feat: (W-004) Engine adapter modules for Lucee, Adobe, and BoxLang#2016
Open
feat: (W-004) Engine adapter modules for Lucee, Adobe, and BoxLang#2016
Conversation
The Wheels framework stores its datasource in application.wheels.dataSourceName, not as the CFML default this.datasource. All queryExecute calls in Job.cfc (15), JobWorker.cfc (14), and JobWorkerSpec.cfc (7) now pass the datasource explicitly. Also replaces MySQL-specific LIMIT :limit with database-agnostic maxrows option in processQueue(). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
queryExecute for DML returns a query with recordCount=0 (result set size, not rows affected). Use the result option to capture DML metadata where the actual affected row count lives. Fixes $claimJob always returning false, which caused processNext to skip all jobs even after successful enqueue. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Two root causes for remaining 20/42 CI failures: 1. $ensureJobTable CREATE TABLE used DATETIME (invalid in PostgreSQL/Oracle) and TEXT (invalid in Oracle). Now detects application.wheels.adapterName and uses TIMESTAMP/CLOB/VARCHAR2 for Oracle, TIMESTAMP for PostgreSQL. 2. $claimJob used queryExecute result option for DML recordCount, which is unreliable across CFML engine + JDBC driver combinations. Replaced with a SELECT verification query -- SELECT recordCount is universally reliable. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Reorder DEFAULT before NOT NULL in CREATE TABLE (Oracle requires this) - Wrap $claimJob UPDATE+SELECT in transaction block to ensure same connection Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…ource MySQL/H2 DATETIME columns round fractional seconds >= 0.5 to the next second. When enqueue() stores runAt=Now() and processNext() checks runAt <= Now() milliseconds later, the rounded-up value appears in the future, making the job invisible. Fix: $now() helper truncates to whole seconds. Also replace application.wheels.adapterName with $detectDatabaseType() using cfdbinfo on the actual datasource — the adapter may have been detected from a different datasource in CI. Add H2-specific TIMESTAMP type in $ensureJobTable(). Test now asserts enqueue persisted before checking processNext result. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…oxLang claim
Oracle requires unquoted identifiers to start with a letter, making _wheels_jobs
invalid. Renamed to wheels_jobs across all SQL, tests, and docs.
Removed transaction {} from $claimJob — auto-commit is sufficient and avoids
failures on BoxLang + PostgreSQL.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…tion SELECT The separate SELECT could hit a different pooled connection that cannot see the uncommitted UPDATE, causing the claim to always fail on BoxLang+Postgres. Using the result option returns the affected-row count from the same connection. Also add a single retry in the test for engines with deferred cross-connection visibility after INSERT. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Direct SELECT verifies INSERT visibility before calling processNext. Custom assertion messages will reveal whether the issue is at the SELECT visibility level or the claim level. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Tests the exact SQL processNext uses and inspects the result struct keys from the claim UPDATE to determine where BoxLang+Postgres fails. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The candidate query is THROWING on BoxLang+Postgres (not returning 0 rows). Added fallback queries to isolate whether the issue is parameter binding, the runAt comparison, or something else. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
A = full query + maxrows, B = without maxrows, C = without IN clause, D = without timestamp param. This will reveal which SQL feature breaks on BoxLang+Postgres. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
BoxLang throws when setMaxRows() is called on a PostgreSQL JDBC PreparedStatement with certain parameter combinations. The for-loop already exits on the first successful claim, so maxrows is unnecessary. Also cleans up diagnostic code from the test. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add job worker daemon with CLI commands (#1913)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…n tests The WHERE path had a bug: line 754 replaced the condition with a ? placeholder before the column lookup, so when the column wasn't found and we logged instead of throwing, the orphaned ? caused a params/where array mismatch downstream. Fix: undo the ? replacement before continuing, letting the raw column name pass through to the database. Also strengthen tests to actually include invalid columns (the original tests only used valid columns, proving nothing about the setting). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Wheels caches SQL shells by hashing model name + arguments. The "still throws" test used the same select string as the preceding "skips invalid" test, so it hit the cached SQL (built without the invalid column) and never reached the column validation code. The throws-when-true behavior is already covered by the existing "throws error for invalid select column" test which uses a different select string and runs before any test disables the setting. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add throwOnColumnNotFound config setting
Plan to restructure the GitHub Actions test matrix from 42 independent jobs (cfengine x dbengine) to 8 engine-grouped jobs. Each CF engine starts once and runs all database test suites sequentially, reducing total CI compute by ~75%. Key changes planned: - Engine-only matrix (8 jobs vs 42) - CFPM install once per Adobe engine instead of 6x - Oracle sleep 120 replaced with sqlplus health check - SQL Server memory 4GB->2GB, Oracle 2GB->1.5GB - JUnit XML reporting via JSON-to-XML conversion - Engine x database summary grid on workflow runs - BoxLang volume mounts re-enabled for local dev parity Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The test dataset is tiny (5 users, 8 authors, 40 photos). 4GB was excessive and prevents running all databases simultaneously on a 7GB GitHub runner. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Allows running all databases simultaneously within 7GB GitHub runner budget. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Volumes were commented out in 9790880 when Dockerfile was changed to COPY all code at build time. This broke local dev (code changes required rebuild). Re-enabling volumes restores live-reload behavior matching other CF engines. Docker gives volumes precedence over COPY, so CI still works correctly. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replaces cfengine x dbengine matrix (42 jobs) with cfengine-only matrix (8 jobs). Each job starts one CF engine + all databases, runs test suites sequentially. Key changes: - CF engine starts once per job instead of 6x (saves 5 startups per engine) - CFPM install runs once per Adobe engine instead of 6x (saves ~50 min compute) - Oracle sleep 120 replaced with sqlplus health check loop - All databases start in parallel during engine warmup - Tests continue after individual database failures (all DBs always run) - JUnit XML output uploaded for each engine-database pair - Summary jobs render engine x database grid on workflow run page - PR annotations via EnricoMi/publish-unit-test-result-action Total compute reduction: ~75% (from ~840 min to ~200 min) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Required by EnricoMi/publish-unit-test-result-action to post test result annotations and PR comments. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…rect Wheels returns HTTP 302 after reload=true completes. The test curl was capturing this 302 as a failure instead of following the redirect. Fix: issue the reload as a separate curl with -L (follow redirects), then run the actual test request without reload. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The previous fix split reload and test into two requests. But the reload request without format=json would run the full test suite in HTML format, then the second request would run it again — doubling execution time. Better approach: include reload=true AND format=json in one curl with -L. Wheels' redirectAfterReload strips reload/password but preserves db= and format=, so the redirect lands on ?db=X&format=json — exactly what we need. Single request, single test execution. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The ?db= parameter switches application.wheels.dataSourceName directly (runner.cfm:70-80). No reload needed because: - $_setTestboxEnv() handles datasource switching per HTTP request - populate.cfm creates identical table schemas across all databases - The old 42-job workflow never used reload=true either reload=true was breaking Adobe CF engines because $handleRestartAppRequest passes "application" as a component reference to cfinvoke, which Adobe CF can't resolve. Removing it fixes all Adobe engines and simplifies the curl. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
v5 uses Node.js 20 which is deprecated and will be forced to Node.js 24 on June 2, 2026. v6 natively supports Node.js 24. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Action version bumps (Node.js 24 support): - actions/checkout v4 → v5 - actions/upload-artifact v4 → v5 - actions/download-artifact v4 → v5 Restart CF engine container between database runs to prevent cross-DB contamination. Without restart, cached model metadata and association methods from one database's test run leak into subsequent runs: - BoxLang: shallow copy in runner.cfm causes datasource name to not actually switch (duplicate key errors on 2nd+ database) - Adobe CF: stale association methods cause "method not found" errors on later databases (e.g. hasAuthor missing on 4th DB) Container restart takes ~10-15s — much cheaper than a full rebuild, and guarantees clean application state per database. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Fixes remaining Node.js 20 deprecation warnings in CI. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…ile handle leak, timestamp
- Change $sentryGetUser() from public to private ($ prefix = internal)
- Wrap initSentry() body in named exclusive lock to prevent double-init races
- Fix HTTP status check: Left(statuscode,3) != "200" instead of find("200",...)
- Fix file handle leak: replace fileOpen/fileReadLine/fileClose with FileRead+ListToArray in try/catch
- Add Z suffix to ISO 8601 timestamp in getTimeVars()
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Remove 'culprit' key from exception event payload (deprecated in SDK v7) - Change exception type from 'exType & " Error"' to just 'exType' to avoid double-labeling (e.g., "Application Error Error") Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Renders JSON path configuration for Hotwire Native apps. Accepts optional settings struct (tabs, etc.) and rules array. Falls back to sensible defaults: all pages use default context with pull-to-refresh; /new$ and /edit$ paths open as modals without pull-to-refresh. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…mp, user context, breadcrumbs Covers: - DSN parsing via public API (valid modern, valid legacy, invalid throws) - ISO 8601 Z-suffix timestamp format - sentrySetUser stores and overwrites request.sentryUserOverride - addBreadcrumb appends in order and includes optional data struct - getEnvironment/getRelease accessors return init values Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Covers all public helpers: hotwireIncludes, turboFrame/turboFrameEnd, all 8 turbo stream actions (verifying template presence/absence on remove and refresh), and all 4 Stimulus attribute helpers. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…E.md - Add plugin.json with correct metadata (mixins: controller,view, wheelsVersion: 3.0) - Fix Basecoat.cfc mixin attribute: remove stale dispatch and microsofttablehelper scopes - Replace Hotwire-content ARCHITECTURE.md with Basecoat-specific architecture doc - Add Phase 2 dialog family (uiDialog, uiDialogFooter, uiDialogEnd) with native <dialog>, ARIA attributes, optional trigger button - Add Phase 3 uiField with full type support: text/email/password/number/tel/url/textarea/select/checkbox/switch; error state; description; auto-generated IDs - Add Phase 4: uiTable family, uiTabs family, uiDropdown (CSS-only details/summary), uiPagination with ellipsis/window logic - Add Phase 5: uiBreadcrumb family, uiSidebar family Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- BasecoatSimpleSpec: Phase 1+2 — button class construction (compound variants, icon-only, anchor, turboConfirm, disabled), badge variants, alert (title/description/destructive), card family, progress, spinner, skeleton, tooltip, separator - BasecoatComplexSpec: Phase 2-5 — dialog ARIA attributes (labelledby/describedby, ID matching, trigger/no-trigger, close button), field layout (label-above vs flex-row for checkbox/switch, error state, description, auto-IDs), table wrappers, tabs data-attributes, dropdown details/summary, pagination (disabled prev/next, window, current page), breadcrumb (aria-label, linked/current), sidebar (active state, icons, section heading) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…wheels-basecoat Rename SentryForWheels/ folder to sentry/, rename SentryForWheels.cfc to Sentry.cfc, update all internal component paths, plugin.json names, box.json slugs, and documentation references. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace the legacy plugins/ directory with a new packages/ → vendor/ activation model. PackageLoader discovers vendor/*/package.json on startup with per-package error isolation, collects mixins into the existing injection pipeline, and supports ServiceProvider and middleware registration. - Add vendor/wheels/PackageLoader.cfc for package discovery and loading - Move sentry, hotwire, basecoat from plugins/ to packages/ - Replace plugin.json with package.json (adds provides.mixins/services/middleware) - Wire $loadPackages() into onapplicationstart.cfc and EventMethods.cfc - Add deprecation warning when plugins/ directory has loaded plugins - Tested: Lucee 6 + Lucee 7 × H2 = 2268 pass, 0 fail, 0 error Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add "Run per-package tests" step to CI that copies each package to vendor/, restarts the engine, runs the package's test specs, then cleans up. Uses h2 for Lucee, sqlite for others. Non-blocking warnings. - Update prepare-base.sh to include packages/ in release artifact. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
SQLite works across all engines (Lucee, Adobe, BoxLang) unlike H2 which is Lucee-only. Standardize on sqlite as the embedded DB for package tests. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add PackageLoaderSpec.cfc with 10 BDD specs covering discovery, error isolation, manifest parsing, and mixin collection - Add test fixtures in tests/_assets/packages/ (goodpkg, brokenpkg, nomanifest, nomixin) - Add componentPrefix param to PackageLoader.init() for test flexibility - Create packages/sentry/CLAUDE.md - Update packages/hotwire/CLAUDE.md and packages/basecoat/CLAUDE.md to reflect package architecture (s/Plugin/Package/) - Add Package System section to main CLAUDE.md with activation docs Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…-basecoat feat: add package system (PackageLoader) and first-party packages
TestBox JSON can have `"failMessage": null` for errored specs (seen on BoxLang + CockroachDB). Python's dict.get() returns None when the key exists with a null value, and ElementTree cannot serialize None as an XML attribute — crashing tostring() after the file was already opened, leaving a truncated 38-byte file that broke publish-unit-test-result-action. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ntax
Converts Hotwire.cfc, Basecoat.cfc, and both Basecoat test specs from
cffunction/cfargument tag syntax to CFScript (component { }, typed params,
savecontent blocks). Matches the style already used in wheels-sentry.
Updates CLAUDE.md coding convention sections accordingly.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…s/docker/testui/picomatch-4.0.4 chore(deps): bump picomatch from 4.0.2 to 4.0.4 in /tools/docker/testui
…-fail CockroachDB was marked as soft-fail in CI because tests failed. Three root causes fixed: 1. Missing CockroachDB in adapter name checks — update.cfc and sql.cfc had hardcoded adapter lists (PostgreSQL,H2,...) that excluded CockroachDB, causing UPDATE SET clauses to generate invalid SQL. 2. Identity select not passing RETURNING result — $executeQuery now passes the cfquery result as returningIdentity to $identitySelect, fixing generated key retrieval on Adobe CF. 3. Boolean type mapping — changed cf_sql_boolean to cf_sql_bit for Adobe CF compatibility (ACF doesn't support cf_sql_boolean in cfqueryparam). Also adds CockroachDB-specific test specs (CRUD, transactions, types) and guards core tests that assume sequential IDs or PostgreSQL transaction semantics. Closes #1972, closes #1974 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ase2 fix: CockroachDB adapter phase 2 — fix SQL generation and remove soft-fail
BoxLang + CockroachDB has pre-existing issues with Long/Integer casting in index metadata (dbinfo, addIndex, removeIndex) and afterFind callbacks. Guard these tests and fix toBeInstanceOf assertion in integration spec. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PR workflow now runs only Lucee 7 + SQLite for fast CI feedback (~3-5 min). Full test suite (5 engines × 7 databases) available via manual workflow_dispatch. - pr.yml: commitlint + label + quick-test (replaces full matrix) - full-tests.yml: manual trigger for complete engine/database matrix - tests.yml: unchanged reusable workflow
…bug-panel-needs-to-be-up feat: (W-001) The developer debug panel needs to be updated. It...
The error page header (_header_simple.cfm) only rendered Logo, Info, Routes, and Migrator tabs. Added the missing API, Guides, and Tests tabs, plus conditional Packages and Plugins tabs matching the normal header. Also made Migrator conditional to match _header.cfm behavior. Closes #1998 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…-debug-panel-breaks-when feat: (W-002) Navbar in Wheels debug panel breaks when controller action...
…sequences (#2006) Lucee's Canonicalize() delegates to Java's URLDecoder.decode(), which throws java.lang.IllegalArgumentException when the input contains %% or any lone % not followed by two hex digits. This crashes form rendering whenever CFWheels re-populates a field (e.g. after validation failure) with a value containing such characters. Passwords are a common real-world trigger: a user entering P%%ssword or 100%%secure will always hit this crash on form re-render. Wrap the Canonicalize() call in try/catch and fall back to the raw input on failure. The raw value is safe because the caller (miscellaneous.cfc:457) always passes it through EncodeForHTMLAttribute() immediately after. The existing null-check guard for Lucee 5 is preserved inside the try block.
Fork PRs receive a read-only GITHUB_TOKEN, so the labeler action cannot write labels and fails with "Resource not accessible by integration". Add continue-on-error so this cosmetic job doesn't block CI. Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ng adapters Centralizes scattered engine-specific checks (StructKeyExists(server, ...)) into a polymorphic adapter pattern, modeled on the existing database adapter hierarchy. Auto-detected at startup and stored in application.wheels.engineAdapter. Phase 1 migrates 6 high-value methods across 4 files: - getResponse(), getStatusCode(), getContentType() — Global.cfc, sse.cfc - getRequestTimeout() — Global.cfc - parseFormKey() — Dispatch.cfc - controllerNameToUpperCamelCase() — Dispatch.cfc New files: - engineAdapters/EngineAdapterInterface.cfc (documented contract) - engineAdapters/Base.cfc (abstract base, Lucee-compatible defaults) - engineAdapters/Lucee/LuceeAdapter.cfc - engineAdapters/Adobe/AdobeAdapter.cfc - engineAdapters/BoxLang/BoxLangAdapter.cfc - tests/specs/engineAdapterSpec.cfc (17 BDD tests) Tested: Lucee 6 (2295 pass/0 fail), Adobe 2025 (2304 pass/0 fail) Closes #1968 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…iew, and dispatch layers Add 16 new adapter methods (identity helpers, Oracle JDBC coercion, dynamic finders, hash normalization, struct defaults, numeric validation, DI completion, method invocation, image/zip handling, glob patterns, query args, port detection, date parsing, image formats) and migrate ~35 engine-specific checks across 23 files to use the centralized adapter instead of scattered StructKeyExists(server) checks. Key migrations: - Model layer: onmissingmethod, validations, miscellaneous, properties, sql, associations, callbacks, create, read - Global.cfc: $image, $zip, $hashedKey, $args, $convertToString, $dbinfo - Dispatch/Controller/Model/EventMethods: onDIComplete pattern consolidated - Mapper: glob pattern matching - Views: assets, links, formsdate, rendering Also replaces IIf() calls in dynamic finders with plain if/else (IIf is unreliable across engines and less readable). Tested: Lucee 6 (2278 pass/0 fail) and Adobe 2025 (2287 pass/0 fail). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
StructKeyExists(server, ...)checks across the codebaseEngineAdapterInterfaceapplication.wheels.engineAdapterDetails
Phase 1 (commit 1): Adapter scaffold + 6 high-value method migrations
getResponse(),getStatusCode(),getContentType()— Global.cfc, sse.cfcgetRequestTimeout()— Global.cfcparseFormKey(),controllerNameToUpperCamelCase()— Dispatch.cfcPhase 2 (commit 2): 16 additional adapter methods + 23-file migration
$image,$zip,$hashedKey,$args,$convertToString,$dbinfoNew files:
vendor/wheels/engineAdapters/EngineAdapterInterface.cfcvendor/wheels/engineAdapters/Base.cfc(abstract base)vendor/wheels/engineAdapters/Lucee/LuceeAdapter.cfcvendor/wheels/engineAdapters/Adobe/AdobeAdapter.cfcvendor/wheels/engineAdapters/BoxLang/BoxLangAdapter.cfcvendor/wheels/tests/specs/engineAdapterSpec.cfc(17 BDD tests)Test plan
Closes #1968
🤖 Generated with Claude Code