Time
S
Nick
Message
07:01 jri joined #dataverse
07:32 jri_ joined #dataverse
09:58 Youssef_Ouahalou joined #dataverse
10:42 jri joined #dataverse
11:38 Youssef_Ouahalou joined #dataverse
11:44 poikilotherm
Guten Morgen Amerika. Wie ist das Wetter?
11:47 donsizemore joined #dataverse
13:22 poikilotherm
Heads up donsizemore14 I'm off to construction site now, but I just created https://github.com/IQSS/dataverse/pull/6800
13:24 poikilotherm
I'm really eager to see more Testcontainers stuff in use and favor that way of doing things. I saw your ansible approach, but from a devs perspective it would be great to have a better tooling for more integration tests around...
13:24 poikilotherm
Obviously these are just my 2 cents...
13:25 poikilotherm
Nonetheless it would be great to form an alliance to push things forward to modern testing approaches together
13:25 poikilotherm
(Whatever those might look like for Dataverse)
13:42 donsizemore
@poikilotherm you're the expert!
14:08 pdurbin joined #dataverse
15:26 jri joined #dataverse
15:44 pdurbin
donsizemore: mornin'. Have you looked at https://github.com/IQSS/dataverse/pull/6800 ? I'm looking now. It's related to the S3 testing you're working on (thanks).
15:45 donsizemore
@pdurbin yes. looks good to me but @poikilotherm will know much better than i
15:48 pdurbin
ok, thanks
15:57 donsizemore
according to my reading yesterday you stand up the container and do the rest in the codebase
15:57 donsizemore
(or you can spin up the container using the codebase, but i thought i'd have ansible prop up whatever it can up front)
16:05 jri joined #dataverse
16:07 pdurbin
Yeah, something like that. It's interesting/weird to me that if you want PostgreSQL, you add it to pom.xml.
16:07 pdurbin
Anyway, I left a couple comments for poikilotherm.
16:30 jri joined #dataverse
17:08 pdurbin
donsizemore: what version of Payara are you running on those test servers, please?
17:09 donsizemore
5.201 i think
17:09 donsizemore
gimme a second and i can verify
17:10 donsizemore
yes 5.201
17:10 pdurbin
Ok. And the "community" edition, I assume.
17:12 donsizemore
correct
17:13 donsizemore
and CentOS 8, though that shouldn't really matter.
17:13 pdurbin
Thanks. I just picked up this issue about docs: https://github.com/IQSS/dataverse/issues/6593
17:13 donsizemore
so far the only non-standard update is Gustavo's patched jakarta.faces.jar
17:15 pdurbin
Right. Like how we patch Weld and Grizzly.
18:19 Gera_fp joined #dataverse
18:21 Gera_fp
Hi all!
18:41 poikilotherm
Hi Gera_fp
18:42 poikilotherm
pdurbin r u still around?
19:24 pdurbin
poikilotherm: hi, just got back. Did you see my review?
19:24 pdurbin
Gera_fp: hi
19:25 poikilotherm
Yeah. If you have a minute, we could talk about it :-)
19:27 pdurbin
Sure. Video chat or just here?
19:29 poikilotherm
I'd prefer IRC because construction site...
19:30 pdurbin
ok
19:32 poikilotherm
It boils down to that one comment about the skip thing, right?
19:32 pdurbin
yeah
19:32 poikilotherm
The other comment was just a typesetting thingy
19:33 pdurbin
yeah, just capitalizing a word for consistency
19:33 poikilotherm
Great. So here's why. If you want to run integration tests with Maven, you should use the failsafe plugin.
19:34 pdurbin
I thought we used surefire or something. I forget.
19:34 poikilotherm
And it's good common practice to run this during the verify phase of Maven
19:34 poikilotherm
That one isn't currently used
19:35 pdurbin
We don't use the verify phase?
19:35 poikilotherm
Yes, you do. We talked about that some time ago, maybe a year or one and a half.
19:35 poikilotherm
Not as far as I know.
19:36 poikilotherm
The IT tests are run manually by running mvn test and specifying the test classes on cmf line
19:36 poikilotherm
Cmd
19:36 pdurbin
I seem to remember playing with the verify phase but let's not worry about that.
19:37 poikilotherm
Which is, by the way, no good. I see IT test classes in the codebase, not included in the test scripts.
19:37 pdurbin
Right, we specify an ugly comma separated list of FooIT,BarIT, etc
19:37 poikilotherm
Yeah, however.
19:37 poikilotherm
I didn't want to touch any of that stuff
19:37 poikilotherm
Because staying out of way was priority for me
19:38 pdurbin
Right. I got excited about your "stay out of the way" goal. Then I saw some skip stuff that confused me.
19:38 poikilotherm
But when introducing more integration tests, I really wanted to push for using the proper tools and BCPs around, which is running IT during verify with failsafe.
19:38 pdurbin
BCPs?
19:39 poikilotherm
Best common practice
19:39 pdurbin
oh
19:39 poikilotherm
Inline with tool documentation etc ;-)
19:40 poikilotherm
So. Back to verify. The circle is gonna close soon, bear with me
19:40 pdurbin
:)
19:41 poikilotherm
When you look at maven, the verify phase is always beyond compile, test and package phases
19:41 pdurbin
verify comes after. ok
19:41 poikilotherm
So when you run verify, all the unit tests will be executed, too.
19:42 poikilotherm
I thought for introducing a new tech, that wouldn't be cool, as it makes you wait.
19:42 poikilotherm
Any time you try out the new cool stuff, you will have to do the unit tests, too.
19:44 poikilotherm
So I wanted to skip those for now. That could be reverted later or left for good. Because the only way to skip ONLY the unit tests is by switching them off via configuration. There is -DskipTests, but this will always skip verify, too!
19:44 poikilotherm
That's why I came up with the variable and everything, because when activating the profile, the global configuration, of which the surefire unit tests are part of, will always be in the loop.
19:45 pdurbin
Has the circle closed?
19:45 poikilotherm
There is simply no other chance to skip unit tests
19:45 poikilotherm
So adding this is also more or less a BCP for projects that might have such constructs
19:45 poikilotherm
I also thought it might come in handy for CI
19:46 poikilotherm
I hope so :-D
19:46 pdurbin
:)
19:46 pdurbin
Is <skip> valid in a pom.xml file?
19:46 poikilotherm
Long story short: make the new profile less bloated by skipping unit tests when activated
19:47 poikilotherm
Definitely
19:47 poikilotherm
Please take a look at the surefire docs
19:47 poikilotherm
It's a configuration setting for the plugin
19:47 poikilotherm
Just like the exclude test stuff above it
19:48 poikilotherm
I could of course add a general note about skipping unit tests in the test guide.
19:49 poikilotherm
In both places: mention auto skip for the TC profile and mention as a handy shortcut during developing.
19:49 poikilotherm
Sometimes you might just want to compile and package without running the unit test again
19:50 pdurbin
Meh. It's fine. I get it. Thanks for the explanation. Are you ready for me to move it to QA?
19:50 poikilotherm
Does all of this make sense now or in any way?
19:51 poikilotherm
Dunno. There is that docs change still missing
19:51 pdurbin
Sure. I get it. You don't want to run the unit tests (which get run all the time) when spinning up containers and doing other magic.
19:52 poikilotherm
Aye. It just adds more time, which might not be cool in all situations.
19:53 pdurbin
I'm looking forward to our first use of Testcontainers. Here we're just getting set up. Small chunks, small diffs, like we ask for. Thank you!
19:54 poikilotherm
If you want to see an example of using it, I'm happy to dig for a link
19:54 poikilotherm
In my dataverse code
19:55 pdurbin
I can wait. :) Do you want me to click commit on my doc suggestion and move it to QA?
19:55 poikilotherm
Which I moved to "rip apart, refactor for EE8" on scolapastas request
19:56 poikilotherm
Can you commit a change from a comment to the git branch??? That's pretty cool!
19:56 pdurbin
sure, you use the ```suggestion
19:56 pdurbin
type something here
19:56 pdurbin
```
19:57 pdurbin
syntax
19:57 pdurbin
Would you rather click the "Commit suggestion" button? To see how it works?
19:57 poikilotherm
Now that's a smart move from GitHub! I didn't knew that!
19:57 poikilotherm
Great, please go ahead.
19:58 poikilotherm
I don't see that button...
19:58 poikilotherm
Looks like only the comment author has it
19:58 pdurbin
done
19:59 poikilotherm
Wow!!!
20:00 poikilotherm
Should we add a note in the docs about the skip thingy?
20:00 poikilotherm
And then resolve the comment?
20:00 pdurbin
I'm putting it in my next review. Please stand by.
20:01 poikilotherm
Can you do that in the same way???
20:01 poikilotherm
That would be f*g awesome... Easy changes via web UI
20:01 pdurbin
no but I'll link to my second review, I guess
20:04 poikilotherm
Marvelous
20:06 pdurbin
done
20:06 pdurbin
yeah, I just linked to the second review
20:07 pdurbin
in the comment about <skip>
20:08 pdurbin
poikilotherm: thanks again!
20:08 poikilotherm
Great. It's so awesome to have you back...
20:08 poikilotherm
Things moving again :-D
20:09 pdurbin
I'm a little blocked on updating docs for Payara though.
20:09 poikilotherm
Thank you so much for supporting and pushing things the right way :-)
20:09 poikilotherm
Uh oh
20:09 Gera_fp
Hi pdurbin, poikilotherm. I have a question respect to a issue in Dataverse version 4.7.1.
20:10 pdurbin
Gera_fp: you're scaring me with your old version but please go ahead. :)
20:10 poikilotherm
Hi Gera_fp
20:10 Gera_fp
It's a old version jeje
20:11 pdurbin
poikilotherm: Gera_fp is from https://data.cimmyt.org
20:11 Gera_fp
Yes, its the main version in CIMMYT.
20:12 poikilotherm
Ah didn't you guys develop the OAuth2 for MS stuff?
20:12 Gera_fp
When I use the OAI. The response is a XML malformed.
20:12 Gera_fp
https://data.cimmyt.org/oai?verb=ListRecords&metadataPrefix=oai_ddi&set=data_cimmyt
20:12 poikilotherm
Just wondering about the version then...
20:13 Gera_fp
You know some information about this error?
20:13 Gera_fp
Yes, We develop this :D
20:15 pdurbin
Gera_fp: I feel like we fixed this. Maybe a while ago. I'm looking.
20:15 Gera_fp
thanks :D
20:18 Gera_fp
poikilotherm, We have the intention of update the Dataverse to a new version but I don't have a date due to the quarantine several plans were delayed :(
20:20 pdurbin
Gera_fp: I can't find anything. Can you please create an issue at https://github.com/IQSS/dataverse/issues ?
20:22 pdurbin
poikilotherm: do you have time to talk about Payara?
20:22 Gera_fp
Ok, I create the issue. Thanks :D
20:23 pdurbin
Gera_fp: thanks!
20:27 poikilotherm
pdurbin hit me
20:27 poikilotherm
I'll try my best. Not in my usual workplace though.
20:28 pdurbin
poikilotherm: so, the way things are going is that we're hacking away on the new Python installer to make it support Payara. But do you plan to run that Python installer? Or are you going to sidestep it and use the Microprofile Config API ?
20:40 poikilotherm
I definitely want to go with config api as it makes things much easier to handle
20:40 poikilotherm
I actually don't need the installer, as I split components anyway in different container images
20:41 poikilotherm
And if you ask me, we really should move a ton of stuff to annotations instead of a script configuring the app server
20:41 poikilotherm
Annotations for JMS etc
20:41 pdurbin
Sure. That all sounds great, but when? I'm working on docs today. :)
20:41 poikilotherm
Those are standardized, calls to an appserver binary are not
20:42 poikilotherm
Well scolapasta just told me that develop is ok to be broken in terms of backward compatibility now
20:43 poikilotherm
Like last Wednesday when we had a VC
20:44 poikilotherm
So maybe instead of working on docs, we should do one step back and look at things again. Most of this stuff is already in issues from two years ago IIRC
20:44 poikilotherm
Because IMHO docs should be done when breaking changes are done, right?
20:45 poikilotherm
Otherwise it would be a great amount of work changing them now and then again after breaking changes
20:45 poikilotherm
And it might lead to the situation that you want to avoid those changes because of lots of work done.
20:45 pdurbin
Well, we want the (future) dev guide to tell developers to use Payara instead of Glassfish. And all the other guides need to be updated as well.
20:45 poikilotherm
Maybe we should start with a list of things we want to break, like we started with some time ago?
20:47 pdurbin
This is what I have so far: https://github.com/IQSS/dataverse/compare/6593-payara-docs
20:47 poikilotherm
Well it's just my 2 cents. IMHO if a bunch of devs does work on this stuff together we can shift things from ye olde ways to modern times.
20:48 poikilotherm
Maybe switch to Jakarta EE in the same go.
20:49 poikilotherm
I will add Jakarta EE 8 and not Java EE 8 for #6694
20:49 poikilotherm
There are some references in the doc you linked to Java EE
20:50 pdurbin
You're so excited about breaking changes and big chunks. :)
20:51 pdurbin
Well, what's the scope of Dataverse 5? Are we switching from Java EE 7 to something newer? :)
20:53 poikilotherm
Yes, Java EE8
20:53 poikilotherm
Which is the same as Jakarta EE 8
20:53 poikilotherm
There is no rela difference between both
20:53 poikilotherm
The big bang is coming with Jakarta EE9
20:54 poikilotherm
But we can state that we are future ready with being a Jakarta EE app not an old Java EE senior
20:54 poikilotherm
Kids like Jakarta EE these days :-)
20:55 poikilotherm
I have to switch to EE8 because scolapasta wants me to use JSON -B from it instead of Jackson.
20:56 poikilotherm
He told me now would be a good time :-D
20:56 pdurbin
https://github.com/IQSS/dataverse/issues/4260 is the Java EE 8 issue but it has the old 5.0 milestone. Let me ask about it.
21:24 pdurbin
Sounds like it may or may not be in Dataverse 5.0.
21:25 poikilotherm
What???
21:25 poikilotherm
No way.
21:25 pdurbin
"surgical"
21:25 poikilotherm
WTF
21:25 poikilotherm
Man you know me, I'm a friendly guy.
21:25 pdurbin
LOL
21:26 poikilotherm
But if I do coding against a standard now and then that gets dropped because you folks want to stay on EE7...
21:27 poikilotherm
Not funny
21:27 pdurbin
I think we just want to keep our app working. And nobody has dropped nothin' yet. Maybe you should leave a comment on that issue.
21:28 poikilotherm
Ah then I misunderstood
21:29 poikilotherm
That issue hasn't been updated in a while
21:29 poikilotherm
I read you asked scolapasta in person and reported back
21:30 poikilotherm
Yeah I will leave a comment about what we came up with last Wednesday
21:30 pdurbin
We're talking in #dv-general (Slack)
21:30 pdurbin
that sounds useful, thanks
21:30 poikilotherm
I summed it up in a comment to #6694 IIRC
21:30 poikilotherm
But you are right, that should also be added there.
21:46 poikilotherm
Leaving construction site, heading home, off to bed with me ;-)
21:46 poikilotherm
Read you tomorrow
21:47 pdurbin
I'm heading out soon too. Take care.
21:51 pdurbin left #dataverse
