On Tue, 17 Jan 2023 at 10:44, Daniel P. Berrangé <berra...@redhat.com> wrote: > Something random that looks like bad text data > > ▶ 166/619 /ptimer/oneshot > policy=no_immediate_����������������������������������������������������������������...snip....�������������
I don't understand this one. The string comes from tests/unit/ptimer-test.c:add_ptimer_tests() but that code creates the test string using g_strlcat() and g_strdup_printf(), so I don't see how it could end up being junk, especially not at that point in the string. I guess it's possible something else in the test is managing to corrupt the g_test test name strings? > And what i think is test float being overly verbose Yes, the float tests are massively over-verbose. This is a problem not just in the length of the test log but also in finding the actual cause of a test failure, because often the float tests were running in parallel with the failed test and print all their very verbose output after the failed test output, so you have to scroll back up for miles to find it. thanks -- PMM
