agentleFS
aviateskaviatesk/JETLS.jl/.agents/skills/write-test/SKILL.md

write-test

Skill319 starsChanged 10 days agoMITView on GitHub

Use when adding or modifying JETLS tests. Covers test file and module structure, `@testset` organization, `let` blocks, `withserver` usage, cached syntax trees in helpers that call features directly, and when subroutine tests are sufficient for language-server features.

8 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 1 is distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when you add new tests or modify existing JETLS test code.

    line 12
  2. Very specific:

    Test code for new language server features should be written in files that define independent module spaces with a test_ prefix.

    line 16
  3. Very specific:

    Use @testset "testset name" to organize tests cleanly.

    line 43
  4. Very specific:

    Use let blocks to avoid unintentionally reusing names across test cases, unless a helper function or do block already provides local scope.

    line 49
  5. Very specific:

    Leave cache_tree0 unset when the FileInfo is only used to convert positions or stands for a file that is not open in the editor, since …

    line 131
  6. Very specific:

    When you add or modify tests, use the run-test workflow to run the most specific relevant test command.

    line 137

“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Write JETLS tests

Use this skill when you add new tests or modify existing JETLS test code.

Test file structure

Test code for new language server features should be written in files that define independent module spaces with a test_ prefix. Include those files from JETLSCore/test/runtests.jl.

This lets each test file run independently from the REPL.

For example, JETLSCore/test/test_completions.jl should look like:

module test_completions
using Test
...
end # module test_completions

Then include it from JETLSCore/test/runtests.jl like this:

@testset "JETLS.jl" begin
    ...
    @testset "completions" include("test_completions.jl")
    ...
end

Organizing test code

Use @testset "testset name" to organize tests cleanly.

For code clarity, avoid placing using, import, or struct definitions inside @testset blocks unless it is specifically necessary. Prefer top-level definitions.

Use let blocks to avoid unintentionally reusing names across test cases, unless a helper function or do block already provides local scope.

Example:

module test_completions

using Test
using JETLSCore: some_completion_func

function testcase_util(s::AbstractString)
    ...
end
function with_testcase(s::AbstractString)
    ...
end

@testset "some_completion_func" begin
    let s = "..."
        ret = some_completion_func(testcase_util(s))
        @test test_with(ret)
    end
    let s = "..."
        ret = some_completion_func(testcase_util(s))
        @test test_with(ret)
    end

    with_testcase(s) do case
        ret = some_completion_func(case)
        @test test_with(ret)
    end
end

end # module test_completions

Language-server feature tests

Testing language server functionality is challenging. To fully test such functionality, you need to start a server loop and send requests that mimic realistic user interactions.

For those tests, use withserver from JETLSCore/test/setup.jl. Refer to JETLSCore/test/test_full_lifecycle.jl as an example.

A good pattern is JETLSCore/test/test_definition.jl: it keeps the full withserver coverage to one request/response sanity test for the DidOpen → analysis → DefinitionRequest path, while most cases use definition_test to call find_definition directly and assert on returned locations. Prefer this split for LSP handlers.

Cached syntax trees

Requests on a synchronized document share the cached syntax tree of its FileInfo without copying it, so no feature may mutate that tree. Under JETLS_TEST_MODE, JETLS records a fingerprint of each cached tree and JETLSCore.check_syntax_tree0(fi) throws if the tree has changed since. withserver runs this check for every open document when the test finishes.

Helpers that construct their own FileInfo and call a feature directly should construct it with cache_tree0 = true, as cache_file_info! does for synchronized documents, and call JETLSCore.check_syntax_tree0(fi) after the feature call:

function find_definition(text::AbstractString, pos::Position)
    server = JETLSCore.Server()
    filename = joinpath(@__DIR__, "testfile_$(gensym(:definition)).jl")
    fi = JETLSCore.FileInfo(#=version=#0, text, filename; cache_tree0 = true)
    furi = filename2uri(filename)
    JETLSCore.store!(server.state.file_cache) do cache
        Base.PersistentDict(cache, furi => fi), nothing
    end
    result = JETLSCore.find_definition(server, furi, fi, pos)
    JETLSCore.check_syntax_tree0(fi)
    return result
end

Leave cache_tree0 unset when the FileInfo is only used to convert positions or stands for a file that is not open in the editor, since JETLS doesn't cache a tree for such files in production either.

After writing tests

When you add or modify tests, use the run-test workflow to run the most specific relevant test command.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.