice it.
Here is that commit with the check you suggested, and a better
explanation in the commit message of what is going on.
I don't know if it's easier for you to just replace patch 8 with this
(there will be a minor textual conflict in the final commit then), or if
you want me to re-post the whole 17-patch series.
Arguably this could be broken into two commits (one to add the simple
"free" and one for this more complicated "detach" case), but that makes
it little more complicated for you to apply. We need some way of
emailing patches coupled with rebase instructions. :)
-- >8 --
Subject: provide a helper to free commit buffer
This converts two lines into one at each caller. But more
importantly, it abstracts the concept of freeing the buffer,
which will make it easier to change later.
Note that we also need to provide a "detach" mechanism for a
tricky case in index-pack. We are passed a buffer for the
object generated by processing the incoming pack.. If we are
not using --strict, we just calculate the sha1 on that
buffer and return, leaving the caller to free it. But if we
are using --strict, we actually attach that buffer to an
object, pass the object to the fsck functions, and then
detach the buffer from the object again (so that the caller
can free it as usual). In this case, we don't want to free
the buffer ourselves, but just make sure it is no longer
associated with the commit.
Note that we are making the assumption here that the
attach/detach process does not impact the buffer at all
(e.g., it is never reallocated or modified). That holds true
now, and we have no plans to change that. However, as we
abstract the commit_buffer code, this dependency becomes
less obvious. So when we detach, let's also make sure that
we get back the same buffer that we gave to the
commit_buffer code.
Signed-off-by: Jeff King <peff@peff.net>
---
builtin/fsck.c | 3 +--
builtin/index-pack.c | 3 ++-
builtin/log.c | 6 ++----
builtin/rev-list.c | 3 +--
commit.c | 13 +++++++++++++
commit.h | 11 +++++++++++
6 files changed, 30 insertions(+), 9 deletions(-)
diff --git a/builtin/fsck.c b/builtin/fsck.c
index fc150c8..8aadca1 100644
--- a/builtin/fsck.c
+++ b/builtin/fsck.c
@@ -310,8 +310,7 @@ static int fsck_obj(struct object *obj)
if (obj->type == OBJ_COMMIT) {
struct commit *commit = (struct commit *) obj;
- free(commit->buffer);
- commit->buffer = NULL;
+ free_commit_buffer(commit);
if (!commit->parents && show_root)
printf("root %s\n", sha1_to_hex(commit->object.sha1));
diff --git a/builtin/index-pack.c b/builtin/index-pack.c
index 18f57de..3f88b12 100644
--- a/builtin/index-pack.c
+++ b/builtin/index-pack.c
@@ -786,7 +786,8 @@ static void sha1_object(const void *data, struct object_entry *obj_entry,
}
if (obj->type == OBJ_COMMIT) {
struct commit *commit = (struct commit *) obj;
- commit->buffer = NULL;
+ if (detach_commit_buffer(commit) != data)
+ die("BUG: parse_object_buffer transmogrified our buffer");
}
obj->flags |= FLAG_CHECKED;
}
diff --git a/builtin/log.c b/builtin/log.c
index 39e8836..226f8f2 100644
--- a/builtin/log.c
+++ b/builtin/log.c
@@ -349,8 +349,7 @@ static int cmd_log_walk(struct rev_info *rev)
rev->max_count++;
if (!rev->reflog_info) {
/* we allow cycles in reflog ancestry */
- free(commit->buffer);
- commit->buffer = NULL;
+ free_commit_buffer(commit);
}
free_commit_list(commit->parents);
commit->parents = NULL;
@@ -1508,8 +1507,7 @@ int cmd_format_patch(int argc, const char **argv, const char *prefix)
reopen_stdout(rev.numbered_files ? NULL : commit, NULL, &rev, quiet))
die(_("Failed to create output files"));
shown = log_tree_commit(&rev, commit);
- free(commit->buffer);
- commit->buffer = NULL;
+ free_commit_buffer(commit);
/* We put one extra blank line between formatted
* patches and this flag is used by log-tree code
diff --git a/builtin/rev-list.c b/builtin/rev-list.c
index 9f92905..e012ebe 100644
--- a/builtin/rev-list.c
+++ b/builtin/rev-list.c
@@ -173,8 +173,7 @@ static void finish_commit(struct commit *commit, void *data)
free_commit_list(commit->parents);
commit->parents = NULL;
}
- free(commit->buffer);
- commit->buffer = NULL;
+ free_commit_buffer(commit);
}
static void finish_object(struct object *obj,
diff --git a/commit.c b/commit.c
index fbdc480..11a05c1 100644
--- a/commit.c
+++ b/commit.c
@@ -245,6 +245,19 @@ int unregister_shallow(const unsigned char *sha1)
return 0;
}
+void free_commit_buffer(struct commit *commit)
+{
+ free(commit->buffer);
+ commit->buffer = NULL;
+}
+
+const void *detach_commit_buffer(struct commit *commit)
+{
+ void *ret = commit->buffer;
+ commit->buffer = NULL;
+ return ret;
+}
+
int parse_commit_buffer(struct commit *item, const void *buffer, unsigned long size)
{
const char *tail = buffer;
diff --git a/commit.h b/commit.h
index 4df48cb..d72ed43 100644
--- a/commit.h
+++ b/commit.h
@@ -51,6 +51,17 @@ int parse_commit_buffer(struct commit *item, const void *buffer, unsigned long s
int parse_commit(struct commit *item);
void parse_commit_or_die(struct commit *item);
+/*
+ * Free any cached object buffer associated with the commit.
+ */
+void free_commit_buffer(struct commit *);
+
+/*
+ * Disassociate any cached object buffer from the commit, but do not free it.
+ * The buffer (or NULL, if none) is returned.
+ */
+const void *detach_commit_buffer(struct commit *);
+
/* Find beginning and length of commit subject. */
int find_commit_subject(const char *commit_buffer, const char **subject);
--
2.0.0.566.gfe3e6b2
next prev parent reply other threads:[~2014-06-12 22:05 UTC|newest]
Thread overview: 60+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-06-09 18:02 [PATCH 0/15] store length of commit->buffer Jeff King
2014-06-09 18:09 ` [PATCH 01/15] alloc: include any-object allocations in alloc_report Jeff King
2014-06-09 18:09 ` [PATCH 02/15] commit: push commit_index update into alloc_commit_node Jeff King
2014-06-10 21:31 ` Junio C Hamano
2014-06-09 18:10 ` [PATCH 03/15] do not create "struct commit" with xcalloc Jeff King
2014-06-10 21:35 ` Junio C Hamano
2014-06-10 21:49 ` Jeff King
2014-06-09 18:10 ` [PATCH 04/15] logmsg_reencode: return const buffer Jeff King
2014-06-10 1:36 ` Eric Sunshine
2014-06-09 18:10 ` [PATCH 05/15] sequencer: use logmsg_reencode in get_message Jeff King
2014-06-10 21:40 ` Junio C Hamano
2014-06-10 21:50 ` Jeff King
2014-06-09 18:11 ` [PATCH 06/15] provide a helper to free commit buffer Jeff King
2014-06-09 18:11 ` [PATCH 07/15] provide a helper to set the " Jeff King
2014-06-09 18:11 ` [PATCH 08/15] provide helpers to access " Jeff King
2014-06-10 8:00 ` Eric Sunshine
2014-06-09 18:12 ` [PATCH 09/15] use get_cached_commit_buffer where appropriate Jeff King
2014-06-09 18:12 ` [PATCH 10/15] use get_commit_buffer to avoid duplicate code Jeff King
2014-06-10 8:01 ` Eric Sunshine
2014-06-10 20:34 ` Jeff King
2014-06-09 18:13 ` [PATCH 11/15] convert logmsg_reencode to get_commit_buffer Jeff King
2014-06-09 18:13 ` [PATCH 12/15] use get_commit_buffer everywhere Jeff King
2014-06-09 22:40 ` Junio C Hamano
2014-06-10 0:02 ` Jeff King
2014-06-10 0:22 ` Jeff King
2014-06-10 16:06 ` Junio C Hamano
2014-06-10 21:27 ` Jeff King
2014-06-09 18:15 ` [PATCH 13/15] commit-slab: provide a static initializer Jeff King
2014-06-09 18:15 ` [PATCH 14/15] commit: convert commit->buffer to a slab Jeff King
2014-06-09 18:15 ` [PATCH 15/15] commit: record buffer length in cache Jeff King
2014-06-10 5:12 ` Christian Couder
2014-06-10 5:27 ` Jeff King
2014-06-10 20:33 ` Jeff King
2014-06-10 21:30 ` Christian Couder
2014-06-10 21:35 ` [PATCH v2 0/17] store length of commit->buffer Jeff King
2014-06-10 21:36 ` [PATCH 01/17] commit_tree: take a pointer/len pair rather than a const strbuf Jeff King
2014-06-10 21:38 ` [PATCH 02/17] replace dangerous uses of strbuf_attach Jeff King
2014-06-10 21:38 ` [PATCH 03/17] alloc: include any-object allocations in alloc_report Jeff King
2014-06-10 21:39 ` [PATCH 04/17] commit: push commit_index update into alloc_commit_node Jeff King
2014-06-10 21:39 ` [PATCH 05/17] do not create "struct commit" with xcalloc Jeff King
2014-06-10 21:39 ` [PATCH 06/17] logmsg_reencode: return const buffer Jeff King
2014-06-10 21:39 ` [PATCH 07/17] sequencer: use logmsg_reencode in get_message Jeff King
2014-06-10 21:40 ` [PATCH 08/17] provide a helper to free commit buffer Jeff King
2014-06-12 18:22 ` Junio C Hamano
2014-06-12 20:08 ` Jeff King
2014-06-12 22:05 ` Jeff King [this message]
2014-06-10 21:40 ` [PATCH 09/17] provide a helper to set the " Jeff King
2014-06-10 21:40 ` [PATCH 10/17] provide helpers to access " Jeff King
2014-06-10 21:40 ` [PATCH 11/17] use get_cached_commit_buffer where appropriate Jeff King
2014-06-10 21:41 ` [PATCH 12/17] use get_commit_buffer to avoid duplicate code Jeff King
2014-06-10 21:41 ` [PATCH 13/17] convert logmsg_reencode to get_commit_buffer Jeff King
2014-06-10 21:41 ` [PATCH 14/17] use get_commit_buffer everywhere Jeff King
2014-06-10 21:42 ` [PATCH 15/17] commit-slab: provide a static initializer Jeff King
2014-06-12 18:15 ` Junio C Hamano
2014-06-12 19:51 ` Jeff King
2014-06-10 21:43 ` [PATCH 16/17] commit: convert commit->buffer to a slab Jeff King
2014-06-10 21:44 ` [PATCH 17/17] commit: record buffer length in cache Jeff King
2014-06-10 21:46 ` [PATCH v2 0/17] store length of commit->buffer Jeff King
2014-06-12 17:22 ` Junio C Hamano
2014-06-13 6:32 ` [PATCH v2] reuse cached commit buffer when parsing signatures Jeff King
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20140612220537.GA14445@sigill.intra.peff.net \
--to=peff@peff.net \
--cc=chriscool@tuxfamily.org \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=jnareb@gmail.com \
--cc=sunshine@sunshineco.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).