20:02 < mgorny> zmedico: why is portage forcing -j1 to make check? 20:13 < scarabeus> mgorny: most packages fail with paralel build in tests 20:14 < mgorny> :P 20:15 < mgorny> but this forces unavoidable de-parallelisation for packages which work fine with parallel-tests 20:29 < scarabeus> which is probably 1 out of 500 packages :D 20:39 < zmedico> mgorny: maybe that's a PMS thing. 20:40 < mgorny> nope, PMS doesn't say a word 20:40 < mgorny> that's why I'm asking 20:40 < zmedico> better ask on the gentoo-dev ml 20:40 < Betelgeuse> mgorny: Is there a default src_test in Portage but PMS has nothing on it? 20:40 < mgorny> Betelgeuse: could you rephrase that? 20:41 < zmedico> does PMS mentions default_src_test? 20:41 < zmedico> if not, it should 20:41 < Betelgeuse> mgorny: It's a PMS bug if Portage provides a default src_test but PMS does not document it 20:41 < Betelgeuse> Assuming the default src_test is an old thing 20:42 < zmedico> ulm: newins patch looks good 20:42 < Betelgeuse> zmedico: How long as there been a default src_test? 20:42 < zmedico> since EAPI 0 :)
There probably should be a code snippet like for other functions.
PMS doesn't have a code snippet for src_test, but a description that seems unambiguous: # The default implementation used when the ebuild lacks the src_test function # must, if tests are enabled, run make check if and only if such a target is # available, or if not run make test, if and only such a target is available. # In both cases, if make returns non-zero the build must be aborted. Taken literally, this would translate into code as follows: src_test() { if make -n check >&/dev/null; then make check || die "make check failed" elif make -n test >&/dev/null; then make test || die "make test failed" fi }
Closing per comment #2.
