Most tests are defined as standalone functions. Some are defined in a class though (e.g. for cleanup purposes). This currently breaks the test runner, as the code to capture console output creates a class for the existing code. We should detect if a class is used in the tests and update the rewriting code accordingly.

In line with our new org-wide policy, the master branch of this repo will be renamed to main. All open PRs will be automatically repointed.

GitHub will show you a notification about this when you look at this repo after renaming:

In case it doesn't, this is the command it suggests:

git branch -m master main
git fetch origin
git branch -u origin/main main

You may like to update the primary branch on your forks too, which you can do under Settings->Branches and clicking the pencil icon on the right-hand-side under Default Branch:

We will post a comment below when this is done. We expect it to happen within the next 12 hours.

Hi, I think there is an issue with F# online editor test.

I've already submitted a bug using the 'Rpeort a bug' function on the editor, apology for the duplication, but I didn't receive any confirmation of the bug I submitted. So I decided to submit the bug report again here as it's easier for me to follow up and see the progress of the bug.

When I am running tests, it always gives me the error message "An ERROR OCCURED" without details:

I can't tell if there is any issue with the code, which is relatively simple, and it runs fine inside Visual Studio 2022, so I assume it's a bug in the online editor test runner.

Here is my code if it helps:

module Accumulate

let rec accumulate (func: 'a -> 'b) (input: 'a list): 'b list = 
    match input with
    | [] -> []
    | h :: t -> (func h) :: (accumulate func t)

My solution passes all tests (un-skipped, unmodified) locally. On the test runner, only 1 test passes. And on the test runner, I get different results (only 2 failures) if I .Trim() the header line of the ledger. But the remaining failures have weird whitespace differences still on various lines. I'm not sure what the differences are, but it seems like there are different whitespace expectations on the test runner than on the unit tests file.

Hi,

I am a mentor for F# track and want to help building the test runner.

We've recently updated the test runner interface specification to allow for two types of test runners, identified via a new version property in the results.json file (see the spec for the version property):

1. version: 1 test runners are quite basic, and only detect if either all tests passed or not. These test runners capture the console output of the test runner and output that. They do not include details on individual tests passing/failing
2. version: 2 test runners contain detailed information on individual tests passing/failing. We've recently added a required field named test_code, which contains the code the test ran to verify the behavior (see the spec for the test_code property).

The test runner's output is currently what I refer to as version 1.5: it contains the individual test information, but does not yet contain the test_code key. If possible, the test runner should be updated to also include the test code, in which case the version property should also be added with 2 as its value.

You are completely free on how you want to extract the test code, either via the AST or via finding the test code in the source code text via string manipulation.

Let me know if there are any questions.

Similar to Daimond but this does not use property testing. It is using the test as members of a type which is unusual. Could the rewrite affect that?

Checked with a student whose tests pass locally and then with my existing submitted solution.

Some use Newtonsoft.Json but the majority of solutions use FSharp.Data/FSharp.Data.JsonExtensions or FSharp.Json and there might be other standard libraries.

These should be added to the TestRunner default packages.

Capturing console output in xUnit 2 is annoying, as you have to work with test classes and constructor injection. Writing the code to support this is annoying, complex and brittle. An option might be to use NUnit as the test framework.
See https://stackoverflow.com/questions/46405753/how-do-i-get-nunit3-console-to-output-my-debug-to-screen-on-a-windows-box

Side note: due to us using FsUnit, the actual test code might not even change much.

Running locally, the tests generally need <80 ms, never more than 300 ms on multiple machines and OSs. Hard to know why they won't run on your server? The code has no loops, no recursion.

module AnnalynsInfiltration

let canFastAttack (knightIsAwake: bool): bool = not knightIsAwake

let canSpy (knightIsAwake: bool) (archerIsAwake: bool) (prisonerIsAwake: bool): bool =
    knightIsAwake || archerIsAwake || prisonerIsAwake

let canSignalPrisoner (archerIsAwake: bool) (prisonerIsAwake: bool): bool =
    prisonerIsAwake && not archerIsAwake

let canFreePrisoner (knightIsAwake: bool) (archerIsAwake: bool) (prisonerIsAwake: bool) (petDogIsPresent: bool): bool =
    (petDogIsPresent && not archerIsAwake) || 
    (prisonerIsAwake && not knightIsAwake && not archerIsAwake)

In line with our new org-wide policy, the master branch of this repo will be renamed to main. All open PRs will be automatically repointed.

GitHub will show you a notification about this when you look at this repo after renaming:

In case it doesn't, this is the command it suggests:

git branch -m master main
git fetch origin
git branch -u origin/main main

You may like to update the primary branch on your forks too, which you can do under Settings->Branches and clicking the pencil icon on the right-hand-side under Default Branch:

We will post a comment below when this is done. We expect it to happen within the next 12 hours.

These are probably the two most used and popular addons when coding in FSharp for adding code and data structures respectively. (Already requested FSharp.Data and Json packages in other just posted issue)

I frequently find that when I submit code for F# exercises that passes all tests locally in a few ms, the tests will timeout on Exercism and I have to resubmit with changes to the source that have no functional impact but bypass the check on exercism submit that the code is not identical. This doesn't seem to happen nearly as frequently with any other languages in my experience.

We've recently updated the test runner interface specification to allow for two types of test runners, identified via a new version property in the results.json file (see the spec for the version property):

1. version: 1 test runners are quite basic, and only detect if either all tests passed or not. These test runners capture the console output of the test runner and output that. They do not include details on individual tests passing/failing
2. version: 2 test runners contain detailed information on individual tests passing/failing. We've recently added a required field named test_code, which contains the code the test ran to verify the behavior (see the spec for the test_code property).

The test runner's output is currently what I refer to as version 1.5: it contains the individual test information, but does not yet contain the test_code key. If possible, the test runner should be updated to also include the test code, in which case the version property should also be added with 2 as its value.

You are completely free on how you want to extract the test code, either via the AST or via finding the test code in the source code text via string manipulation.

Let me know if there are any questions.

Ensures that all Skip properties are removed, regardless of the type of attribute. Without this, the diamond exercise's tests won't be unskipped (they use property-based testing).

See exercism/csharp-test-runner#67 for the C# equivalent PR.

Test runner fails with the following error if the test file is big in size.

{
  "status": "error",
  "message": "BeerSongTest.fs(99,1): error FS0010: Incomplete structured construct at or before this point in expression. Expected ']' or other token. \nBeerSongTest.fs(71,9): error FS0598: Unmatched '[' \nBeerSongTest.fs(70,5): error FS0588: The block following this 'let' is unfinished. Every code block is an expression and must have a result. 'let' cannot be the final code element in a block. Consider giving this block an explicit result. \nBeerSongTest.fs(98,11): error FS0514: End of file in string begun at or before here",
  "tests": []
}

To produce, try running it for beer-song exercise.

For the Diamond exercise in F#, the submission below fails with an empty error message, stating only "We received the following error when we ran your code", but no actual reason.

Looking at the dev console in my browser, I see multiple calls to the "test_run_ endpoint returning

{"test_run":null}

with status 304. If I modify my submission slightly by defining

let space = '.' // instead of let space = ' '

then I receive a "test_run" response with status 200 and some expected test failure messages (after some empty 304s as above).

The code submission:

module Diamond

let alphabet = [|'A'..'Z'|]
let space = ' '

let createUpperHalfRow row size =
    let center = size / 2
    if row = 0 then
        Array.init size (fun i -> if i = center then alphabet.[row] else space)
    else
        Array.init size (fun i -> if i = center - row || i = center + row then alphabet.[row] else space)


let createUpperHalf size =
    Array.init (size / 2 + 1) (fun i -> createUpperHalfRow i size)

let mirrorLowerHalf upperHalf =
    let length = Array.length upperHalf
    Array.init (length - 1) (fun i -> upperHalf.[length - i - 2])

let make letter = 
    let size = (Array.findIndex ((=) letter) alphabet) * 2 + 1
    let upperHalf = createUpperHalf size
    let lowerHalf = mirrorLowerHalf upperHalf

    lowerHalf
    |> Array.append upperHalf
    |> Array.map System.String
    |> String.concat System.Environment.NewLine