AL Runner v2.2 and v2.3: Coverage, Quick Scripts, and a Watch Mode You Can Trust
AL Runner v2.2 and v2.3 added code coverage reporting, a quick-script execute command for testing snippets without a codeunit, and fixed a watch mode that was recompiling mid-save during file bursts. The releases also included roughly 130 behavioral fixes to align the runner with real Business Central, with startup performance improvements.
Read it hereOpen in Claudehttps-stefanmaron-com-posts-al-runner-coverage-execute-and-a-guard-that-worked--6c4b521103.md
post Stefan Maron
"If the runner and real BC disagree, the sandbox result wins and the runner gets fixed to match it, not the other way around."
Key points
- The –coverage flag gives a statement-level report of which AL lines tests actually ran, based on the compiler’s own internal tracking
- The execute command runs a bare AL snippet or a full pasted codeunit without needing a test codeunit
- Watch mode now waits until file changes settle (capped at 10 seconds) instead of a fixed quarter-second delay, avoiding false failures on branch switches or bulk edits
- Editing an existing object’s body in watch mode now re-tests without a full app rebuild, about 18x faster on a 2,000-object app
- About 130 fixes align runner behavior with real BC (TestPage, manual event subscribers, transactions, report filters), and startup overhead dropped by roughly a quarter
Quotes
- “If the runner and real BC disagree, the sandbox result wins and the runner gets fixed to match it, not the other way around.” (Ensures runner behavior aligns with actual BC behavior, making tests reliable predictors of production behavior)
Context
- Features: code coverage reporting, quick execute command, watch mode, AL test execution, statement coverage, file change detection
- Versions: 27.0, 28.0, 28.1, 28.2, 28.3, 28.4
Source: Stefan Maron, community blog. Summary, key points and quotes are derived (CONTENT-NOTICE.md); read the original for the full text.