Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jul 31 2018 17:03

    scottkurz on 7-remove-SpringBoot-parent-pom

    Update README.md Add warning a… Update snapshot versions Merge pull request #1 from cher… and 2 more (compare)

  • Jul 26 2018 20:25

    scottkurz on first-test

    Add SpringBoot 1.5 app as first… (compare)

Scott Kurz
@scottkurz
In trying to review #220, I just found another issue with the precommit hook it seems ;(
Scott Kurz
@scottkurz
There's a mistake in the case ... newList != newlist
so the formatter wasn't checking all it should
gotta love a compiler
Scott Kurz
@scottkurz
Adam Wisniewski
@awisniew90
@chyt - prototyping what you had mentioned the other day about possibly having a derby booster. I'm picturing this as just a booster (i.e. "booster-derby") who's pom.xml has a dependency on apache derby at a certain version of our choosing. I imagine that if someone were to use this, they would also need to add our jdbc booster dependency to their project. We wouldn't have the derby booster implicitly pull in the jdbc booster right? Or maybe?
Adam Wisniewski
@awisniew90
When someone has a chance to review, I have a PR for using a user defined derby dependency here: OpenLiberty/boost#229 .
I also created a new issue under the JDBC epic for creating boosters for different database driver dependencies.
Scott Kurz
@scottkurz
So #229 you're making into a booster-derby dependency??
let's title them so it differentiates btw. a booster-derby vs. derby dependency
Adam Wisniewski
@awisniew90
oh. sorry. no #229 is just a straight derby dependency
I opened issue #228 for a booster-derby (and other database driver boosters)
Scott Kurz
@scottkurz
:+1:
Mike Andrasak
@uberskigeek
Some thoughts on overriding the server port at execution time. The spring boot plugin will override the server port with either of the following "mvn spring-boot:run -Dserver.port=8888" or "mvn spring-boot:run -Drun.jvmArguments='-Dserver.port=8888' . In either case if the parm is not specified on the next run then the server defaults to the original port. Liberty provides an environment variable JVM_ARGS and when exporting JVM_ARGS='-Dserver.port=8888' then execution mvn boost:run .... The server will use the port override. This will remain in effect until the environment variable is unset.
Scott Kurz
@scottkurz
So @uberskigeek it sounds like you're heading in a direction realizing that boost:run needs to take the -Dprop=val on the command-line and set that on JVM_ARGS as it starts/runs the server.
Sounds like we're close... though you still have to worry about merge into an existing variable value
Scott Kurz
@scottkurz
So @awisniew90 .. after looking over the JDBC work.. it seems like we could extract out another issue to allow the use of a user-provided DB URL (which, like for Derby, could be a file location). For the JPA case, we also want to support this same user-provided DB URL but also consider how to expose options for whether the tables need to be created/dropped or not..... thinking quickly makes me thing there are three options: a) create (OK, if tables are there, create if not), b) drop-then-create (I need new tables) , c) don't create (if someone else didn't create the tables then don't, it's an error).
I think there's some overlap with the URL, say in Derby, where I believe you can put the create into the URL.
Do the a), b) and c) need to be applied to the first case? Not sure... perhaps.
Scott Kurz
@scottkurz
Maybe not the ideal time of week/year to bring up something new... but wondering what anyone thought about binding a default phase of start, stop, to pre/post-integration-test, respectively....so you could do something like this:
                <execution>
                    <goals>
                        <goal>start</goal>
                        <goal>package</goal>
                        <goal>stop</goal>
                    </goals>
                </execution>
```
Scott Kurz
@scottkurz
Maybe I'll use this downtime to merge a mass format https://github.com/OpenLiberty/boost/pull/231/checks of Java anyway.
Would be nice to have something for Groovy or at least static code analysis
Will let the checks run first though
Scott Kurz
@scottkurz
That only caught about 20 files but still
Scott Kurz
@scottkurz
Was going a bit crazy trying to figure out why the Travis builds for simple formatting changes were failing. It's because the springBoot2 image was just updated to point to 18004. I couldn't recreate locally until doing docker pull open-liberty:springBoot2...then the same "test-spring-boot-docker" test fails.
Scott Kurz
@scottkurz
OK.. I figured out it's the url http://localhost:9080/spring/ which is the problem...only http://localhost:9080/spring works. I wonder if this is an OL change in 18004?
Cheryl King
@cherylking
@scottkurz Should you open an issue in OL? That sounds like a regression.
Scott Kurz
@scottkurz
@cherylking ... I did send a message to Tom and @anjumfatima90 is looking into it now.
Cheryl King
@cherylking
:thumbsup:
Scott Kurz
@scottkurz
So @ajm01 or anyone.. did we come up with a useful string to put in the Travis after_failure: entry to dump our logs?
Scott Kurz
@scottkurz
E.g. the gradle build fails here: https://github.com/OpenLiberty/boost/pull/235/checks any ideas?
That build took about 2 hrs... I removed the exclusions we'd "temporarily" made for 18002 to allow more builds through... maybe I should put them back. Guess I'll do that and try again
Matt Bowersox
@mattbsox
You can try rerunning that specific build in travis. Might be related to the Java 9 issue we ran into with 18.0.0.3
Scott Kurz
@scottkurz
BTW..about the after_failure comment.. I don't think I was keeping up..we already use the travis.sh to collect some logs here... so not sure what to do about a test failing
Scott Kurz
@scottkurz
Thx for the suggestion @mattbsox ....so you were wondering if this failure in 18004 was related? Hmm...well.. I thought 20 builds (2 hrs) was too much so I force-pushed over that anyway...we'll see what happens now. There's still one with the failing config: oraclejdk9 \["RUNTIME=wlp RUNTIME\_VERSION=18.0.0.4 BUILD\_TOOL=gradle"\]
My last rev is down to 18..still pretty long though.
Adam Wisniewski
@awisniew90
I've added a few unittests to test the JDBC / datasource configuration. The PR for adding a defined derby dependency should be ready for review now for whenever someone gets a chance #229
For usage documentation, I can't remember what we decided on before the break, but I think adding a section in the github wiki makes the most sense. The alternative would be adding some .md files in the repo, but I think that could get messy.
Adam Wisniewski
@awisniew90
@scottkurz - I added a "usage" page in our wiki where we can keep track of externals we add like new boosters, plugin goals, etc. Let me know if this is more or less what you were thinking:
https://github.com/OpenLiberty/boost/wiki/Using-Boost-in-a-Jakarta-EE-application
Cheryl King
@cherylking
@awisniew90 Looks like a good start to me.
Adam Wisniewski
@awisniew90
The travis failures for #229 seemed pretty random. I restarted the failing tasks and all is passing now.
Adam Wisniewski
@awisniew90
I'm picking up #205 for using DB2 w/ username and password. The issue also says "w/ container-mgd alias." What does container-mgd alias mean in regards to a datasource?
Adam Wisniewski
@awisniew90
Scratch that. Talked to Scott and created/picked up #240
Adam Wisniewski
@awisniew90
@chyt - FWIW, going back to our conversation about the JDBC / JPA booster interaction, looks like spring boot is transitively defining their jdbc starter when a user specifies the jpa starter.
Adam Wisniewski
@awisniew90
PR for initial boost property function when someone gets a chance to review: OpenLiberty/boost#247
Adam Wisniewski
@awisniew90
Have a quick PR for adding an IT for db2 support: OpenLiberty/boost#250
Ozzy Osborne
@BarDweller
Heads up that the link to https://openliberty.io/blog/2018/09/12/build-and-push-spring-boot-docker-images.html on the page https://www.ibm.com/blogs/bluemix/2018/11/open-liberty-loves-spring/ has an extra -plugin in the link that prevents it working =).
YK Chang
@yeekangc
Client is asking for Spring on Open Liberty with Gradle. They want a packaged jar with OL as the runtime in it. Is the Boost plugin for Gradle ready? That is fine to share with client? Not sure if it's alpha or beta or ...