First of all, thanks Ananth and Cai for help!
Jan, I need your help ;)
looking at different reports I can't understand how to interpret them.
To the point, I do not understand if the overall results are good or bad.
The first question, are these tests supposed to be "stable"?
For example,
> Those are the test results on i686 F12 hosts with and without
> CONFIG_UTRACE. Interesting thing is that the results on quite
> different on two Intel hosts.
Yes! I'd say the results are just "reversed". We have some PASS->FAIL
changes on "ProLiant DL360 G4p", and on "HP Workstation XW4200" machine
the _same_ tests show FAIL->PASS change. I can't imagine how utrace-ptrace
can explain this difference if results are stable.
I spent several hours trying to figure out how can I fix the failures,
but since I never used gdb this is very much nontrivial to me. Because
I just don't understand whats going on and what any particular test
actually does.
So. Given that the number of test is huge, and (I guess) we can't
hope utrace-ptrace can pass 100% of tests, I am asking you to tell
me which failures are "important". Then I'll try to fix them (most
probably I will ask a lot of stupid questions before I will be able
to do this ;).
IOW. Only you and Roland can say whether utrace-ptrace is ready
for use from gdb-testsuite's pov. Please tell me which failures
should be fixed to conclude that utrace-ptrace "works".
Oleg.
