🤖 This is an automated response from Repo Assist.

The linter tests are currently disabled — linterTests createServer is commented out in Program.fs (line 119):

//linterTests createServer

So they don't run in CI at all, which explains why the flakiness hasn't been blocking builds.

The likely root cause of the original flakiness:

The test in ExtensionsTests.fs subscribes to linterDiagnostics "Script.fsx" and waits for the first non-empty event from the "F# Linter" diagnostics stream. But FSharpLint processes rules independently and can publish results in multiple batches — so the first event may contain fewer than the 9 expected diagnostics, causing Seq.zip to silently truncate the comparison, or the comparison to fail if order differs between runs.

A more robust approach would be to collect diagnostics until a stable state (e.g. a fixed timeout after which no new linter diagnostics arrive), then assert on the full set. The test infrastructure's Observable.debounce or accumulation patterns could help here.

If someone wants to re-enable these tests, the fix would be in linterTests in ExtensionsTests.fs — change from "wait for first non-empty linter event" to "collect all linter diagnostics after a settling period".

Generated by Repo Assist

Generated by Repo Assist · ◷

To install this agentic workflow, run
gh aw add githubnext/agentics/workflows/repo-assist.md@d1d884596e62351dd652ae78465885dd32f0dd7d

The Linter Diagnostics test is very flaky, it needs to be looked at #700

Description

Metadata

Metadata

Assignees

Labels

Type

Fields

Projects

Milestone

Relationships

Development

Issue actions