Tarantool development patches archive
 help / color / mirror / Atom feed
From: Konstantin Osipov <kostja@tarantool.org>
To: "v.shpilevoy@tarantool.org" <v.shpilevoy@tarantool.org>
Cc: tarantool-patches@freelists.org, vdavydov.dev@gmail.com
Subject: Re: [tarantool-patches] [PATCH 1/3] netbox: allow to create a netbox connection from existing socket
Date: Tue, 27 Mar 2018 00:55:49 +0300	[thread overview]
Message-ID: <20180326215549.GA5011@atlas> (raw)
In-Reply-To: <9C4A63BA-42EB-4EB6-A19C-39102FED2F0F@tarantool.org>

* v.shpilevoy@tarantool.org <v.shpilevoy@tarantool.org> [18/03/22 22:54]:
> > 22 марта 2018 г., в 22:33, Konstantin Osipov <kostja@tarantool.org> написал(а):
> > 
> > * Vladislav Shpilevoy <v.shpilevoy@tarantool.org <mailto:v.shpilevoy@tarantool.org>> [18/03/22 22:18]:
> > 
> > The patch generally looks good to me, I can only nitpick on new
> > things. 
> > 
> > How about renaming do_connect to connect_impl and wrap_connect to
> > wrap_socket or simply wrap, or maybe bless or imbue?
> 
> Ok.
> 
> > 
> >> +        if existing_connection then
> >> +            connection = existing_connection
> >> +            existing_connection = nil
> > 
> > Why is it necessary to set it to nil?
> 
> On reconnect it will call protocol_sm again, and it must no get
> old existing_connection again.

This code is a mess and you make it even more messy.

What frustrates me most is that it's a trivial refactoring and
you prefer explaining why a fix is impossible rather than making
a good fix.

You *can* change the responsibilities of all of these functions so
that everything works nicely without extra branches and having to
pass existing_connection through several layers of calls.


> > 
> >> +            assert(greeting)
> >> +            assert(greeting.protocol ~= 'Lua console')
> > 
> > Please keep in mind that in Lua there are no asserts - it's a
> > runtime check.
> kk
> Ok.
> 
> >> +
> >> +local function wrap_socket(connection, greeting, url, opts)
> >> +    if connection == nil or type(greeting) ~= 'table' then
> >> +        error('Usage: netbox.wrap_socket(socket, greeting, [opts])')
> >> +    end
> >> +    if greeting.protocol == 'Lua console' then
> >> +        error('Can not wrap console socket')
> >> +    end
> >> +    opts = opts or {}
> >> +    if not opts.user and not opts.password then
> >> +        opts.user, opts.password = url.login, url.password
> >> +    end
> >> +    return do_connect(url.host, url.service, opts, connection, greeting)
> >> +end
> > 
> > Looks like you should make the next step, i.e. take the piece of
> > do_connect which establishes a connection and move it to
> > connect(), and then you can the remains of do_connect into wrap()
> > and call wrap() connect().
> > 
> > Did you try to do it? Had any trouble?
> 
> Connection establishment must be in connect inside
> create_transport, because even if an existing socket was
> wrapped, it still can have opts.reconnect_after option and the
> following scenario:
> socket = ...
> con = netbox.wrap_socket(socket, ..., {reconnect_after = ...})
> --
> -- use connection ...
> --
> 
> Suppose it is broken due to error. Then netbox will try to
> establish a new connection after 'reconnect_after' seconds.

You can leave reconnect in protocol_sm, but you can make sure
protocol_sm is never called without an existing connection.
You can make the first connect directly in create_transport(),
or better yet, as I first suggested, outside, removing the need
for connect_impl().


-- 
Konstantin Osipov, Moscow, Russia, +7 903 626 22 32
http://tarantool.org - www.twitter.com/kostja_osipov

  parent reply	other threads:[~2018-03-26 21:55 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-22 19:17 [PATCH 0/3] console: do not use netbox for console text connections Vladislav Shpilevoy
2018-03-22 19:17 ` [PATCH 1/3] netbox: allow to create a netbox connection from existing socket Vladislav Shpilevoy
2018-03-22 19:33   ` [tarantool-patches] " Konstantin Osipov
2018-03-22 19:50     ` v.shpilevoy
2018-03-22 19:57       ` [tarantool-patches] " v.shpilevoy
2018-03-23 10:01         ` [tarantool-patches] " v.shpilevoy
2018-03-26 21:55       ` Konstantin Osipov [this message]
2018-03-22 19:17 ` [PATCH 2/3] console: do not use netbox for console text connections Vladislav Shpilevoy
2018-03-22 19:36   ` [tarantool-patches] " Konstantin Osipov
2018-03-22 19:17 ` [PATCH 3/3] netbox: deprecate console support Vladislav Shpilevoy
2018-03-22 19:37   ` [tarantool-patches] " Konstantin Osipov

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=20180326215549.GA5011@atlas \
    --to=kostja@tarantool.org \
    --cc=tarantool-patches@freelists.org \
    --cc=v.shpilevoy@tarantool.org \
    --cc=vdavydov.dev@gmail.com \
    --subject='Re: [tarantool-patches] [PATCH 1/3] netbox: allow to create a netbox connection from existing socket' \
    /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