put all websocket sessions in a single static map

[nverwer]

This fixes #305: The console:log function could only send to one channel, because it could only reach one session. By putting all sessions in a static map, console:log can find the correct session belonging to the channel it wants to log to.

[duncdrum]

@nverwer thank you for your contribution, much appreciated.

My only concern is that the desired behaviour isn't tested with the PR, so it will be tricky to maintain going forward. If you need help with creating a test within the project set-up let us know and we can figure out where to best to include it.

I've added the PR to the agenda of next weeks community call.

[nverwer]

Thank you, @duncdrum . I am working on different projects, and it might take some time for me to figure out how to make an effective test. I will have a look as soon as I find the time for it.

[line-o]

I don't think this seven character fix is blocked by a missing test. I would still want to get a test for it but this should not prevent this from going in.

[duncdrum]

@nverwer we discussed this PR in the community call if you can't find the time for adding the test. I'll add one via cypress once I can find the time.

[adamretter]

@duncdrum I remember the conclusion at the Community Call differently. I recall that we discussed that this could be tested manually, as we don't have the necessary test infrastructure (that @nverwer would need) available in Java within the project; that we could be pragmatic, and merge it if there is no issue identified from manual testing. I also on the call confirmed that I had previously tested this manually already and seen no issue.

[duncdrum]

@adamretter It is correct that you asserted that this couldn't be tested on the existing infrastructure, we took a look and disagreed. This can be tested from within the existing test infrastructure, and so it should be. You are of course free to expand the

necessary test infrastructure available in Java

in a separate PR.

[adamretter]

we took a look and disagreed

There is no infrastructure in Monex for either Unit or Integration tests written in Java at present. @duncdrum How do you see this working?

[duncdrum]

@adamretter you don't have to test this from within Java. We have integration test configured so I m planning to expand those.

[nverwer]

Dear all, I am on a biking holiday without a computer. I will be back next week, and see if I can come up with a unit test. Op di 15 jul 2025 23:33 schreef Duncan Paterson ***@***.***>:

…

*duncdrum* left a comment (eXist-db/monex#306) <#306 (comment)> @adamretter <https://github.com/adamretter> you don't have to test this from within Java. We have integration test configured so I m planning to expand those. — Reply to this email directly, view it on GitHub <#306 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIMEOQGIGS5DRLSKOQ4CIL3IVXTPAVCNFSM6AAAAACAOPATUSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANZVG44DCMZSGM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[nverwer]

I will be back next week, and see if I can come up with a unit test.

When I wrote that, I clearly had no idea what writing a test would involve, because I have never used Cypress.
With the help of an AI, I made a test, but running it (with mvn release:prepare) results in errors that I am unable to fix.

I pushed my test, so that someone who knows more than I do about this stuff may try to run it.

[duncdrum]

@nverwer thank you. I ll refactor this, I was out and about.

[duncdrum]

now with tests and green

[duncdrum]

@nverwer the changes to the test code are all in here 9c64469 I took the chance to upgrade the framework, and had to rebase.

@eXist-db/core we should cut a release after this is merged, the original change fixes another failing test with console:log

[adamretter]

Test looks good. However the default port should not be 8080. Also that is already solved in this earlier PR #298 which will need to be merged before this one, and then this one will need to be rebased.

mixing concerns

[adamretter]

Needs to have the port issue fixed as it hardcodes it in many places.

[line-o]

I will pull this in as #298 is blocked

[adamretter]

@line-o this PR should not get pulled in as is. It needs to be broken into 2 PRs. This one should only have the Websocket fix and test, it should not have the Cypress upgrade and other test refactorings in it

[line-o]

I would pull in both PRs so I don't see the point of splitting them first.