wordy: update to latest

[ErikSchierboom]

This demonstrates how easy it is to handle errors

[github-actions]

This PR touches files which potentially affect the outcome of the tests of an exercise. This will cause all students' solutions to affected exercises to be re-tested.

If this PR does not affect the result of the test (or, for example, adds an edge case that is not worth rerunning all tests for), please add the following to the merge-commit message which will stops student's tests from re-running. Please copy-paste to avoid typos.

[no important files changed]

For more information, refer to the documentation. If you are unsure whether to add the message or not, please ping @exercism/maintainers-admin in a comment. Thank you!

[tasxatzial]

Nice. The error handling is an open and stale issue atm. See #670

The reason I haven't yet synced exercises that include errors is that there isn't a consensus on the topic. There are many ways to handle errors, but my personal preference is to throw some kind of exception, usually either a generic Exception or an IllegalArgumentException. However, the error messages from the canonical data should probably be incorporated into the tests. For instance, this exercise includes two distinct error messages: 'unknown operation' and 'syntax error.' It is expected that participants should write code that can distinguish between these messages.

The main question, then, is: What kind of exceptions should be returned for each message?

[ErikSchierboom]

The main question, then, is: What kind of exceptions should be returned for each message?

Good question. I don't have much Java experience so I'm more than happy to go with what you'd like.

[tasxatzial]

Then we can keep the IllegalArgumentException for all errors and also include the error message in the tests.

[ErikSchierboom]

I'll need to fix the example implementation. Do the tests look good to you?

[tasxatzial]

Not 100% sure, but I think you've written them correctly.

[ErikSchierboom]

@tasxatzial I've fixed the example implementation, although I feel like the canonical data is slightly weird. In particular:

{
      "uuid": "8a7e85a8-9e7b-4d46-868f-6d759f4648f8",
      "description": "Non math question",
      "property": "answer",
      "input": {
        "question": "Who is the President of the United States?"
      },
      "expected": {
        "error": "unknown operation"
      }
    },

why isn't that a syntax error?

This test case makes sense for an unknown operation:

{
      "uuid": "3ad3e433-8af7-41ec-aa9b-97b42ab49357",
      "description": "unknown operation",
      "property": "answer",
      "input": {
        "question": "What is 52 cubed?"
      },
      "expected": {
        "error": "unknown operation"
      }
    },

FYI It is fine to deviate/tweak the canonical data if we feel that it benefits the exercise. We should at least consider that here I think

[tasxatzial]

Agree, this looks like a syntax error. So this is a case for a manual tweak?

[ErikSchierboom]

Indeed. I'll update

[ErikSchierboom]

Updated

[tasxatzial]

Tested and seems to be working fine.

There's a sort of weird situation going on with the error handling in tests, that I feel is worth noting. The tests will pass only if the error thrown is a string that contains the string found in the tests.

Also, calling Character/isDigit without wrapping it in anonymous function is a new feature in Clojure 1.12.0. @ErikSchierboom I'm thinking of bumping leiningen configs to 1.12.0 to allow a bit more flexibility if someone is using it locally.

[ErikSchierboom]

The tests will pass only if the error thrown is a string that contains the string found in the tests.

I've made the tests more specific, where it has to match the string exactly.

I'm thinking of bumping leiningen configs to 1.12.0 to allow a bit more flexibility if someone is using it locally.

Sounds good!

[tasxatzial]

I've made the tests more specific, where it has to match the string exactly.

The error messages is a topic I raised in discord to get an idea what people would do if the message is modified for some reason. Here's the thread https://discord.com/channels/854117591135027261/1303832856534974464/1303832859815186474

What do you think the best approach would be regarding the check for the error messages?

[ErikSchierboom]

What do you think the best approach would be regarding the check for the error messages?

I think that the chance of the error messages changing is low enough that it is fine to check for the specific message. I value people using the right error message for their errors enough that we should probably check for it. Are you okay with that?

[tasxatzial]

I value people using the right error message for their errors enough that we should probably check for it.

No problem. If the message changes the regex can be adjusted to avoid invalidating solutions.