Hello all-
I've been looking at some test failures in the stable tree (I haven't looked at whether they're present in the development tree yet) occurring in the 7.8/SSL testsuite. The failures happen on netbsd and illumos at a minimum, and I'm not able to see what's actually wrong, as the a/b values are equal.
Examples can be found in this verifylog.txt: https://paste.sr.ht/~hww3/6fca76d6c19743525e0b1f7325fea4213b100ecd#http://pi...
The original file is: http://pike.lysator.liu.se/generated/pikefarm/results/8.0/6652_154/verifylog...
So, if o->a==o->b, then why is the test failing? These are not new failures, but if there's a simple solution, I'd like to fix them.
Any thoughts?
Bill
How do you know the values are equal? Note that the log only shows the first 4000 characters of the strings... (The code doing the truncation is in log_msg_result(), in lib/modules/Tools.pmod/Testsuite.pmod, if you want to try without.)
Just to close the loop here, it looks like these failures have the same root cause I was tracking down during the last release: some platforms just need a little more time to flush their buffers. I increased the number of trips through the backend and they went away. Should fix about 26 tests across multiple platforms. Have a few other low hanging fruit to fix and then I’ll cut a release candidate.
On a side note, I don’t suppose there’s any way to report the number of test failures in a xenofarm result? I think that would be handy...
Bill
On Jan 14, 2022, at 6:54 PM, Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum 10353@lyskom.lysator.liu.se wrote:
How do you know the values are equal? Note that the log only shows the first 4000 characters of the strings... (The code doing the truncation is in log_msg_result(), in lib/modules/Tools.pmod/Testsuite.pmod, if you want to try without.)
pike-devel@lists.lysator.liu.se
-
H William Welliver -
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum -
william＠welliver.org