Tarantool development patches archive
 help / color / mirror / Atom feed
From: Alexander Turenko <alexander.turenko@tarantool.org>
To: Konstantin Osipov <kostja.osipov@gmail.com>
Cc: tarantool-patches@freelists.org,
	Vladimir Davydov <vdavydov.dev@gmail.com>
Subject: Re: [tarantool-patches] [PATCH v3 6/7] Add merger for tuples streams (C part)
Date: Thu, 25 Apr 2019 16:32:27 +0300	[thread overview]
Message-ID: <20190425133227.mvell3bwcbhzqkyb@tkn_work_nb> (raw)
In-Reply-To: <20190425114343.GM29257@atlas>

On Thu, Apr 25, 2019 at 02:43:43PM +0300, Konstantin Osipov wrote:
> * Alexander Turenko <alexander.turenko@tarantool.org> [19/04/10 18:23]:
> > +enum { MERGER_SOURCE_REF_MAX = INT_MAX };
> 
> It's merge_source, not merger, source, no? Here and in the rest of
> the code.

merge source or merger's source. I found it neat to have one prefix
'merger_' for all related structures and functions.

But okay, I'll check it with Vladimir and if he agree will change those
names.

> 
> > +void
> > +merger_context_delete(struct merger_context *ctx)
> 
> Please decide it's either simply merger or merge_context.
> merger_context does not make any sense to me.

merger is a special kind of a source, the different thing.

> 
> what's in the context anyway? Can you have multiple contexts in a
> single merge? If no, why not simply merger? Why do you need both,
> a merger and a context?
> 
> > +/**
> > + * Holds immutable parameters of a merger.
> > + */
> > +struct merger_context {
> > +	struct key_def *key_def;
> > +	box_tuple_format_t *format;
> > +	/* Reference counter. */
> > +	int refs;
> > +};
> 
> Uhm.. Is it a comparison/sort context, not merger context then?

key_def is here for comparisons, yep, but format is for creating tuples
(with needed offsets). I'm like 'merger_' prefix, because it makes clear
that it is something that a merger need to perform its work.

The name is very common, yep. 'Comparison/sort context' would be more
specific, but not exact: it would arise a question: 'why a format is a
part of comparison context'?

I'll discuss this point with Vladimir too.

> Please explain why you need to ref it btw. I would simply pass it
> into Lua via ffi, to avoid any references, and make it clear that
> the life cycle of this object is bound to the merge window.

So you'll need to construct a format in Lua somehow. Moreover, you'll
need to create *right* format in Lua to make comparisons fast. Looks as
a way more complex approach.

Maybe I didn't get your idea.

  reply	other threads:[~2019-04-25 13:32 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-10 15:21 [PATCH v3 0/7] Merger Alexander Turenko
2019-04-10 15:21 ` [PATCH v3 1/7] Add luaL_iscallable with support of cdata metatype Alexander Turenko
2019-04-18 17:30   ` [tarantool-patches] " Konstantin Osipov
2019-04-30 12:45   ` Vladimir Davydov
2019-04-10 15:21 ` [PATCH v3 2/7] Add functions to ease using Lua iterators from C Alexander Turenko
2019-04-18 17:31   ` [tarantool-patches] " Konstantin Osipov
2019-04-30 12:46   ` Vladimir Davydov
2019-04-10 15:21 ` [PATCH v3 3/7] lua: optimize creation of a tuple from a tuple Alexander Turenko
2019-04-18 17:32   ` [tarantool-patches] " Konstantin Osipov
2019-04-30 12:50   ` Vladimir Davydov
2019-04-30 15:07     ` Alexander Turenko
2019-04-10 15:21 ` [PATCH v3 4/7] lua: add non-recursive msgpack decoding functions Alexander Turenko
2019-04-18 17:35   ` [tarantool-patches] " Konstantin Osipov
2019-04-18 18:30     ` Alexander Turenko
2019-04-18 18:33       ` Konstantin Osipov
2019-04-18 18:44         ` Alexander Turenko
2019-04-30 13:03   ` Vladimir Davydov
2019-04-30 18:38     ` Alexander Turenko
2019-04-10 15:21 ` [PATCH v3 5/7] net.box: add skip_header option to use with buffer Alexander Turenko
2019-04-18 17:37   ` [tarantool-patches] " Konstantin Osipov
2019-04-18 18:39     ` Alexander Turenko
2019-04-30 13:16   ` Vladimir Davydov
2019-04-30 18:39     ` Alexander Turenko
2019-04-10 15:21 ` [PATCH v3 6/7] Add merger for tuples streams (C part) Alexander Turenko
2019-04-25 11:43   ` [tarantool-patches] " Konstantin Osipov
2019-04-25 13:32     ` Alexander Turenko [this message]
2019-04-25 13:45       ` Konstantin Osipov
2019-04-25 15:32         ` [tarantool-patches] " Alexander Turenko
2019-04-25 16:42           ` Konstantin Osipov
2019-04-30 15:34   ` Vladimir Davydov
2019-05-07 22:14     ` Alexander Turenko
2019-04-10 15:21 ` [PATCH v3 7/7] Add merger for tuple streams (Lua part) Alexander Turenko
2019-04-25 11:46   ` [tarantool-patches] " Konstantin Osipov
2019-04-25 12:53     ` Alexander Turenko
2019-04-25 13:30       ` Konstantin Osipov
2019-04-30 17:37   ` Vladimir Davydov
2019-04-30 21:09     ` [tarantool-patches] " Konstantin Osipov
2019-05-02  9:48       ` Vladimir Davydov
2019-05-07 22:14     ` Alexander Turenko

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=20190425133227.mvell3bwcbhzqkyb@tkn_work_nb \
    --to=alexander.turenko@tarantool.org \
    --cc=kostja.osipov@gmail.com \
    --cc=tarantool-patches@freelists.org \
    --cc=vdavydov.dev@gmail.com \
    --subject='Re: [tarantool-patches] [PATCH v3 6/7] Add merger for tuples streams (C 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