Multiple test runners not working

The same problem affects c# track.

I think it’s broken for all test runners. I can get it to work occasionally but get this error most of the time.

I’ll use the CLI for now but they still count as failing.

Things should come back online over the next 20mins. Sorry for the pain. You can track the conversation in more depth in #social on our Discord:

1 Like

Same here for Rust.

Javascript as well.

It was this, yes.

1 Like

Speaking of Discord social, I would like it a lot more if people used threads as in Slack. I don’t really like scrolling through a bunch of stream-of-consciousness stuff I’m not interested in. If threads were used I could skip that stuff. Just sayin’.

It might be helpful to mention it where it is happening. Exactly the right audience.

Yeah, I just posted in server-feedback.

We discussed this at length before launching Discord and it was clear that people who use Discord felt Discord should be used in the “Discord way” which is much more realtime conversation, without the idea that people would go back through history, and thus more like IRC. The feeling was that we’d constantly be having to “correct” everyone that used it to use it “our way” with threads rather than the normal Discord way “with mainline conversation”.

I felt the same as you did, Bob, which is threads make things much saner for me, but my conclusion was that trying to push against the tide (of normal Discord usage) would probably just make it a painful experience for everyone and probably mean it was a ghost-town (like Slack became too).

So I use it to read things if I’m around and don’t really go back in history and read older stuff. But then use the forum for longer-form discussions where I don’t want to miss things.

1 Like

Perhaps an issue is that sometimes things are put in social that might better belong elsewhere. I remember that we did have a channel or two in Slack that was threadless for the reason/s others have expressed. I suppose I’ll just avoid social for the most part going forward and hope I don’t miss something “non-social” that would be of Exercism interest.

I’m trying to always put anything official on the forum so it should be safe to do that.

2 Likes

The reason I suggested mentioning in in social there is because people can socially decide that it is the right thing to do. As opposed to a general server policy. Sometimes people just start to do things because they did not realize what it was doing to impact others.

But no point in fighting a battle there if everyone else is fine with the status quo. But it might encourage some people to thread if they think the discussion start is going to be more focused.

I’m not sure what the official status of the original issue is, i.e. ‘Multiple test runners not working’, (the thread seems to have drifted into discussing Discord) however I’ve been seeing the same issue for the last few days.

I’m currently working in the Julia track. When I submit my local solution to the website quite often (but not always) I’ll see an initial message indicating that my solution passed all the tests. Perhaps, after as little as a minute, I’ll see the little red dot along with the generic error message reported above - somehow my tests have gone from passing to failed :face_with_raised_eyebrow:. If I then work in the editor with my submission unchanged, run the tests I find that everything works and I can submit my solution.

As a workaround I’ll try copy-pasting my local solutions into the editor and submit via the website instead.

@blogscot I think this might just be an issue with Julia. I’ve heard similar things. We never tracked it down, but I’ll revisit this when I get chance once we’ve got a load of our existing PRs merged in the next few days. If you could provide any exercises where this has happened, we’ll take a look. Thanks :)

Certainly not just Julia. See also #5826.

In the last few days I’ve seen this on vbnet (multiple exercises) and PureScript (hello-world).

With purescriot I’m almost positive it is due to the test runner being slow. This could also be the case with the vbnet one.

That could explain why it seems to be more common on low-traffic tracks with few/no maintainers. I’d previously seem it on CFML, which fits the same pattern.

Though it’s a bit of a problem if they can’t run hello-world!

For the Julia Hamming exercise I ended up having 4 iterations. I deleted the first and third iterations as they were reported as passed-then-failed. The second and fourth iterations are re-submissions of those solutions without changes.

Additionally, on Discord, Adrien ANTON LUDWIG is reporting a similar sounding bug on the Unison track.

1 Like

3 posts were split to a new topic: Could local tests count as verified when submitted?