From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp29.i.mail.ru (smtp29.i.mail.ru [94.100.177.89]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by dev.tarantool.org (Postfix) with ESMTPS id 2C6C1469719 for ; Sun, 11 Oct 2020 18:35:54 +0300 (MSK) References: <6643130f-6b57-7c15-fb79-4bdd4873f561@tarantool.org> <20201009151021.GB31117@tarantool.org> From: Vladislav Shpilevoy Message-ID: <570611d8-9433-0637-719c-1557c67d2b31@tarantool.org> Date: Sun, 11 Oct 2020 17:35:52 +0200 MIME-Version: 1.0 In-Reply-To: <20201009151021.GB31117@tarantool.org> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Subject: Re: [Tarantool-patches] [PATCH v3 0/2] vinyl: rework upsert operation List-Id: Tarantool development patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Nikita Pettik Cc: tarantool-patches@dev.tarantool.org >> Does this patchset change format of vinyl statements on disk? If yes, I guess >> it breaks backward compatibility. Did you check how does it behave when booted >> from files of a previous version? Does it crash? Does it report an error? How >> a user can upgrade his vinyl files to the new format? > > No, everything goes fine except the fact that upsert statements restored > from disk have ops encoded into single array and as a result being applied > may break associative property (like it was before this patch). I've tried > to add a test, but since squash operation modifies content of run/xlog files, > such test can be run only once. If it is vital, I guess it is possible to > copy data directory before each test run.. Yes, it is possible. At least it would be better than not having a test at all.