> On Dec 8, 2020, at 2:12 PM, Thiago Henrique Hupner <thi...@gmail.com> wrote: > > Hey David! > > Good to be remembered ;)
:) > Thanks for the link. > > I've been contributing to the Piranha server lately helping to pass the > Servlet TCK, so I guess I've got some experience with the TCK. > I really would like to help. Probably I'll need some help setting up the > full TCK environment because I only have the Servlet and Pages TCK in my > machine. There'll be many people who will need help with the setup, so feel free to be extra noisy and ask a lot of questions. All that will help improve the documentation so more people can join in the fun. Here's what we have so far: - https://github.com/apache/tomee-tck#getting-setup Meanwhile, you can get the test output and server logs from tck.work. For example here are the test failures for com.sun.ts.tests.servlet.spec.defaultcontextpath which has a table with "Package" and "Test": - https://tck.work/tomee/tests?path=com.sun.ts.tests.servlet.spec.defaultcontextpath&build=1607347669779&status=FAILED If you click the link under "Package" you get these server logs and Maven output: - https://tck.work/tomee/api/build/file/1607347669779/247 If you click the link under "Test" you get the TCK test output: - https://tck.work/tomee/api/testlog/247/1607347937486 With that you've got the name of the test, all forms of server output and the test output. It's enough to grab the test source and try and figure out on which line the test is failing and what expectation isn't met. People are always afraid of being too noisy and "bothering" others, but honestly hitting the list with even those findings is a great start. It's enough to start a conversation on what the fix might be and where it should be (in the TCK or in TomEE). There are 300 failing tests, so if someone took even a small peak at each one and posted some quick analysis of each one that'd be no small contribution. On that note, there is a feature to help deal with the volume of failures and focus on what might have the highest impact. Here's a report of all the exceptions thrown by the failing Servlet tests and the number of tests impacted: - https://tck.work/tomee/issues?path=com.sun.ts.tests.servlet&issue=&build=1607347669779&status=FAILED Indeed, actually debugging a test to figure out what the fix is definitely requires getting the TCK setup yourself. > > And sure it will help me give a better view of the web platform and how to > set up things for Piranha in the future too :) > > One question I have is with respect to issuing a pull request: I should > first open an issue at Jira then open a PR? This still confuses me a little. First is hit the list with a short "I'm looking to hack on a PR for X, any thoughts/tips before I start?" After that it doesn't really matter what comes next PR or Issue. We used to use the Issues to automatically generate the changelog each release. As long as we ensure there is a JIRA before we do the release, that's the final deadline. Unfortunately, that system decayed due to change in JIRA APIs over time and isn't being used currently and as a result there is no changelog each release. There's a whole other area/opportunity for contribution there. Sometimes I question the need for JIRA anymore, but that's a whole other story. :) -David > Em ter., 8 de dez. de 2020 às 18:42, David Blevins <david.blev...@gmail.com> > escreveu: > >> Hey Thiago! >> >> Thought I saw your name on today's JakartaOne Livestream. Here's a site >> that has the latest TomEE TCK results: >> >> - https://tck.work/tomee/projects >> >> We're at the stage where we're down to the last few hundred failures. >> This is where things get really detailed and is actually a great place to >> dive in and help. In the earlier phases getting tests to pass is often, >> "we need someone to create/integrate an implementation of X spec" which is >> definitely heavy work. >> >> At this stage the issues are more like, "the spec says this, we do that, >> the TCK seems to do something entirely different." In this phase it's a >> lot of reading test code, reading specifications to try and understand, >> having discussions here to get some consistent perspective then having >> conversations with other people on the Jakarta EE side. >> >> It's one large learning opportunity for all of us, myself included. If >> anyone is excited about potentially becoming an expert in Jakarta EE >> through helping with the TCK, this is the phase to dig in. >> >> If this interests you, I recommend keeping an eye out for any threads with >> the "[TCK]" prefix and asking questions to get more information and help >> with the research. If you're feeling really aggressive, you can even find >> a failing test, dig into the source code and see if you can understand what >> the test is asking for that we aren't doing then post some details on what >> you find (even if what you find is more questions -- that seems to be the >> typical result). >> >> >> -- >> David Blevins >> http://twitter.com/dblevins >> http://www.tomitribe.com >> >>> On Oct 30, 2020, at 6:56 AM, Thiago Henrique Hupner <thi...@gmail.com> >> wrote: >>> >>> Hi all, >>> >>> Is there any place where we can see the current results of the TCK? >>> >>> If not, it would be possible to use the Github Actions to run it? >>> >>> Thanks >>> >>> Thiago >> >>
