Automation Test Reports

Automation Test Reports — Modernized (TestLink 2.0.1, Refs #1062 / #1921)

Net-new Automation Test Reports screen for TestLink, added to the Reports menu. TestLink 1.9.20 shipped no automation reporting at all: automated results were written into the same tl executions table as manual ones and could only be counted through the generic metrics/charts reports, which never split automated from manual execution, never exposed the automated test-case inventory, the per-build automation trend or a failed-automation worklist. This is therefore a 2.0.1 enhancement rather than a legacy migration — nothing is dropped, because nothing existed.

Why this screen

Issue #1062 asks for a dedicated Reports › Automation Test Reports page that aggregates results received from automation frameworks / CI / external runners: overall status, results grouped by test plan/build/platform, manual-vs-automated coverage, a trend across builds and a worklist of failures. This screen delivers the part that maps onto the existing schema (no new tables): the manual-vs- automated split of the plan's linked test cases, the latest execution status per test case version, a per-build trend and a filterable worklist with CSV export.

Modern stack

Layer	File
Screen	gui/templates/results/automationReports.html (Dashio standalone HTML/JS/CSS)
BFF	api/automationreports/index.php (GET|HEAD ?action=init|report)
Wiring	cfg/reports.cfg.php automation_reports + api/aside/index.php link_report_automation_reports + $actions->automationReports in lib/functions/common.php
i18n	autom.* (57 keys) + footers.automationReports in all 10 client bundles; $TLS_link_report_automation_reports in all 19 server locales

BFF contract

Method / route	Purpose	Notes
GET|HEAD ?action=init&tplan_id=N[&tproject_id=P]	context + filter options	returns context{tplan_id,tplan_name,tproject_id,tproject_name} and filters{builds[],platforms[]}
GET|HEAD ?action=report&tplan_id=N[&tproject_id=P][&build_id=B][&platform_id=PL]	the aggregate	returns context, filters, filter, summary, build_trend[], rows[], rows_truncated

Error contract: 400 invalid_request, 401 not_authenticated / session_expired, 403 no_right, 404 tplan_not_found / project_mismatch, 405 method_not_allowed.

A build_id / platform_id that does not belong to the owning test project is refused with 400 invalid_request (never silently ignored — that would leak counts from a foreign project).

Semantics of the aggregate

• The unit is the test-case version linked to the test plan (testplan_tcversions.tcversion_id), because execution_type — manual vs automated — is a property of a version (tcversions.execution_type; 1 = manual, 2 = automated).
• A linked version is counted once even when the plan links it on several platforms.
• The latest recorded execution of a version (per version + platform, across all builds unless a build is filtered) supplies the reported status; a version with no execution is not run. This mirrors the "absolute latest execution" convention of the legacy neverRun / resultsByStatus reports.
• summary.automation_coverage = automated linked versions / linked versions.
• build_trend is the same latest-per-(version, build) status counted per build (the "trend across builds" the enhancement asks for).

Report content

• Coverage strip — linked test cases, automated, manual, % automated, executed, not run.
• Status by automation type — one stacked bar per automation type (passed / failed / blocked / not run) drawn with plain CSS.
• Trend across builds — a compact table: build, passed, failed, blocked, executed.
• Worklist — a sortable DataTable: ID, test case, suite, type, status, build, platform, last execution, tester.
• Filters — build and platform (project-scoped).
• Export CSV — the worklist (UTF-8 BOM, ;-separated, with a spreadsheet formula-injection guard).

Interactions

• Refresh — re-runs the report with the current filters.
• Apply — re-runs the report with the selected build/platform.
• Export CSV — downloads the worklist; a Nothing to export yet toast when the report is empty.
• Close — closes the window (same idiom as the sibling result screens).
• Locale switcher — the standard TLi18n switcher; the whole page follows the selected bundle (English + Romanian translated; the remaining bundles fall back to English, matching the rest of the results family).
• State cards — every BFF error code maps to a friendly title/body card, with the raw server message shown as an escaped detail.

Hardening

• Session auth (401 anonymous) + bffSameOriginGuard() + bffEnforceSession().
• testplan_metrics checked on the owning test project AND test plan with getAccess = true (a right held only globally does not reach a private test project), mirroring the accepted siblings api/keywordbarchart, api/prioritybarchart, api/toplevelsuitesbarchart.
• The plan is resolved first (opaque 404 tplan_not_found); the client's tproject_id is honoured only after the right check (project_mismatch).
• Strict positive-integer readers for every id (1abc, 1.9, -3, '' refused).
• html_entity_decode() on stored names; every BFF string is escaped client-side before it reaches the DOM; explicit $_GET reads (not $_REQUEST); Cache-Control: no-store.

Verification

Fixture tmp/fixtures_1921.sql: public project AR Project (9100), plan AR Plan (9130) with 4 linked versions (2 manual, 2 automated) + empty plan AR Empty Plan (9131), build AR Build 1/AR Build 2, platform AR Linux. Expected and measured (no filter): linked 4, automated 2, manual 2, %automated 50, passed 1, failed 2, blocked 0, not run 1, executed 3; trend Build 1 → P1/F2/E3, Build 2 → P1/F0/E1. Build filter → executed 1, not run 3. Empty plan → empty state. ai/verify_i18n_coverage.sh 9/9 PASS. Event Viewer clean.