git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ronnie Sahlberg <sahlberg@google.com>
To: git@vger.kernel.org
Cc: mhagger@alum.mit.edu, Ronnie Sahlberg <sahlberg@google.com>
Subject: [PATCH 20/31] refs.c: check for lock conflicts already in _update
Date: Wed, 14 May 2014 15:13:19 -0700	[thread overview]
Message-ID: <1400105610-21194-21-git-send-email-sahlberg@google.com> (raw)
In-Reply-To: <1400105610-21194-1-git-send-email-sahlberg@google.com>

Check for lock conflicts in _update and flag the transaction as errored
instead of waiting until _commit for doing these checks. If there was a lock
failure we check if this was due to a previous update in the same transaction
or not. This so that we can preserve the current error messages on lock failure
and log "Multiple updates for ref '%s' not allowed." when we have multiple
updates in the same transaction and "Cannot lock the ref '%s'." for any other
reason that the lock failed.

This also means that we no longer need to sort all the refs and check for
duplicates during _commit so all that code can be removed too.

Signed-off-by: Ronnie Sahlberg <sahlberg@google.com>
---
 refs.c | 85 +++++++++++++++++++++++++++++++++---------------------------------
 1 file changed, 42 insertions(+), 43 deletions(-)

diff --git a/refs.c b/refs.c
index 93f01e8..76cab6e 100644
--- a/refs.c
+++ b/refs.c
@@ -3415,6 +3415,7 @@ int transaction_update_sha1(struct ref_transaction *transaction,
 			    struct strbuf *err)
 {
 	struct ref_update *update;
+	int i;
 
 	if (have_old && !old_sha1)
 		die("BUG: have_old is true but old_sha1 is NULL");
@@ -3438,19 +3439,49 @@ int transaction_update_sha1(struct ref_transaction *transaction,
 	 * we do not need to check against this ref for name collissions
 	 * during locking.
 	 */
-	if (update->flags & REF_ISPACKONLY)
+	if (update->flags & REF_ISPACKONLY) {
 		add_delname(transaction, update->refname);
-	else {
-		update->lock = lock_ref_sha1_basic(update->refname,
-						   (update->have_old ?
-						    update->old_sha1 :
-						    NULL),
-						   update->flags,
-						   &update->type,
-						   transaction->delnames,
-						   transaction->delnum);
+		return 0;
 	}
-	return 0;
+
+	update->lock = lock_ref_sha1_basic(update->refname,
+					   (update->have_old ?
+					    update->old_sha1 :
+					    NULL),
+					   update->flags,
+					   &update->type,
+					   transaction->delnames,
+					   transaction->delnum);
+	if (update->lock)
+		return 0;
+
+	/* If we could not lock the ref it means we either collided with a
+	   different command or that we tried to perform a second update to
+	   the same ref from within the same transaction.
+	*/
+	transaction->status = REF_TRANSACTION_ERROR;
+
+	/* -1 is the update we just added. Start at -2 and find the most recent
+	   previous update for this ref.
+	*/
+	for (i = transaction->nr - 2; i >= 0; i--) {
+		if (transaction->updates[i]->update_type != UPDATE_SHA1) {
+			continue;
+		}
+		if (!strcmp(transaction->updates[i]->refname,
+			    update->refname))
+			break;
+	}
+	if (err)
+		if (i >= 0) {
+			const char *str =
+				"Multiple updates for ref '%s' not allowed.";
+			strbuf_addf(err, str, update->refname);
+		} else {
+			const char *str = "Cannot lock the ref '%s'.";
+			strbuf_addf(err, str, update->refname);
+		}
+	return 1;
 }
 
 int transaction_create_sha1(struct ref_transaction *transaction,
@@ -3525,32 +3556,6 @@ int update_ref(const char *action, const char *refname,
 	return 0;
 }
 
-static int ref_update_compare(const void *r1, const void *r2)
-{
-	const struct ref_update * const *u1 = r1;
-	const struct ref_update * const *u2 = r2;
-	return strcmp((*u1)->refname, (*u2)->refname);
-}
-
-static int ref_update_reject_duplicates(struct ref_update **updates, int n,
-					struct strbuf *err)
-{
-	int i;
-	for (i = 1; i < n; i++) {
-		if (updates[i]->update_type != UPDATE_SHA1)
-			continue;
-		if (!strcmp(updates[i - 1]->refname, updates[i]->refname)) {
-			const char *str =
-				"Multiple updates for ref '%s' not allowed.";
-			if (err)
-				strbuf_addf(err, str, updates[i]->refname);
-
-			return 1;
-		}
-	}
-	return 0;
-}
-
 int transaction_commit(struct ref_transaction *transaction,
 			   struct strbuf *err)
 {
@@ -3569,12 +3574,6 @@ int transaction_commit(struct ref_transaction *transaction,
 		return 0;
 	}
 
-	/* Copy, sort, and reject duplicate refs */
-	qsort(updates, n, sizeof(*updates), ref_update_compare);
-	ret = ref_update_reject_duplicates(updates, n, err);
-	if (ret)
-		goto cleanup;
-
 	/* Acquire all ref locks while verifying old values */
 	for (i = 0; i < n; i++) {
 		struct ref_update *update = updates[i];
-- 
2.0.0.rc3.506.g3739a35

  parent reply	other threads:[~2014-05-14 22:14 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-05-14 22:12 [PATCH 00/31] Finish implementing ref and reflog transactions Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 01/31] refs.c make ref_transaction_create a wrapper to ref_transaction_update Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 02/31] refs.c: make ref_transaction_delete a wrapper for ref_transaction_update Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 03/31] refs.c: rename the transaction functions Ronnie Sahlberg
2014-05-16 21:15   ` Junio C Hamano
2014-05-19 23:11     ` Ronnie Sahlberg
2014-05-19 23:25       ` Junio C Hamano
2014-05-14 22:13 ` [PATCH 04/31] refs.c: add a new update_type field to ref_update Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 05/31] refs.c: add a function to append a reflog entry to a fd Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 06/31] refs.c: add a transaction function to append a reflog entry Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 07/31] refs.c: add a flag to allow reflog updates to truncate the log Ronnie Sahlberg
2014-05-16 21:20   ` Junio C Hamano
2014-05-19 23:27     ` Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 08/31] refs.c: only write reflog update if msg is non-NULL Ronnie Sahlberg
2014-05-16 21:24   ` Junio C Hamano
2014-05-19 22:55     ` Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 09/31] refs.c: allow multiple reflog updates during a single transaction Ronnie Sahlberg
2014-05-16 21:35   ` Junio C Hamano
2014-05-16 22:01     ` Eric Sunshine
2014-05-19 22:58       ` Ronnie Sahlberg
2014-05-19 23:06     ` Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 10/31] reflog.c: use a reflog transaction when writing during expire Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 11/31] refs.c: null terminate the string in copy_msg Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 12/31] refs.c: track the refnames we are deleting in the transaction structure Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 13/31] refs.c: update the list of deleted refs during _update instead of _commit Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 14/31] refs.c: return error instead of dying when locking fails during transaction Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 15/31] refs.c: lock the ref during _update instead of during _commit Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 16/31] refs.c: add an error argument to create/delete/update just like commit Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 17/31] refs.c: make _update_reflog take an error argument Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 18/31] refs.c: return immediately from _commit if the transaction has an error Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 19/31] tests: move tests for -z update/delete/verify to after the ref is created Ronnie Sahlberg
2014-05-14 22:13 ` Ronnie Sahlberg [this message]
2014-05-14 22:13 ` [PATCH 21/31] refs.c allow multiple updates of the same ref in a transaction Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 22/31] refs.c: release all remaining locks during transaction_free Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 23/31] reflog.c: use the existing transaction to also lock and update the ref Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 24/31] refs.c: make unlock_ref static Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 25/31] refs.c: make close_ref static Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 26/31] refs.c: make commit_ref static Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 27/31] refs.c: remove the function lock_any_ref_for_update Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 28/31] refs.c: make struct ref_lock private to refs.c Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 29/31] refs.c: allow passing raw git_committer_info as email to _update_reflog Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 30/31] refs.c: move ref_update and other definitions to earlier in the file Ronnie Sahlberg
2014-05-14 22:13 ` [PATCH 31/31] refs.c: use the transaction to manage the reflog in rename_refs Ronnie Sahlberg

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=1400105610-21194-21-git-send-email-sahlberg@google.com \
    --to=sahlberg@google.com \
    --cc=git@vger.kernel.org \
    --cc=mhagger@alum.mit.edu \
    /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).