On Tue, 20 Jan 2015 12:51:08 +0100 Ancor Gonzalez Sosa <[email protected]> wrote:
> On 01/20/2015 12:24 PM, Ancor Gonzalez Sosa wrote: > > On 01/15/2015 05:53 PM, Ancor Gonzalez Sosa wrote: > >> This change introduced in our test:unit task has caused the tests > >> in yast-yast2 to break. > >> https://github.com/yast/yast-rake/commit/b00ca0e8c8208b05698675b0cb4752552de710c9 > >> > >> It's actually not a fault of the change in the task. The problem > >> is that the tests are fragile because they pollute the global > >> namespace in several ways. To be honest, I'd expect test from some > >> other repositories to also break. Specially those written when we > >> were still newbies with RSpec. > >> > >> I did some changes in order to fix it but found another problem. > >> Here is the patch, not intended to enhance or refactor the tests > >> in any way, just to fix the minimum needed to make them run again. > >> https://github.com/ancorgs/yast-yast2/commit/a5c09480081be09d894f2df3a8b36a32dc3c26c4 > >> > >> The weird thing is that I needed to introduce the line marked with > >> a FIXME, which means that closing the default SCR is not as secure > >> as all our tests assume. Looking at the yast2-core source code I'd > >> say that the line is actually needed and we need to update quite > >> some tests and the documentation. But it's the core, so of course > >> I could be wrong. > > > > That's the question I was trying to open in the phone call. We need > > to fix the broken tests in yast-yast2 and, for sure, in many other > > modules. How to proceed about the branches? > > > > 1) For maintenance branches like SLE12 and 13.2 > > > > Do we want to fix it? > > > > They work with the version of yast-rake present in SLE12 and 13.2, > > but we cannot run "rake test:unit" for those branches with > > yast-rake>0.1.8 (in Tumbleweed at the moment). I think as long as fix works with both version of yast-rake we should backport it to maintenance branches, so it will work without problems. For me it is bug, not feature of old versions :) > > > > 2) For master > > > > We have to fix it (not a question) but I think that fixing, for > > example, chrooting of SCR in every single module is wrong. We need > > to extract this functionality to the ruby bindings and then use > > them in the different modules. So in my opinion it's time to > > introduce RSpec helpers for SCR in the ruby bindings. > > Like this, that I will commit to the proper repositories/branches once > we answer the question above. > > https://github.com/ancorgs/yast-yast2/blob/29bfc826c0df463e77fbf0512c272167a5860412/library/general/test/SCRStub.rb I agree. It should be in ruby bindings, but as I commented, it is not so easy to do it properly generic. So maybe we start with small one and then extending it. I prefer to have it close for changes and open for extension principle here, as this should be really backward compatible, otherwise every change can cause breakage of all testsuite, which I would like to avoid. > > > > > 3) For both (or only for master if we decide not to fix maint. > > branches) > > > > Do we want to fix all the modules right now or only yast-yast2 and > > wait to see errors in the rest? Keep in mind that in order to > > experiment the errors you need to be using yast-rake>0.1.8 so the > > errors could persist unnoticed for quite some time unless we all > > update the dev tools. > > > > Cheers. As long as we can release and deliver yast to users and customers it is not so critical. So I think we can go with evolution way, not revolution which changes it in all modules. Maybe we just need to check it for new code...BTW what version use travis? It should use latest, not? Josef -- To unsubscribe, e-mail: [email protected] To contact the owner, e-mail: [email protected]
