From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Tue, 30 Apr 2019 12:48:17 +0300 From: Konstantin Osipov Subject: Re: [tarantool-patches] Re: [PATCH v4 3/3] box: introduce multikey indexes in memtx Message-ID: <20190430094817.GA5685@atlas> References: <20190429160615.5a2uedzpzj6rw76k@esperanza> <0ea77506-42b9-1e08-cda2-062d2948a0ff@tarantool.org> <20190430084353.zq2uxs23bxwxo26i@esperanza> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20190430084353.zq2uxs23bxwxo26i@esperanza> To: Vladimir Davydov Cc: Kirill Shcherbatov , tarantool-patches@freelists.org List-ID: * Vladimir Davydov [19/04/30 12:43]: > > > That would emphasize that it inserts a single multikey entry. > > > > I like "memtx_tree_index_replace_multikey_step". What do you think? > > Better _one or _do_replace IMO :) We typically use _one suffix for > splitting functions like that in the code; _step sounds like it is > about iterators while it is not. _one or replace_one_key > > >> +/** > > >> + * Check if child_field may be attached to parent_field, > > >> + * update the parent_field container type if required. > > >> + */ > > >> +static int > > >> +tuple_field_ensure_child_compatibility(struct tuple_field *parent, > > >> + struct tuple_field *child) > > > > > > Wouldn't tuple_format_check_field be a better name? > > I am not shure. > > This code not only check some cases, but also may change field type > > when it is possible. > > "enshure compatibility" -- make everything okey or die then split it into two functions? -- Konstantin Osipov, Moscow, Russia, +7 903 626 22 32