-
Notifications
You must be signed in to change notification settings - Fork 81
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
TCK Bugs? #217
Comments
hey @MattGill98 - thanks a lot for the feedback. For the first issue, I believe it's a gap in the MP Config TCK, as its design outlines that I agree with your observation of the second issue. I propose that we update the path of Thoughts? |
Hi Arthur, Thank you for your prompt response. That's not a problem. That being the case, I'll raise an issue for the Config API project! That would work absolutely fine. I'll make a pull request making this change. Kind regards, Matt |
Hi @arthurdm, Just following up on the first issue again. I found that the issue raised here on the Config API project specified that for a WAR file the Kind regards, Matt |
My understanding of the Config API is that for a war file the properties should be under WEB-INF/classes/META-INF see eclipse/microprofile-config#268 |
thanks for the discussion guys. My view on this subject is that the MP OpenAPI spec has defined They can too, of course, put their properties files in anywhere else where MP Config picks them up (i.e they can pass the key/value pairs as env variables too), but I don't want to give the impression that From a I hope that clarifies it, but if not, I will be happy to continue the discussion here or in the next OpenAPI hangout. |
I am pretty sure there is an explicit test in the Config api TCK that checks that in a war the properties file is in WEB-INF/classes/META-INF so not sure this will work |
Hi @arthurdm, Thanks for your reply. I see that the |
Right - I agree that the MP Config TCK is checking for the My argument is that the MP OpenAPI spec (not MP Config spec) defined that |
Just curious, but what was the thought process for that particular directory to be chosen? META-INF in a war archive is a rather unusual location really. |
hey @arjantijms - it was deemed a reasonable place for metadata (i.e. OpenAPI description) of the app. |
PR #218 (first issue) was merged - thx again @MattGill98 I have opened #221 to continue the work outlined here in terms of clarifying the location of the specs. Thanks everyone for the feedback! Closing this issue. |
Just a note - I think you misunderstood the MP config design because it clearly says that the path should be on the classpath, while in a WAR, this is in |
Hi @arthurdm
Are there any minutes of that decision, and the people who deemed that, had any of them used or worked with war archives before? My feeling says someone may have thought about a regular .jar file and didn't know exactly what a .war file is. As I think this is a spec bug, I'll file a separate issue for this where it might be better to continue this conversation. |
Hi, I was hoping someone wouldn't mind helping me clarify a few problems I've encountered when working with the TCK.
Firstly, the adding of the
microprofile-config.properties
seems strange. In the Config API TCK, this file is added using this line:This adds the file to
WEB-INF/classes/META-INF
, which is on the application class path.In the OpenAPI TCK however, this file is added using this line:
This adds the file to
META-INF
, which isn't on the application class path. Since the Config API doesn't test for this, it's behaviour is presumably undefined. Should this be an added test to the Config API TCK, or a required change in the OpenAPI TCK?My second problem is regarding the AirlinesAppTest. This application seems to be invalid, and as such may not deploy to all application servers. The endpoint in question involves these two methods in the
ReviewResource
:There's no problem with this from the perspective of the OpenAPI, as they will appear as different endpoints. For the JAX-RS server though these endpoints are functionally the same, and so may not deploy correctly on all application servers. In my opinion this should not cause a failed test however, and maybe one of these endpoints should be fixed to have a different
@Produces
type. Is this an accurate evaluation?Thanks in advance,
Matt
The text was updated successfully, but these errors were encountered: