Package: autopkgtest Version: 5.4.2 Tags: patch
>From ee5188ba223d31cbdca70f0601c7683090cddfa9 Mon Sep 17 00:00:00 2001 From: Ian Jackson <ijack...@chiark.greenend.org.uk> Date: Fri, 3 Aug 2018 03:32:50 +0100 Subject: [PATCH] doc/README.package-tests.rst: document hint-testsuite-triggers
There is currently no official way to influence which packages trigger a retest on ci.d.n. However, there is a trick, involving an unrunnable test which exists only to get its Depends into Testsuite-Triggers. This is used by the dgit packagae. In lieu of a better mechanism, this trick ought to be documented. Documenting it also means that future ci machinery can know what this strange test actually means. Signed-off-by: Ian Jackson <ijack...@chiark.greenend.org.uk> --- doc/README.package-tests.rst | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/doc/README.package-tests.rst b/doc/README.package-tests.rst index 384eacd..44a9197 100644 --- a/doc/README.package-tests.rst +++ b/doc/README.package-tests.rst @@ -269,6 +269,28 @@ skippable that ``skippable`` tests never exit with status 77 for reasons that should be treated as a failure. +hint-testsuite-triggers + This test exists purely as a hint to suggest when rerunning the + tests is likely to be useful. Specifically, it exists to + influence the way dpkg-source generates the Testsuite-Triggers + .dsc header from test metadata: the Depends for this test are + to be added to Testsuite-Triggers. (Just as they are for any other + test.) + + The test with the hint-testsuite-triggers restriction should not + actually be run. Future systems which understand per-test update + triggering should treat the Depends of the test with this + restriction, as triggering packages for all tests in this + debian/tests/control. + + The packages listed as Depends for this test are usually indirect + dependencies, updates to which are considered to pose a risk of + regressions in other tests defined in this package. + + There is currently no way to specify this hint on a per-test + basis; but in any case the debian.org machinery is not able to + think about triggering individual tests. + Defined features ---------------- -- 2.11.0
-- Ian Jackson <ijack...@chiark.greenend.org.uk> These opinions are my own. If I emailed you from an address @fyvzl.net or @evade.org.uk, that is a private address which bypasses my fierce spamfilter.