Branch: refs/heads/master Home: https://github.com/xwiki/xwiki-dev-llm Commit: c4388f99cf42edcd47d6e9235de7525c0a421a19 https://github.com/xwiki/xwiki-dev-llm/commit/c4388f99cf42edcd47d6e9235de752... Author: Vincent Massol <vincent@massol.net> Date: 2026-08-27 (Thu, 27 Aug 2026) Changed paths: M .claude-plugin/marketplace.json M kimi.plugin.json M opencode.jsonc M xwiki/.claude-plugin/plugin.json M xwiki/okf/index.md M xwiki/okf/testing/strategy.md M xwiki/skills/xwiki-review/SKILL.md M xwiki/skills/xwiki-test-guidelines/SKILL.md Log Message: ----------- [Misc] Record that a functional test is a scenario with a single fixture * Add the scenario rule to okf/testing/strategy.md: a *IT pays a wiki start, a browser start and a page load per navigation, so the number of fixtures drives its runtime, not the number of assertions. Write as few @Test methods as the scenario allows, extend an existing method or *IT rather than rebuilding the same fixture, and note that @Order fixes execution order without sharing a fixture so it is not a substitute. Sourced from the dev wiki's functional-testing best practices, which the plugin only linked and never stated. * Add the matching procedural bullet to the xwiki-test-guidelines skill, so the rule applies when writing a test and not only when reviewing one. * Add the check to the xwiki-review Tests angle: a new @Test method or *IT class that rebuilds a fixture an existing one already builds is a finding. * Update the strategy summary and the OKF map entry, and bump the plugin version. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Commit: 63d3343a254b846586abcdc75652f160c676698a https://github.com/xwiki/xwiki-dev-llm/commit/63d3343a254b846586abcdc75652f1... Author: Vincent Massol <vincent@massol.net> Date: 2026-08-27 (Thu, 27 Aug 2026) Changed paths: M xwiki/okf/index.md M xwiki/okf/testing/strategy.md M xwiki/skills/xwiki-test-guidelines/SKILL.md Log Message: ----------- [Misc] Pivot the scenario rule from method count to duplicate fixtures * Restated the rule as "no two @Test methods build the same fixture" instead of "as few methods as possible, one if you can". The defect the rule targets is a second method paying for a fixture an existing one already builds, not method count in itself, and the previous wording read as a push towards a single method however long it grew. * Added the counterweight: a method nobody can follow end to end, or one whose failure no longer says which behaviour broke, has been merged too far. A distinct fixture justifies a distinct method, a merely distinct assertion does not. * Mirrored the change in the xwiki-test-guidelines bullet, the strategy summary and the OKF map. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Commit: f2679a39ede9629ffd5aea87e0587668cc8fd8fc https://github.com/xwiki/xwiki-dev-llm/commit/f2679a39ede9629ffd5aea87e05876... Author: Vincent Massol <vincent@massol.net> Date: 2026-08-27 (Thu, 27 Aug 2026) Changed paths: M .claude-plugin/marketplace.json M kimi.plugin.json M opencode.jsonc M xwiki/.claude-plugin/plugin.json M xwiki/okf/index.md M xwiki/okf/testing/strategy.md M xwiki/skills/xwiki-review/SKILL.md M xwiki/skills/xwiki-test-guidelines/SKILL.md Log Message: ----------- [Misc] Record that a functional test is a scenario with a single fixture Compare: https://github.com/xwiki/xwiki-dev-llm/compare/707cdb63c1c7...f2679a39ede9 To unsubscribe from these emails, change your notification settings at https://github.com/xwiki/xwiki-dev-llm/settings/notifications