Tarantool development patches archive
 help / color / mirror / Atom feed
From: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
To: sergepetrenko <sergepetrenko@tarantool.org>, georgy@tarantool.org
Cc: tarantool-patches@dev.tarantool.org
Subject: Re: [Tarantool-patches] [PATCH 4/5] vclock: ignore 0th component in comparisons.
Date: Sun, 22 Dec 2019 18:59:17 +0100	[thread overview]
Message-ID: <cf520b03-e153-a9e6-1e2a-ee3d76196da0@tarantool.org> (raw)
In-Reply-To: <ac5d46cee3386df55d8290ca6b3c5acecfbd8ca6.1576443171.git.sergepetrenko@tarantool.org>

Thanks for the patch!

See 2 comments below.

On 15/12/2019 21:58, sergepetrenko wrote:
> 0th vclock component will be used to count replica-local rows of an
> anonymous replica. These rows won't be replicated and different
> instances will have different values in vclock[0]. So ignore 0th
> component in comparisons.

1. Lets add 'Part of #3186'.

> ---
>  src/box/vclock.h    | 7 +++++++
>  test/unit/vclock.cc | 8 ++++----
>  2 files changed, 11 insertions(+), 4 deletions(-)
> 
> diff --git a/src/box/vclock.h b/src/box/vclock.h
> index b5eddcf8b..5dddbd173 100644
> --- a/src/box/vclock.h
> +++ b/src/box/vclock.h
> @@ -281,6 +281,13 @@ vclock_compare(const struct vclock *a, const struct vclock *b)
>  
>  	for (size_t replica_id = bit_iterator_next(&it); replica_id < VCLOCK_MAX;
>  	     replica_id = bit_iterator_next(&it)) {
> +		/*
> +		 * Ignore 0-th component in comparisons.
> +		 * It is empty for normal replicas and should
> +		 * be ignored for anonymous ones.
> +		 */
> +		if (replica_id == 0)
> +			continue;
>  
>  		int64_t lsn_a = vclock_get(a, replica_id);
>  		int64_t lsn_b = vclock_get(b, replica_id);

2. Consider this refactoring in order not to compare replica_id
with 0 on each iteration:

================================================================================

diff --git a/src/box/vclock.h b/src/box/vclock.h
index 5dddbd173..fc6aeb724 100644
--- a/src/box/vclock.h
+++ b/src/box/vclock.h
@@ -279,16 +279,15 @@ vclock_compare(const struct vclock *a, const struct vclock *b)
 	struct bit_iterator it;
 	bit_iterator_init(&it, &map, sizeof(map), true);
 
-	for (size_t replica_id = bit_iterator_next(&it); replica_id < VCLOCK_MAX;
-	     replica_id = bit_iterator_next(&it)) {
-		/*
-		 * Ignore 0-th component in comparisons.
-		 * It is empty for normal replicas and should
-		 * be ignored for anonymous ones.
-		 */
-		if (replica_id == 0)
-			continue;
-
+	size_t replica_id = bit_iterator_next(&it);
+	/*
+	 * Ignore 0-th component in comparisons. It is empty for
+	 * normal replicas and should be ignored for anonymous
+	 * ones.
+	 */
+	if (replica_id == 0)
+		replica_id = bit_iterator_next(&it);
+	for (; replica_id < VCLOCK_MAX; replica_id = bit_iterator_next(&it)) {
 		int64_t lsn_a = vclock_get(a, replica_id);
 		int64_t lsn_b = vclock_get(b, replica_id);
 		le = le && lsn_a <= lsn_b;

================================================================================

  reply	other threads:[~2019-12-22 17:59 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-15 20:56 [Tarantool-patches] [PATCH 0/5] introduce anonymous replicas sergepetrenko
2019-12-15 20:58 ` [Tarantool-patches] [PATCH 1/5] box: update comment describing join protocol sergepetrenko
2019-12-22 17:58   ` Vladislav Shpilevoy
2019-12-23 21:12     ` Sergey Petrenko
2019-12-15 20:58 ` [Tarantool-patches] [PATCH 2/5] replication: do not decode replicaset uuid when processing a subscribe sergepetrenko
2019-12-15 20:58 ` [Tarantool-patches] [PATCH 3/5] applier: split join processing into two stages sergepetrenko
2019-12-22 17:59   ` Vladislav Shpilevoy
2019-12-23 22:10     ` Sergey Petrenko
2019-12-24 15:50       ` Vladislav Shpilevoy
2019-12-15 20:58 ` [Tarantool-patches] [PATCH 4/5] vclock: ignore 0th component in comparisons sergepetrenko
2019-12-22 17:59   ` Vladislav Shpilevoy [this message]
2019-12-23 21:26     ` Sergey Petrenko
2019-12-23 22:58       ` Sergey Petrenko
2019-12-26  4:43   ` Konstantin Osipov
2019-12-26  5:02     ` Konstantin Osipov
2019-12-27 12:56       ` Sergey Petrenko
2019-12-27 13:31         ` Konstantin Osipov
2019-12-27 13:48           ` Sergey Petrenko
2019-12-27 14:40             ` Konstantin Osipov
2019-12-15 20:58 ` [Tarantool-patches] [PATCH 5/5] replication: introduce anonymous replica sergepetrenko
2019-12-16 13:28   ` Serge Petrenko
2019-12-20 12:06     ` Serge Petrenko
2019-12-22 17:58   ` Vladislav Shpilevoy
2019-12-25 12:40     ` Sergey Petrenko
2019-12-25 18:23       ` Vladislav Shpilevoy
2019-12-26 16:08         ` Sergey Petrenko
2019-12-15 21:00 ` [Tarantool-patches] [PATCH 0/5] introduce anonymous replicas Sergey Petrenko
2019-12-18  7:49 ` Georgy Kirichenko
2019-12-20 12:07   ` Serge Petrenko
2019-12-20 12:17     ` Serge Petrenko
2019-12-22 17:59 ` Vladislav Shpilevoy

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=cf520b03-e153-a9e6-1e2a-ee3d76196da0@tarantool.org \
    --to=v.shpilevoy@tarantool.org \
    --cc=georgy@tarantool.org \
    --cc=sergepetrenko@tarantool.org \
    --cc=tarantool-patches@dev.tarantool.org \
    --subject='Re: [Tarantool-patches] [PATCH 4/5] vclock: ignore 0th component in comparisons.' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox