From: Vladimir Davydov <vdavydov.dev@gmail.com>
To: Kirill Shcherbatov <kshcherbatov@tarantool.org>
Cc: tarantool-patches@freelists.org, kostja@tarantool.org
Subject: Re: [PATCH v5 8/9] box: introduce offset slot cache in key_part
Date: Tue, 4 Dec 2018 18:51:51 +0300 [thread overview]
Message-ID: <20181204155151.mfkugyfwukvdkcrs@esperanza> (raw)
In-Reply-To: <20181203210418.bpjwulkewjnoyqoh@esperanza>
On Tue, Dec 04, 2018 at 12:04:18AM +0300, Vladimir Davydov wrote:
> On Mon, Nov 26, 2018 at 01:49:42PM +0300, Kirill Shcherbatov wrote:
> > diff --git a/src/box/alter.cc b/src/box/alter.cc
> > index 029da02..6291159 100644
> > --- a/src/box/alter.cc
> > +++ b/src/box/alter.cc
> > @@ -856,7 +856,10 @@ alter_space_do(struct txn *txn, struct alter_space *alter)
> > * Create a new (empty) space for the new definition.
> > * Sic: the triggers are not moved over yet.
> > */
> > - alter->new_space = space_new_xc(alter->space_def, &alter->key_list);
> > + alter->new_space =
> > + space_new_xc(alter->space_def, &alter->key_list,
> > + alter->old_space->format != NULL ?
> > + alter->old_space->format->epoch + 1 : 1);
>
> Passing the epoch around looks ugly. Let's introduce a global counter
> and bump it right in tuple_format_new(). If you worry about disk_format
> and mem_format having different epochs in vinyl, it's not really a
> problem: make epoch an optional argument to tuple_format_new(); if the
> caller passes an epoch, use it, otherwise generate a new one; in vinyl
> reuse mem_format->epoch for disk_format.
>
> > diff --git a/src/box/key_def.h b/src/box/key_def.h
> > index 7731e48..3e08eb4 100644
> > --- a/src/box/key_def.h
> > +++ b/src/box/key_def.h
> > @@ -95,6 +95,14 @@ struct key_part {
> > char *path;
> > /** The length of JSON path. */
> > uint32_t path_len;
> > + /**
> > + * Source format for offset_slot_cache hit validations.
> > + * Cache is expected to use "the format with the newest
> > + * epoch is most relevant" strategy.
> > + */
> > + struct tuple_format *format_cache;
>
> Why did you decide to store tuple_format in key_part instead of epoch,
> as you did originally? I liked that more.
Discussed verbally, agreed that:
- Storing a pointer to the format in key_part is potentially dangerous,
because pointers may coincide after reallocation, resulting in
invalid behavior.
- We should make the epoch counter global rather than passing it around
on space creation. That is, tuple_format_new() will bump a global
int64_t counter and assign the new value to the new format.
- In vinyl disk_format and mem_format must have different epochs,
because they have different field maps so no optional argument
argument is required for tuple_format_new() - it will always assign a
new epoch to the new format.
- 'Epoch' isn't a very good name. We should probably call it 'version'
or 'format_version' or 'tuple_format_version', depending on the
context.
next prev parent reply other threads:[~2018-12-04 15:51 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-26 10:49 [PATCH v5 0/9] box: indexes by JSON path Kirill Shcherbatov
2018-11-26 10:49 ` [PATCH v5 1/9] box: refactor json_path_parser class Kirill Shcherbatov
2018-11-26 12:53 ` [tarantool-patches] " Kirill Shcherbatov
2018-11-29 15:39 ` Vladimir Davydov
2018-11-26 10:49 ` [PATCH v5 2/9] lib: implement JSON tree class for json library Kirill Shcherbatov
2018-11-26 12:53 ` [tarantool-patches] " Kirill Shcherbatov
2018-11-29 17:38 ` Vladimir Davydov
2018-11-29 17:50 ` Vladimir Davydov
2018-12-04 15:22 ` Vladimir Davydov
2018-12-04 15:47 ` [tarantool-patches] " Kirill Shcherbatov
2018-12-04 17:54 ` Vladimir Davydov
2018-12-05 8:37 ` Kirill Shcherbatov
2018-12-05 9:07 ` Vladimir Davydov
2018-12-05 9:52 ` Vladimir Davydov
2018-12-06 7:56 ` Kirill Shcherbatov
2018-12-06 7:56 ` [tarantool-patches] Re: [PATCH v5 2/9] lib: make index_base support for json_lexer Kirill Shcherbatov
2018-11-26 10:49 ` [PATCH v5 3/9] box: manage format fields with JSON tree class Kirill Shcherbatov
2018-11-29 19:07 ` Vladimir Davydov
2018-12-04 15:47 ` [tarantool-patches] " Kirill Shcherbatov
2018-12-04 16:09 ` Vladimir Davydov
2018-12-04 16:32 ` Kirill Shcherbatov
2018-12-05 8:37 ` Kirill Shcherbatov
2018-12-06 7:56 ` Kirill Shcherbatov
2018-12-06 8:06 ` Vladimir Davydov
2018-11-26 10:49 ` [PATCH v5 4/9] lib: introduce json_path_cmp routine Kirill Shcherbatov
2018-11-30 10:46 ` Vladimir Davydov
2018-12-03 17:37 ` [tarantool-patches] " Konstantin Osipov
2018-12-03 18:48 ` Vladimir Davydov
2018-12-03 20:14 ` Konstantin Osipov
2018-12-06 7:56 ` [tarantool-patches] Re: [PATCH v5 4/9] lib: introduce json_path_cmp, json_path_validate Kirill Shcherbatov
2018-11-26 10:49 ` [tarantool-patches] [PATCH v5 5/9] box: introduce JSON indexes Kirill Shcherbatov
2018-11-30 21:28 ` Vladimir Davydov
2018-12-01 16:49 ` Vladimir Davydov
2018-11-26 10:49 ` [PATCH v5 6/9] box: introduce has_json_paths flag in templates Kirill Shcherbatov
2018-11-26 10:49 ` [PATCH v5 7/9] box: tune tuple_field_raw_by_path for indexed data Kirill Shcherbatov
2018-12-01 17:20 ` Vladimir Davydov
2018-11-26 10:49 ` [PATCH v5 8/9] box: introduce offset slot cache in key_part Kirill Shcherbatov
2018-12-03 21:04 ` Vladimir Davydov
2018-12-04 15:51 ` Vladimir Davydov [this message]
2018-11-26 10:49 ` [PATCH v5 9/9] box: specify indexes in user-friendly form Kirill Shcherbatov
2018-12-04 12:22 ` Vladimir Davydov
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=20181204155151.mfkugyfwukvdkcrs@esperanza \
--to=vdavydov.dev@gmail.com \
--cc=kostja@tarantool.org \
--cc=kshcherbatov@tarantool.org \
--cc=tarantool-patches@freelists.org \
--subject='Re: [PATCH v5 8/9] box: introduce offset slot cache in key_part' \
/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