From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Tue, 8 May 2018 18:49:34 +0300 From: Konstantin Osipov Subject: Re: [tarantool-patches] [PATCH 4/8] netbox: extend codec with 'decode' methods Message-ID: <20180508154934.GD4867@atlas> References: <5196d19705a653bfbf600cf787e05f4f75df99f6.1523903144.git.v.shpilevoy@tarantool.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <5196d19705a653bfbf600cf787e05f4f75df99f6.1523903144.git.v.shpilevoy@tarantool.org> To: tarantool-patches@freelists.org Cc: vdavydov.dev@gmail.com List-ID: * Vladislav Shpilevoy [18/04/16 21:44]: > Netbox has a table 'method_codec' that is used to encode a > request by a method name. But a response is decoded out of codec. > It leads to > 1) decoding into Lua tables before decoding into tuples where > needed - it is double decoding and produces a lot of garbage; > 2) each method contains hacks like one_tuple(), or single tuple > check. > > These things can not be fixed with no real codec instead of > encoder only. > > Also global table with decoders is needed for #3107, where > a request could be sent async with no fiber blocking. An async > response when received already does not have a call context - it > has only method name. > > Needed for #3107 > --- > src/box/lua/net_box.lua | 116 +++++++++++++++++++++++++++++------------------- > test/box/net.box.result | 14 ++++++ > test/box/sql.result | 2 + > 3 files changed, 87 insertions(+), 45 deletions(-) > > diff --git a/src/box/lua/net_box.lua b/src/box/lua/net_box.lua > index 4ed2b375d..3868cdf1c 100644 > --- a/src/box/lua/net_box.lua > +++ b/src/box/lua/net_box.lua > @@ -50,7 +50,34 @@ local E_PROC_LUA = box.error.PROC_LUA > > -- utility tables > local is_final_state = {closed = 1, error = 1} > -local method_codec = { > + > +local function decode_nil(...) end > +local function decode_nothing(...) return ... end decode_nothing -> decode_nop > +local function decode_one_tuple(response) > + if response[1] then > + return box.tuple.new(response[1]) > + end > +end decode_one_tuple -> decode_tuple Why do you need the if () guard? > +local function decode_single_tuple(response) > + if response[2] then > + return nil, box.error.MORE_THAN_ONE_TUPLE > + end > + if response[1] then > + return box.tuple.new(response[1]) > + end > +end decode_single_tuple -> decode_get() > +local function decode_select(response) > + setmetatable(response, sequence_mt) > + for i, v in pairs(response) do > + response[i] = box.tuple.new(v) > + end > + return response > +end Why are you using pairs() rather than ipairs()? How does this avoid double decode as you claim in changeset comment? > +local function decode_count(response) > + return response[1] > +end > + > @@ -336,7 +385,10 @@ local function create_transport(host, port, user, password, callback, > -- Decode xrow.body[DATA] to Lua objects > body, body_end_check = decode(body_rpos) > assert(body_end == body_end_check, "invalid xrow length") > - return res -- the length of xrow.body > - elseif not err then > - setmetatable(res, sequence_mt) > - local postproc = method ~= 'eval' and method ~= 'call_17' > - if postproc then > - local tnew = box.tuple.new You removed this local variable from the decoder - it was here for a reason. Please put it back in your implementation of decoder. > diff --git a/test/box/net.box.result b/test/box/net.box.result > index cf7b27f0b..6a3713fc0 100644 > --- a/test/box/net.box.result > +++ b/test/box/net.box.result > @@ -416,9 +416,11 @@ cn.space.net_box_test_space:select({234}, { iterator = 'LT' }) > ... > cn.space.net_box_test_space:update({1}, { { '+', 2, 2 } }) > --- > +- null net.box calling convention should be the same as local box, if a tuple is not found, it should return nothing, not null. -- Konstantin Osipov, Moscow, Russia, +7 903 626 22 32 http://tarantool.io - www.twitter.com/kostja_osipov