From 1392521a8e1a5ef484ab3e2505d88be1a0ec50b3 Mon Sep 17 00:00:00 2001 From: Michael Dieringer <65093775+MichaelDieringer@users.noreply.github.com> Date: Tue, 8 Sep 2026 19:02:03 +0200 Subject: [PATCH] Address second round of Jesper Schulz-Wedde's review on PR #159 - commit-shared-test-fixture-inside-lazy-initialize.md: fundamentally rewritten. AutoCommit is the documented default TransactionModel, not AutoRollback. Explains the real mechanism (Commit() protects a fixture from the test method's own later deliberate rollback, per Codeunit.Run/ TransactionModel-property semantics) and the TestIsolation dependency (Disabled/Codeunit survive across methods, Function does not). Fixtures rewritten to demonstrate the actual failure/success shape. - transactionmodel-attribute-governs-test-transactions.md: now states the AutoCommit default explicitly and agrees with the article above, closing the contradiction Jesper flagged between the two testing articles. - Deleted confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text (.md/.good.al/.bad.al): the underlying platform bug (microsoft/ ALAppExtensions#23935) was closed as completed in Feb 2024; cannot be reproduced or bc-version-pinned on any currently supported version. - table-relation-test-exclude-known-invalid-relations-via-event.md: added the [Scope('OnPrem')] boundary verified against BCApps' Table Relation Test codeunit. - use-assert-isfalse-not-asserterror-for-boolean-checks.md: added a Scope section resolving the overlap with asserterror-needs-expectederror-and-code. - al-testing-review.md: fixed the shared-fixture cue to catch the actual anti-pattern instead of the compliant shape, added the missing cue for use-assert-isfalse-not-asserterror-for-boolean-checks, wired precedence between it and the generic asserterror rule, and removed the cue for the deleted article. - Added in-file Source provenance (specific fluxxus.nl post per article, with what was independently verified vs. taken from the post) to the three surviving externally-inspired articles, per Jesper's request that provenance live in the knowledge file itself, not only the PR description. Co-Authored-By: Claude Sonnet 5 --- ...test-fixture-inside-lazy-initialize.bad.al | 17 ++++++++++- ...est-fixture-inside-lazy-initialize.good.al | 15 +++++++++- ...red-test-fixture-inside-lazy-initialize.md | 16 +++++++--- ...onfirmhandler-sees-substituted-text.bad.al | 9 ------ ...nfirmhandler-sees-substituted-text.good.al | 9 ------ ...re-confirmhandler-sees-substituted-text.md | 30 ------------------- ...clude-known-invalid-relations-via-event.md | 8 +++-- ...del-attribute-governs-test-transactions.md | 8 +++-- ...alse-not-asserterror-for-boolean-checks.md | 8 +++++ microsoft/skills/review/al-testing-review.md | 6 ++-- 10 files changed, 65 insertions(+), 61 deletions(-) delete mode 100644 microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.bad.al delete mode 100644 microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.good.al delete mode 100644 microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.md diff --git a/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.bad.al b/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.bad.al index 50f5d54..171ce5d 100644 --- a/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.bad.al +++ b/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.bad.al @@ -1,9 +1,12 @@ codeunit 50142 "Sample Test Library" { + Subtype = Test; + var Initialized: Boolean; + RollBackMsg: Label 'Revert back the tables to their original state.'; - procedure Initialize() + local procedure Initialize() begin if Initialized then exit; @@ -16,4 +19,16 @@ codeunit 50142 "Sample Test Library" begin // insert master/setup data shared across every test in this codeunit end; + + [Test] + procedure FirstTestUsesSharedFixture() + begin + Initialize(); + + // exercise/verify against the shared fixture, then make scratch changes of its own + + asserterror Error(RollBackMsg); + // the deliberate rollback above also erases the never-committed fixture; + // Initialized still reads true on the next test, but the rows are gone + end; } diff --git a/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.good.al b/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.good.al index bb5e470..1aa16c8 100644 --- a/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.good.al +++ b/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.good.al @@ -1,9 +1,12 @@ codeunit 50142 "Sample Test Library" { + Subtype = Test; + var Initialized: Boolean; + RollBackMsg: Label 'Revert back the tables to their original state.'; - procedure Initialize() + local procedure Initialize() begin if Initialized then exit; @@ -17,4 +20,14 @@ codeunit 50142 "Sample Test Library" begin // insert master/setup data shared across every test in this codeunit end; + + [Test] + procedure FirstTestUsesSharedFixture() + begin + Initialize(); + + // exercise/verify against the shared fixture, then make scratch changes of its own + + asserterror Error(RollBackMsg); + end; } diff --git a/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.md b/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.md index dd46878..3321e05 100644 --- a/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.md +++ b/microsoft/knowledge/testing/commit-shared-test-fixture-inside-lazy-initialize.md @@ -1,7 +1,7 @@ --- bc-version: [all] domain: testing -keywords: [initialize, isinitialized, shared-fixture, commit, autorollback, lazy-initialization] +keywords: [initialize, isinitialized, shared-fixture, commit, autocommit, asserterror, testisolation, lazy-initialization] technologies: [al] countries: [w1] application-area: [all] @@ -11,16 +11,24 @@ application-area: [all] ## Description -A test codeunit that creates master/setup data once, guarded by an `IsInitialized` flag, to avoid repeating expensive setup across many `[Test]` methods depends on that data surviving into every later test. Each `[Test]` method runs under `AutoRollback` by default, so data inserted during the first test's call to `Initialize()` rolls back at the end of that test. `IsInitialized` is a variable, not persisted data, so it still reads `true` on the next test — but the fixture rows it points to are already gone. +A test method with no `[TransactionModel(...)]` attribute defaults to `AutoCommit` (see `transactionmodel-attribute-governs-test-transactions.md`): a method that completes without error commits automatically at its own boundary, with no explicit `Commit()` needed. So a lazy/shared `Initialize()` — guarded by an `IsInitialized` flag, creating master/setup data once to avoid repeating expensive setup across many `[Test]` methods — does not need `Commit()` just to survive into the next test method; under the default model it already will. (Declaring `[TransactionModel(AutoRollback)]` instead is not compatible with this pattern at all: `AutoRollback` assumes the code under test never commits, and a `Commit()` call under it raises a runtime error.) + +What an early `Commit()` inside `Initialize()` actually guards against is the test method's *own later, deliberate* rollback — the BCApps cleanup idiom of ending a test with `asserterror Error(SomeLabel)` to undo demo-data mutations that method made, so the run doesn't permanently dirty the database. Per the documented `Codeunit.Run` transaction semantics, changes are committed at the end of an execution "unless an error occurs" — an unhandled error rolls back whatever wasn't already committed. `Commit()` closes out the fixture's own transaction immediately, so it is unaffected by whatever the rest of that method does afterward, including that end-of-test error. Without the early `Commit()`, the same deliberate rollback wipes out the fixture too, even though `IsInitialized` still reads `true` on the next test, since it's a plain variable, not persisted data. BCApps' `codeunit 134915 "ERM Online Mapping Setup"` shows exactly this shape: no `TransactionModel` attribute, `Commit()` inside a lazy `Initialize()`, and the test itself ends with `asserterror Error(RollBackMessage)`. + +Protecting the fixture from that same-method rollback is necessary but not sufficient for the fixture to reach a *later* test method — that also depends on the executing test runner's `TestIsolation`. Under `Disabled` (the property's own documented default) or `Codeunit` (used by BCApps' own `TestRunner`, `CLITestRunner`, and `SnapTestRunner` codeunits), nothing rolls back until the whole test codeunit finishes, so the already-committed fixture survives across every method run before then. Under `Function`, the runner rolls back all database changes — explicitly including ones already committed via `Commit()` — after every single test method; no amount of committing inside `Initialize()` makes a fixture shared across methods survive that regime, because the whole premise of a lazy, once-per-codeunit fixture doesn't hold when every method is isolated from every other. ## Best Practice -Call `Commit()` at the end of a lazy/shared `Initialize()` procedure, once the shared fixture data is created, so it survives past the first test's rollback boundary. Pair this with a `TestIsolation`-enabled test runner so the committed fixture is still cleaned up at the end of the full run. +When a test method's own cleanup relies on ending in a deliberate error to roll back its scratch changes, call `Commit()` once, inside the lazy `Initialize()` guard, right after the shared fixture is created — before that cleanup-triggering error can run. This pattern only delivers a fixture shared across test methods when the executing runner's `TestIsolation` is `Disabled` or `Codeunit`; do not recommend it, or pair it with, a `Function`-isolated runner — that configuration undoes the committed fixture after every method regardless. See sample: `commit-shared-test-fixture-inside-lazy-initialize.good.al`. ## Anti Pattern -A shared `Initialize()` guarded by `IsInitialized` that creates fixture records but never commits. The first test that runs it passes; every later test in the same codeunit either fails to find the fixture data or silently re-triggers setup logic that `IsInitialized` was meant to skip. +A shared `Initialize()` guarded by `IsInitialized` that creates fixture records without committing, in a test method that ends with a deliberate `asserterror Error(...)` to undo its own scratch changes, run under a `Disabled`- or `Codeunit`-isolated test runner. That rollback also erases the never-committed fixture; the next test still finds `IsInitialized = true` but the rows it depends on are gone. (Under a `Function`-isolated runner the fixture is lost regardless of `Commit()`, for the unrelated reason above — that is a runner-configuration problem, not this anti-pattern.) See sample: `commit-shared-test-fixture-inside-lazy-initialize.bad.al`. + +## Source + +The shared/lazy `Initialize()` pattern and its `Commit()` call are drawn from Luc van Vugt's "Let's talk about Shared Fixture and how to profit from this with the Dynamics NAV Test Toolkit": https://www.fluxxus.nl/index.php/bc/let39s-talk-about-shared-fixture-and-how-to-profit-from-this-with-the-dynamics-nav-test-toolkit/. That post shows the `Commit()` call in its `Initialize()` example but does not explain the transaction mechanics behind it; the `AutoCommit`-default, `Codeunit.Run`-error, and `TestIsolation`-level analysis above is this article's own, verified independently against Microsoft's TransactionModel/TestIsolation documentation and BCApps' `codeunit 134915 "ERM Online Mapping Setup"` source, not taken from the post. diff --git a/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.bad.al b/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.bad.al deleted file mode 100644 index acb825f..0000000 --- a/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.bad.al +++ /dev/null @@ -1,9 +0,0 @@ -codeunit 50140 "Sample Confirm Usage" -{ - procedure ConfirmDeletion(RecordCount: Integer): Boolean - var - ConfirmMsg: Label 'Do you want to delete %1 records?'; - begin - exit(Confirm(ConfirmMsg, false, RecordCount)); - end; -} diff --git a/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.good.al b/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.good.al deleted file mode 100644 index 3a529fb..0000000 --- a/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.good.al +++ /dev/null @@ -1,9 +0,0 @@ -codeunit 50140 "Sample Confirm Usage" -{ - procedure ConfirmDeletion(RecordCount: Integer): Boolean - var - ConfirmMsg: Label 'Do you want to delete %1 records?'; - begin - exit(Confirm(StrSubstNo(ConfirmMsg, RecordCount), false)); - end; -} diff --git a/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.md b/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.md deleted file mode 100644 index f92a783..0000000 --- a/microsoft/knowledge/testing/confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.md +++ /dev/null @@ -1,30 +0,0 @@ ---- -bc-version: [all] -domain: testing -keywords: [confirm, confirmhandler, strsubstno, placeholder, question, substitution] -technologies: [al] -countries: [w1] -application-area: [all] ---- - -# Build the Confirm message with StrSubstNo, or a ConfirmHandler sees the raw template - -## Description - -`Confirm`'s placeholder-substitution overload — `Confirm('text %1', false, Value)` — substitutes the placeholder only for the dialog a real user sees. Inside a `[ConfirmHandler]`, the `Question` parameter received is the literal, unsubstituted template string (`'text %1'`), not the value-filled text. This is a reported platform defect (microsoft/ALAppExtensions#23935), not documented or intended behavior — treat it as a known, unconfirmed-fix issue rather than a permanent platform rule; if it is ever fixed, this workaround becomes unnecessary rather than wrong. Notably, `Message`/`[MessageHandler]` substitutes correctly — only `Confirm` is affected, which is itself evidence this is a bug rather than a deliberate design choice. A test that asserts `Question` against the expected substituted message either fails outright or silently checks the wrong thing. - -## Best Practice - -When a `ConfirmHandler` needs to assert on the actual message text, build the string with `StrSubstNo(Text, Value)` in the production code first, and pass the already-substituted string to `Confirm()` with no further placeholder arguments. - -See sample: `confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.good.al`. - -## Anti Pattern - -Calling `Confirm('text %1', false, Value)` and then asserting the substituted text against `Question` inside a `[ConfirmHandler]`. `Question` holds the raw `'text %1'` template, so the assertion never matches the intended message. - -See sample: `confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text.bad.al`. - -## Source - -Reported by Luc van Vugt: https://github.com/microsoft/ALAppExtensions/issues/23935 — an internal Microsoft bug was filed from that report; the issue's fix status is not confirmed as of this writing. diff --git a/microsoft/knowledge/testing/table-relation-test-exclude-known-invalid-relations-via-event.md b/microsoft/knowledge/testing/table-relation-test-exclude-known-invalid-relations-via-event.md index a3af2be..f680f1a 100644 --- a/microsoft/knowledge/testing/table-relation-test-exclude-known-invalid-relations-via-event.md +++ b/microsoft/knowledge/testing/table-relation-test-exclude-known-invalid-relations-via-event.md @@ -11,11 +11,11 @@ application-area: [all] ## Description -Codeunit 134926 "Table Relation Test" (shipped in BCApps' test app — only consumers that depend on the BC test libraries can subscribe to it) reads Table Relations Metadata tenant-wide across every installed app, not just the current one, and fails the moment a related field's type or length doesn't match what the relation requires — the related field must match the largest related field's length, and its type must match (except a field may relate to both `Code` and `Text`, which resolves to `Text`). A field with a legitimate, intentional relation shape has no per-field override in its own object definition; the check runs with no built-in escape hatch. +Codeunit 134926 "Table Relation Test" (shipped in BCApps' test app — only consumers that depend on the BC test libraries can subscribe to it) reads Table Relations Metadata tenant-wide across every installed app, not just the current one, and fails the moment a related field's type or length doesn't match what the relation requires — the related field must match the largest related field's length, and its type must match (except a field may relate to both `Code` and `Text`, which resolves to `Text`). A field with a legitimate, intentional relation shape has no per-field override in its own object definition; the check runs with no built-in escape hatch. The validation test method itself is `[Scope('OnPrem')]`: it only runs from an on-premises test surface, not from a cloud-targeted test app, so this whole exception mechanism — and the check it works around — is only reachable where that test can actually execute. ## Best Practice -Subscribe to `OnAfterRemoveTableRelation` and call the codeunit's own `RemoveTableRelation(TableRelationsMetadata, TableID, FieldID, RelatedTableID, RelatedFieldID)` to strike the one known-valid relation before the test evaluates it, scoped as narrowly as the exception actually is. +Subscribe to `OnAfterRemoveTableRelation` and call the codeunit's own `RemoveTableRelation(TableRelationsMetadata, TableID, FieldID, RelatedTableID, RelatedFieldID)` to strike the one known-valid relation before the test evaluates it, scoped as narrowly as the exception actually is. Because the test itself is `[Scope('OnPrem')]`, do not recommend subscribing to it as a way to guard a cloud-targeted app's test suite — the subscription has no effect where the test never runs. See sample: `table-relation-test-exclude-known-invalid-relations-via-event.good.al`. @@ -24,3 +24,7 @@ See sample: `table-relation-test-exclude-known-invalid-relations-via-event.good. Excluding an entire table's relations (or disabling the whole test codeunit) to work around one known exception. This discards the check's coverage for every other relation on that table, or in the app, not just the one that needed an exception. See sample: `table-relation-test-exclude-known-invalid-relations-via-event.bad.al`. + +## Source + +The `OnAfterRemoveTableRelation` exclusion technique is drawn from Luc van Vugt's "How-to: Test your Table Relations (2)": https://www.fluxxus.nl/index.php/bc/how-to-test-your-table-relations-2/. The codeunit/event signature, the `[Scope('OnPrem')]` boundary, and the tenant-wide `Table Relations Metadata` scope described above were verified directly against BCApps' `codeunit 134926 "Table Relation Test"` source, not taken from the post. diff --git a/microsoft/knowledge/testing/transactionmodel-attribute-governs-test-transactions.md b/microsoft/knowledge/testing/transactionmodel-attribute-governs-test-transactions.md index ae24f1f..0f4f3bd 100644 --- a/microsoft/knowledge/testing/transactionmodel-attribute-governs-test-transactions.md +++ b/microsoft/knowledge/testing/transactionmodel-attribute-governs-test-transactions.md @@ -11,11 +11,11 @@ application-area: [all] ## Description -`[TransactionModel(...)]` declares how a test method interacts with the database's write transaction. The attribute applies only to methods inside a codeunit with `SubType = Test` and takes one of three values: `AutoRollback`, `AutoCommit`, or `None`. The "a call to `Commit` produces a runtime error" behavior is specific to the *explicitly declared* `AutoRollback` attribute — it is not what an undeclared/default test method does. BCApps' own canonical pattern for a lazily-initialized shared fixture (see `codeunit 134915 "ERM Online Mapping Setup"`) declares no `TransactionModel` attribute at all, calls `Commit()` inside its `Initialize()` helper, and cleans up manually with a deliberate `asserterror Error(...)` at the end rather than relying on automatic rollback — this is a legitimate, common pattern, not a bug. When a test method *does* declare `AutoRollback` explicitly, the choice must match the code being exercised: per the platform reference, "if the code that you test includes calls to the COMMIT Method, then set the TransactionModel property on the test method to AutoCommit." Applying `AutoRollback` to a test that drives code which calls `Commit` produces a runtime error on the first Commit, not a meaningful assertion failure. +`[TransactionModel(...)]` declares how a test method interacts with the database's write transaction. The attribute applies only to methods inside a codeunit with `SubType = Test` and takes one of three values: `AutoRollback`, `AutoCommit`, or `None`. **`AutoCommit` is the documented default** — a test method with no `[TransactionModel(...)]` attribute at all runs under `AutoCommit`, not `AutoRollback` and not `None` (Microsoft's TransactionModel property reference states this explicitly: "AutoCommit is the default value"). The "a call to `Commit` produces a runtime error" behavior is specific to the *explicitly declared* `AutoRollback` attribute. BCApps' own canonical pattern for a lazily-initialized shared fixture (see `codeunit 134915 "ERM Online Mapping Setup"`) declares no `TransactionModel` attribute at all — so it runs under the `AutoCommit` default — calls `Commit()` inside its `Initialize()` helper, and cleans up manually with a deliberate `asserterror Error(...)` at the end rather than relying on automatic rollback; this is a legitimate, common pattern, not a bug. Per the same reference, under `AutoCommit` an error, even one caught by `asserterror`, still rolls back the transaction — but "only to the point at which `Commit` was called" if the code being tested committed first. When a test method *does* declare `AutoRollback` explicitly, the choice must match the code being exercised: per the platform reference, "if the code that you test includes calls to the COMMIT Method, then set the TransactionModel property on the test method to AutoCommit." Applying `AutoRollback` to a test that drives code which calls `Commit` produces a runtime error on the first Commit, not a meaningful assertion failure. ## Best Practice -When declaring `[TransactionModel(...)]` explicitly, pick `AutoRollback` for a test whose own logic and the code it exercises make no `Commit` call, `AutoCommit` when the code under test genuinely calls `Commit` — posting routines, job-queue handlers, integration flows — and make the test exercise that commit path, and `None` for a read-only test or one that drives UI code without writing from the test method itself. Pair `AutoCommit` with a `TestIsolation`-enabled test runner so committed changes are reverted at a higher scope. A lazily-initialized shared fixture that commits once and relies on a manual `asserterror`-based cleanup, with no `TransactionModel` attribute declared at all, is a distinct and equally valid pattern — do not treat the absence of the attribute as equivalent to declaring `AutoRollback`. +Leave `[TransactionModel(...)]` undeclared to get the `AutoCommit` default when the codeunit's own tests rely on that default's behavior — for example a lazily-initialized shared fixture that commits once and cleans up its own scratch changes with a manual `asserterror`-based rollback (see `commit-shared-test-fixture-inside-lazy-initialize.md`); do not treat that absence as equivalent to declaring `AutoRollback`. When declaring `[TransactionModel(...)]` explicitly instead, pick `AutoRollback` for a test whose own logic and the code it exercises make no `Commit` call, `AutoCommit` when the code under test genuinely calls `Commit` — posting routines, job-queue handlers, integration flows — and make the test exercise that commit path, and `None` for a read-only test or one that drives UI code without writing from the test method itself. Pair an intentional, suite-wide reliance on `AutoCommit` with a `TestIsolation`-enabled test runner so committed changes are reverted at a higher scope. See sample: [`transactionmodel-attribute-governs-test-transactions.good.al`](transactionmodel-attribute-governs-test-transactions.good.al). @@ -24,3 +24,7 @@ See sample: [`transactionmodel-attribute-governs-test-transactions.good.al`](tra Declaring `[TransactionModel(AutoRollback)]` explicitly on a test method without checking whether the tested business logic calls `Commit`. The test throws at the first Commit, leaving no verdict on the behavior it intended to verify; in a CI run this looks like a flake or a setup bug, not a specification mismatch. The mirror-image anti-pattern is defaulting to `AutoCommit` across the suite "to avoid the error" — without a `TestIsolation` runner this permanently dirties the test database between runs and produces order-dependent test outcomes. Flagging a `Commit()` call in a test method that declares no `TransactionModel` attribute at all is not this anti-pattern — that shape does not error, and is BCApps' own documented pattern for shared lazy fixtures. See sample: [`transactionmodel-attribute-governs-test-transactions.bad.al`](transactionmodel-attribute-governs-test-transactions.bad.al). + +## Source + +The `AutoCommit`-is-default claim and the exact rollback-to-last-`Commit` mechanics are quoted from Microsoft's TransactionModel Property reference: https://learn.microsoft.com/en-us/previous-versions/dynamicsnav-2018-developer/TransactionModel-Property. The current AL [TransactionModel attribute](https://learn.microsoft.com/dynamics365/business-central/dev-itpro/developer/attributes/devenv-transactionmodel-attribute) page describes the same three values but never states a default; this older property reference is the citable source for that fact. diff --git a/microsoft/knowledge/testing/use-assert-isfalse-not-asserterror-for-boolean-checks.md b/microsoft/knowledge/testing/use-assert-isfalse-not-asserterror-for-boolean-checks.md index 926e434..df1f791 100644 --- a/microsoft/knowledge/testing/use-assert-isfalse-not-asserterror-for-boolean-checks.md +++ b/microsoft/knowledge/testing/use-assert-isfalse-not-asserterror-for-boolean-checks.md @@ -24,3 +24,11 @@ See sample: `use-assert-isfalse-not-asserterror-for-boolean-checks.good.al`. `asserterror Assert.IsTrue(SomeFunc(), Msg);` to verify `SomeFunc()` is `false`. It passes today because `Assert.IsTrue` happens to raise an error on failure, but it verifies the assertion helper's error-raising behavior, not the value under test. See sample: `use-assert-isfalse-not-asserterror-for-boolean-checks.bad.al`. + +## Source + +Drawn from Luc van Vugt's "TDD in NAV – ASSERTERROR or IsFalse": https://www.fluxxus.nl/index.php/bc/tdd-in-nav-asserterror-or-isfalse/. The post's own example and reasoning — reserve `asserterror` for the product code actually raising an error, use `Assert.IsFalse`/`Assert.IsTrue` to check a boolean the test framework itself computes — carries over directly; the overlap with `asserterror-needs-expectederror-and-code.md` below is this repository's own addition, not from the source. + +## Scope + +This rule and `asserterror-needs-expectederror-and-code.md` can both match `asserterror Assert.IsTrue(SomeFunc(), Msg);` with nothing after it — the generic rule sees a bare `asserterror`, this one sees `asserterror` wrapping an `Assert.IsTrue`/`Assert.IsFalse` call used to invert a boolean. This rule wins for that shape: the fix is to replace the construct with a direct `Assert.IsFalse`/`Assert.IsTrue` call, not to add `Assert.ExpectedError`/`Assert.ExpectedErrorCode` after it. `asserterror-needs-expectederror-and-code.md` still applies on its own to every other bare `asserterror`, including one guarding `Assert.IsTrue`/`Assert.IsFalse` where the intent genuinely is to assert that the guarded call itself raises an error (for example, asserting that a validation helper errors before it can even return a boolean). diff --git a/microsoft/skills/review/al-testing-review.md b/microsoft/skills/review/al-testing-review.md index 33016f1..d7087d0 100644 --- a/microsoft/skills/review/al-testing-review.md +++ b/microsoft/skills/review/al-testing-review.md @@ -49,9 +49,9 @@ The following targeted checks cover every current `testing` article. Treat each - An `AutoCommit` test runs under a `Subtype = TestRunner` codeunit that omits `TestIsolation` or sets it to `Disabled`, leaving committed data between tests — `testisolation-belongs-on-the-test-runner`. Require runner/repository context; a standalone test file cannot prove which runner executes it. - A permission-sensitive test uses `TestPermissions = Disabled`, claims to test a restricted user without `"Permissions Mock"`/`"Library - Lower Permissions"`, or declares `[TestPermissions(...)]` without applying that context — `permission-tests-must-lower-the-execution-context`. - Test fixture code manually calls `Init`/`Insert`, invents keys or prerequisite records, or bypasses available `LibrarySales`, `LibraryPurchase`, `LibraryERM`, `LibraryInventory`, `LibraryRandom`, or equivalent library codeunits — `use-library-codeunits-for-test-fixtures`. -- `asserterror` is added or changed without a following `Assert.ExpectedError`, `Assert.ExpectedErrorCode`, or a purpose-built assertion such as `ExpectedTestFieldError` — `asserterror-needs-expectederror-and-code`. -- A `[ConfirmHandler]` asserts `Question` against a substituted message from a `Confirm` call using its placeholder-substitution overload — `confirm-needs-strsubstno-before-confirmhandler-sees-substituted-text`. -- A shared/lazy `Initialize()`-style fixture helper calls `Commit()` — `commit-shared-test-fixture-inside-lazy-initialize`. +- `asserterror` is added or changed without a following `Assert.ExpectedError`, `Assert.ExpectedErrorCode`, or a purpose-built assertion such as `ExpectedTestFieldError` — `asserterror-needs-expectederror-and-code`. Exclude `asserterror Assert.IsTrue(...)` / `asserterror Assert.IsFalse(...)` guarding a `Boolean`-returning call — that shape belongs to `use-assert-isfalse-not-asserterror-for-boolean-checks` instead, which wins for it. +- `asserterror` wraps `Assert.IsTrue(BooleanExpression, ...)` (or the `IsFalse` mirror) solely to invert the boolean result of the guarded call, rather than to assert that call itself raises an error — `use-assert-isfalse-not-asserterror-for-boolean-checks`. +- A shared/lazy `Initialize()`-style fixture helper creates fixture data without a following `Commit()`, in a test method whose body later forces its own rollback (for example `asserterror Error(...)` used for end-of-test cleanup) — `commit-shared-test-fixture-inside-lazy-initialize`. The presence of `Commit()` after the fixture is the compliant shape, not the signal to look for; the missing-`Commit()` shape combined with a later deliberate rollback is the anti-pattern. Require runner/repository context for the `TestIsolation` value: a standalone test file cannot prove which runner executes it, and under `Function`-level isolation this whole pattern is moot regardless of `Commit()` — do not raise the finding when the executing runner's `TestIsolation` is known to be `Function`. - Changed code subscribes to `OnAfterRemoveTableRelation`, calls `RemoveTableRelation`, or references `Codeunit "Table Relation Test"`/134926 — `table-relation-test-exclude-known-invalid-relations-via-event`. - A test path raises UI and `[HandlerFunctions(...)]` does not match the invoked handlers, or the test has no meaningful evidence of the UI result (for example, it treats a Boolean set before the action as proof of success) — `ui-handlers-in-tests`. A capture/reset/assert-after-`RunModal` pattern is valid. Enqueue/dequeue and `AssertEmpty` are required only when order, count, text, replies, or a scripted sequence is part of the contract. Only nonoptional handlers have to execute: a listed handler declared `[SendNotificationHandler(true)]` or `[RecallNotificationHandler(true)]` is optional by design, so do not treat it as unmatched when the run never raises the notification.