On Thu, Jun 06, 2019 at 08:33:26PM +0200, PiBa-NL wrote: > Hi Olivier, > > Op 6-6-2019 om 18:20 schreef Olivier Houchard: > > Hi Pieter, > > > > On Wed, Jun 05, 2019 at 09:00:22PM +0200, PiBa-NL wrote: > > > Hi Olivier, > > > > > > It seems this commit ea8dd94 broke something for my FreeBSD11 system. > > > Before that commit (almost) all vtest's succeed. After it several cause > > > core-dumps. (and keep doing that including the current HEAD: 03abf2d ) > > > > > > Can you take a look at the issue? > > > > > > Below in this mail are the following: > > > - gdb# bt full of one of the crashed tests.. > > > - summary of failed tests > > > > > > Regards, > > > PiBa-NL (Pieter) > > > > > Indeed, there were a few issues. I know pushed enough patches so that I only > > fail one reg test, which also failed before the offending commit > > (reg-tests/compression/basic.vtc). > > Can you confirm it's doing better for you too ? > > > > Thanks ! > > > > Olivier > > Looks better for me :). > > Testing with haproxy version: 2.0-dev5-7b3a79f > 0 tests failed, 0 tests skipped, 36 tests passed > > This includes the /compression/basic.vtc for me. > > p.s. > This result doesn't "always" happen. But at least it seems 'just as good' as > before ea8dd94. For example i still see this on my tests: > 1 tests failed, 0 tests skipped, 35 tests passed > ########################## Gathering results ########################## > ###### Test case: > ./work/haproxy-7b3a79f/reg-tests/http-rules/converters_ipmask_concat_strcmp_field_word.vtc > ###### > ## test results in: > "/tmp/haregtests-2019-06-06_20-22-18.5Z9PR6/vtc.4579.056b0e93" > ---- c1 0.3 EXPECT resp.status (504) == "200" failed > But then another test run of the same binary again says '36 passed'.. so it > seems some tests are rather timing sensitive, or maybe a other variable > doesn't play nice.. Anyhow the core-dump as reported is fixed. Ill try and > find why the testresults are a bit inconsistent when running them > repeatedly.. Anyhow ill send a new mail for that if i find something > conclusive :). >
Hi Pieter, Yeah I had this one randomly failing for quite some time, on Linux and on FreeBSD, haven't bothered to investigate yet. Regards, Olivier
