> Why can't this check be done in parallel with other tasks? Because when we contribute code to Ignite we must follow code style.
> I am not against the code-style check at all. Yes, you are :) But it’s ok > I just want to get some freedom for dev branches only, if it is possible of > course. It’s not about freedom - you are free to run any checks you want on TC. It’s about project codebase consistency. No long time ago we don’t have TC bot or checkstyle plugin. That leads us to new and new tests failures and code style violations. You can take a look at the commits that introduced checks and how many files were touched. We just ignore our own rules without checks. https://github.com/apache/ignite/commit/2fbbb676386515ea881e4e61f08864d6bc93225a https://github.com/apache/ignite/commit/2a85925f1705fbad36b5421c0ca5cf9de9a29658 https://github.com/apache/ignite/commit/d63f4d3569dcb387394d367a7f00aaf35f1b288e > 7 июля 2020 г., в 12:52, Вячеслав Коптилин <slava.kopti...@gmail.com> > написал(а): > > Hi Maxim, > >> Why the auto-formatting procedure cannot be used for any PR you'd like to > run on TC? >> Even more, you can remove all the rules from checkstyle XML config and do > all you want in the PR branch > Yes, I can enable auto-formatting, which should be checked anyway, I can > edit pom.xml every time, etc. > > My question is the following: Why can't this check be done in parallel with > other tasks? > > Yep, I see the argument mentioned by Nikolay - TeamCity capacity (we should > re-run all tests even though we rename a variable), but I don't see that > our TC servers are overwhelmed for now. > Perhaps, I am wrong and this argument (TC capacity) is more significant > than it seems to me at first glance. > > Folks, please don't get me wrong. I am not against the code-style check at > all. I just want to get some freedom for dev branches only, if it is > possible of course. > > Thanks, > Slava. > > вт, 7 июл. 2020 г. в 12:21, Maxim Muzafarov <mmu...@apache.org>: > >> Slava, >> >> Why the auto-formatting procedure cannot be used for any PR you'd like >> to run on TC? Even more, you can remove all the rules from checkstyle >> XML config and do all you want in the PR branch. >> >> On Tue, 7 Jul 2020 at 12:17, Вячеслав Коптилин <slava.kopti...@gmail.com> >> wrote: >>> >>> Hi Nikolay, >>> >>>> Why do you need to Run All on TC resources for this task? >>> For instance, it may be a race that cannot be reproduced locally on my >> own >>> hardware. >>> (By the way, even though if I just want to start Cache1 suite, the >>> Code-Style check executes anyway). >>> >>>> If we don’t want to follow the code style or considered it as a «waste >> of >>> the time» we should change it. >>> This is absolutely not what I'm trying to say. I do not think, that code >>> style is useless. I just want to say that this check can be done in >>> parallel for dev branches, for example. >>> >>> Thanks, >>> S. >>> >>> вт, 7 июл. 2020 г. в 11:49, Nikolay Izhikov <nizhi...@apache.org>: >>> >>>>> Let's assume that I want to implement a dirty fix of a bug, check a >>>> reproducer from the user list, etc. >>>> >>>> Why do you need to Run All on TC resources for this task? >>>> >>>>> I do not want to waste my time fixing all code style violations. >>>> >>>> I assume that you have the Ignite project locally because you edit >> Ignite >>>> source code. >>>> If yes, then IntelliJ IDEA has an automatic code formatting and does >> all >>>> the work for you. >>>> >>>>> Does this use-case look reasonable? >>>> >>>> Yes. >>>> But, I still don’t see what is the issue here. >>>> >>>> If we don’t want to follow the code style or considered it as a «waste >> of >>>> the time» we should change it. >>>> As simple as that. >>>> >>>> But, we should force the code style as early as we can. >>>> I think we should fail maven local build on code style violations. >>>> >>>>> 7 июля 2020 г., в 11:28, Вячеслав Коптилин <slava.kopti...@gmail.com >>> >>>> написал(а): >>>>> >>>>> Nikolay, >>>>> >>>>> There is *NO *intention to ignore code style violations and do merge >> PRs >>>>> into the master branch without fixing them. >>>>> >>>>> Let's assume that I want to implement a dirty fix of a bug, check a >>>>> reproducer from the user list, etc. >>>>> And I do not have the intention to merge that into the master branch >> as >>>> is, >>>>> however, I do not want to waste my time fixing all code style >> violations. >>>>> Does this use-case look reasonable? >>>>> >>>>> Thanks, >>>>> Slava. >>>>> >>>>> вт, 7 июл. 2020 г. в 11:07, Nikolay Izhikov <nizhi...@apache.org>: >>>>> >>>>>> Slava. >>>>>> >>>>>> All contributors should follow our code style during their >> contribution. >>>>>> For now, we have an automatic PR check that is integrated to the >> GitHub >>>>>> interface. >>>>>> >>>>>> I don’t see any issue here. >>>>>> All open-source project that I know, uses the same approach. >>>>>> >>>>>> Anyway, If some of the experienced community members is interested >> in >>>> the >>>>>> particular contribution he or she can help a newcomer with the code >>>> style - >>>>>> GitHub interface has the edit button even for someone else’s PR’s >>>>>> >>>>>> >>>>>>> 7 июля 2020 г., в 11:01, Вячеслав Коптилин < >> slava.kopti...@gmail.com> >>>>>> написал(а): >>>>>>> >>>>>>> Hello Nikolay, >>>>>>> >>>>>>>> Because any code style violations should be fixed before the >> merge. >>>>>>>> Therefore after the fix, you must rerun TC. >>>>>>>> This means that running heavy suites when code style is violated >> is a >>>>>>> waste of the resources. >>>>>>> This makes sense, however, to be honest, I don't see that our team >> city >>>>>>> servers are really busy. >>>>>>> >>>>>>>> Why do you want to violate code style rules in your PR? >>>>>>> Please take a look at the original e-mail from Ilya: >>>>>>>> This means that I have completely lost an option to run tests >> against >>>>>>> pull >>>>>>>> requests by new contributors - they usually compile but will not >> pass >>>>>>>> Checkstyle. That's a blocker. >>>>>>> >>>>>>> This issue has also been discussed here: >>>>>>> >>>>>> >>>> >> http://apache-ignite-developers.2346864.n4.nabble.com/DISCUSSION-Separate-code-sanity-check-and-build-task-td47003.html >>>>>>> >>>>>>> Thanks, >>>>>>> Slava. >>>>>>> >>>>>>> >>>>>>> вт, 7 июл. 2020 г. в 10:47, Nikolay Izhikov <nizhi...@apache.org>: >>>>>>> >>>>>>>> All checks just force rules we agreed on. >>>>>>>> >>>>>>>>> Why this check is so important? Why do you think it is more >> important >>>>>>>> than all other tasks/tests? >>>>>>>> >>>>>>>> Because any code style violations should be fixed before the >> merge. >>>>>>>> Therefore after the fix, you must rerun TC. >>>>>>>> This means that running heavy suites when code style is violated >> is a >>>>>>>> waste of the resources. >>>>>>>> >>>>>>>> Why do you want to violate code style rules in your PR? >>>>>>>> >>>>>>>> I think instead of hiding style errors we should make our code >> style >>>>>>>> comfortable for everyone. >>>>>>>> Can you, please, write - what part of the code style hurts you? >>>>>>>> >>>>>>>> >>>>>>>>> 7 июля 2020 г., в 10:39, Вячеслав Коптилин < >> slava.kopti...@gmail.com >>>>> >>>>>>>> написал(а): >>>>>>>>> >>>>>>>>> Hello Maxim, >>>>>>>>> >>>>>>>>>> Why do you think that splitting the checkstyle build is better >>>> option >>>>>>>>> than fixing code style issues reporting by the checkstyle plugin? >>>>>>>>> Why do you think that Ilya talks that code style errors should >> not be >>>>>>>> fixed? >>>>>>>>> >>>>>>>>> It looks weird to me that we do not even start the tests if check >>>> style >>>>>>>>> plugin reports violations. >>>>>>>>> Why can't this check be done in parallel with the tests and >> reported >>>> by >>>>>>>>> mtcga bot? >>>>>>>>> Why this check is so important? Why do you think it is more >> important >>>>>>>> than >>>>>>>>> all other tasks/tests? >>>>>>>>> >>>>>>>>> Thanks, >>>>>>>>> Slava. >>>>>>>>> >>>>>>>>> пн, 6 июл. 2020 г. в 20:34, Maxim Muzafarov <mmu...@apache.org>: >>>>>>>>> >>>>>>>>>> Hello Ilya, >>>>>>>>>> >>>>>>>>>> Why do you think that splitting the checkstyle build is better >>>> option >>>>>>>>>> than fixing code style issues reporting by the checkstyle >> plugin? >>>>>>>>>> >>>>>>>>>> On Mon, 6 Jul 2020 at 19:43, Ilya Kasnacheev < >>>>>> ilya.kasnach...@gmail.com >>>>>>>>> >>>>>>>>>> wrote: >>>>>>>>>>> >>>>>>>>>>> Hello! >>>>>>>>>>> >>>>>>>>>>> I have just noticed today that Checkstyle will fail Apache >> Ignite >>>>>>>> build: >>>>>>>>>>> >>>>>>>>>> >>>>>>>> >>>>>> >>>> >> https://ci.ignite.apache.org/buildConfiguration/IgniteTests24Java8_BuildApacheIgnite/5443282?buildTab=log&focusLine=3&linesState=683.4289 >>>>>>>>>>> >>>>>>>>>>> This means that I have completely lost an option to run tests >>>> against >>>>>>>>>> pull >>>>>>>>>>> requests by new contributors - they usually compile but will >> not >>>> pass >>>>>>>>>>> Checkstyle. That's a blocker. >>>>>>>>>>> >>>>>>>>>>> Can we please split Checkstyle as a separate build which is >>>> triggered >>>>>>>>>> with >>>>>>>>>>> Run All? >>>>>>>>>>> I think we even have >>>>>>>>>>> >>>>>>>>>> >>>>>>>> >>>>>> >>>> >> https://ci.ignite.apache.org/buildConfiguration/IgniteTests24Java8_CheckCodeStyle?mode=builds#all-projects >>>>>>>>>>> >>>>>>>>>>> WDYT? >>>>>>>>>>> >>>>>>>>>>> Regards, >>>>>>>>>>> -- >>>>>>>>>>> Ilya Kasnacheev >>>>>>>>>> >>>>>>>> >>>>>>>> >>>>>> >>>>>> >>>> >>>> >>
