I slightly increased the default stack size limit of the forked test jvm in the related groovy script. It now works with Java 11 too.
To summarize: The documentation is updated and setting up the EE9.1 TCK is quite easy. It basically boils down to: (0) Ensure to have Maven and Java (11/14). (1) Run on a Linux / Mac OS system. (2) Adjust the settings.xml according to the template (3) Run the setup script (only oncE) (4) Check that everything is available (5) Add the TCK install into a local git repository to reset easily without delete & extract it again and again (as a TCK run modifies the isntallation) (6) Run the TCK and have fun. Nevertheless, we need to fix the javamail (javax vs jakarta) thing first, but afterwards any new volunteer can happily start TCK work. Gruß Richard Am Dienstag, dem 26.04.2022 um 18:34 +0000 schrieb Zowalla, Richard: > Thanks & You are right! Running with JDK 14 works. > > JDK 17 has some issues with illegal access within the groovy code > used. > > With JDK 14 I am hitting the Javamail issue - so the setup procedere > basically works. That are good news. > > Gruß > Richard > > > Am Dienstag, dem 26.04.2022 um 19:48 +0200 schrieb Jean-Louis > Monteiro: > > Sorry for the late reply. > > I managed to reproduce your issue. It does not seem to be a regular > > stackoverflow (loop without exit condition). > > Debugging the code with some additional traces, it looks like we > > are > > loading jar after jar and the stack isn't big enough to handle it > > in > > Java > > 11. > > > > I moved to a newer JDK like 14 or 17 and it worked. > > -- > > Jean-Louis Monteiro > > http://twitter.com/jlouismonteiro > > http://www.tomitribe.com > > > > > > On Tue, Apr 26, 2022 at 7:42 PM Zowalla, Richard < > > richard.zowa...@hs-heilbronn.de> wrote: > > > > > It looks like the SQL is not the real issue here - although it is > > > a > > > bit > > > confusing because the select statements look ok to me. > > > > > > However, after same additional digging / fiddling arround, I am > > > at > > > a > > > point, there the JVM for the tests does not startup due to an > > > StackOverflowError [1]. > > > > > > I tried remote debugging via "-dj" but have no glue why I am > > > running to > > > this StackOverflowError during the "ant.java(...)" call in the > > > related > > > groovy command. I bet, that I miss a little detail - so happy to > > > any > > > additional input ;) > > > > > > Gruß > > > R > > > > > > > > > [1] https://gist.github.com/rzo1/0af90723dc0c6689a6f7d7bae9d2e8da > > > > > > > > > Am Dienstag, dem 26.04.2022 um 09:59 +0000 schrieb Zowalla, > > > Richard: > > > > Hi, > > > > > > > > I changed the script pointing to the released ee9.1 tck now and > > > > updated > > > > the README accordingly. If the instructions are followed > > > > carefully, > > > > it > > > > will lead to a fresh installation. > > > > > > > > In the process, I noticed, that Glassfish 6.0.0 isn't available > > > > anymore > > > > and is replaced by 6.2.5 - JL mentioned on Slack, that 6.2.5 > > > > should > > > > also work. > > > > > > > > I think, that we do not really need a public git repository. > > > > Setting > > > > it > > > > up with the script is really straight forward. However, putting > > > > the > > > > TCK > > > > installation into a local git repository makes totally sense in > > > > order > > > > to avoid delete / unpack cycles due to the modifications on the > > > > installation, if the TCK is run. > > > > > > > > I added some instructions in the README. > > > > > > > > Currently, I am struggling with > > > > > > > > [ERROR] java.sql.SQLSyntaxErrorException: Syntax error: > > > > Encountered > > > > "Simple_Select_Query" at line 1, column 1. > > > > [INFO] 1130 of 1307 SQL statements executed successfully > > > > > > > > when running > > > > > > > > rm -rf target/ && ./runtests -Dhttps.protocols=TLSv1.1,TLSv1.2 > > > > -- > > > > ee91 > > > > --env -nc -c -U -w tomee-plume > > > > com.sun.ts.tests.ejb30.lite.appexception.singleton.annotated.Cl > > > > ie > > > > nt > > > > > > > > as the initial setup of the derby database seem to fail - don't > > > > have > > > > much ideas but using the search engine of trust I found an > > > > entry, > > > > that > > > > this also happened on a tck.work build long time ago. > > > > > > > > I guess, that I am just missing a little thing to fix it for my > > > > setup. > > > > > > > > Currently, the TCK runs will fail anyway as we need to fix java > > > > mail > > > > first but getting the same exception / outcome as JL would be > > > > beneficial as it would prove, that the setup works and others > > > > can > > > > simply replicate - if this work, I will re-evaluate the > > > > instructions > > > > for debugging ;) > > > > > > > > > > > > Gruß & thanks > > > > Richard > > > > > > > > > > > > Am Montag, dem 25.04.2022 um 14:13 -0700 schrieb David Blevins: > > > > > Thanks for digging into this, Richard! > > > > > > > > > > Agree with your prior statements about updating the README, > > > > > branching > > > > > for TomEE 8 and making TomEE 9 the master branch in the > > > > > tomee- > > > > > tck > > > > > project. > > > > > > > > > > > > > > > > On Apr 24, 2022, at 11:14 PM, Zowalla, Richard < > > > > > > richard.zowa...@hs-heilbronn.de> wrote: > > > > > > > > > > > > Ok, I found it: > > > > > > > > > > > > - > > > > > > https://lists.apache.org/thread/sfbvz76c527wojym71csv72k8j2qrrg7 > > > > > > > > > > > > If we adjust the script to target the (released) ee9.1 tck > > > > > > (instead > > > > > > of > > > > > > the snapshots), it will certainly help. (This was the > > > > > > script > > > > > > I > > > > > > had > > > > > > in > > > > > > mint) > > > > > > > > > > Agree. Maybe it needs a new name as well so it's more clear > > > > > that > > > > > it > > > > > can be used to create a fresh setup. > > > > > > > > > > > But while searching the list archive, I also found this > > > > > > post > > > > > > on > > > > > > happy > > > > > > TCK work: > > > > > > > > > > > > - > > > > > > https://lists.apache.org/thread/vgthyd8d7rww2p398r9swx8y37mwxf9c > > > > > > > > > > > > Sadly the setup images from IntelliJ IDEA are gone (maybe > > > > > > you > > > > > > have > > > > > > them > > > > > > somewhere, David?). > > > > > > > > > > Unfortunately, I don't have those images anymore. We can > > > > > maybe > > > > > describe the menu steps instead of showing images. Getting > > > > > that on > > > > > the website in the developer area would be good. > > > > > > > > > > > Perhaps we should provide a similar jakartaeetck-9.x > > > > > > installation > > > > > > via a > > > > > > Git repository? > > > > > > > > > > > > Then we could adjust the "update" script and make it an > > > > > > "install" > > > > > > script. > > > > > > > > > > The trick here is licensing of the TCK source code, which is > > > > > EPL. Here's the source of truth on "is this license ok" > > > > > questions: > > > > > > > > > > - > > > > > https://www.apache.org/legal/resolved.html#weak-copyleft-licenses > > > > > > > > > > Effectively, we can use EPL in binary form. In source form > > > > > it's ok > > > > > in small amounts where it generally can't be avoided. An > > > > > entire > > > > > repository of EPL source code living under > > > > > https://github.com/apache/ > > > > > is likely not ok. > > > > > > > > > > It doesn't mean the idea is shot, but it does mean that repo > > > > > would > > > > > have to live outside the Apache org in github. > > > > > > > > > > > That would ease the necessary setup steps. Wdyt? > > > > > > > > > > Definitely agree it's a great idea. Perhaps we could create > > > > > an > > > > > org > > > > > of our own? > > > > > > > > > > > > > > > -David > > > > >
smime.p7s
Description: S/MIME cryptographic signature
