From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp43.i.mail.ru (smtp43.i.mail.ru [94.100.177.103]) (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 B02204696C3 for ; Wed, 22 Apr 2020 11:11:12 +0300 (MSK) Date: Wed, 22 Apr 2020 11:11:11 +0300 From: Kirill Yukhin Message-ID: <20200422081111.gcuhjhrfpf2sgqbf@tarantool.org> References: <6e8d54dae772d09c170f26612a3a028b7a38b037.1587476678.git.sergeyb@tarantool.org> <1587499465.734344969@f324.i.mail.ru> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable In-Reply-To: <1587499465.734344969@f324.i.mail.ru> Subject: Re: [Tarantool-patches] [PATCH v4 2/10] gitlab-ci: enable static analysis with luacheck List-Id: Tarantool development patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Alexander Tikhonov Cc: tarantool-patches@dev.tarantool.org, v.shpilevoy@tarantool.org, o.piskunov@tarantool.org Hello, On 21 =D0=B0=D0=BF=D1=80 23:04, Alexander Tikhonov wrote: >=20 > Sergey, thank you for the patch, I have some questions on it. >=20 > Do we really want to have separate stage before the man stages with testi= ng=C2=A0? As I see you want to have the ability to avoid of testing if the = initial luacheck test fails, but does it mean that issues found by the luac= heck will always need changes in the code that always will change the rest = of the jobs, I mean is it strong dependency between this check and the othe= rs that we can=E2=80=99t do the rest of the testing=C2=A0w/o luacheck ? We = had the opinion that our testing should produce fails as fast as possible t= hat is why we decided to remove different stages, and we can see the early = fails=C2=A0as fast as possible, because currently all the testing start tog= ether from the very start, but in your fix will have more that 2+ minutes d= elay before it, while we had strong suggestion from Kirill Y. to make the t= esting fast as much as possible and try to make it parallel up to 2 minutes= =2E Right, for now we have overall time of testing up to 20 minutes, but in= general fastest jobs finishes in 1+ minutes and we can see the issues much= faster than waiting the whole testing process. Please use lines of sane length. -- Regards, Kirill Yukhin