Re: [gpsd-dev] New slowdown option for regression tests

[Hal Murray]

>> Did you tweak any of the other timings?
> What other timings?  There were only two, WRITE_PAD and CLOSE_DELAY.

I'm just fishing for ideas.  Did you change anything in the test framework 
other than adding the slow option?

A few days ago, every time I ran scons check on that machine it got an error 
on one particular test, and it always got the same error pattern.

Running the same tests today, sometimes it works.  When it fails, it's on the 
same test, but the error pattern is not consistent.  Earlier today, I had 0 
errors out of 5 runs.  A few minutes ago, I had 4 errors out of 5 tries and 
the details were all different.  The next batch was 3 errors out of 5 tries.

That's in fast mode.  I haven't seen any problems in slow mode, but I haven't 
tried very hard.

Maybe it was warmer a few days ago.  ???

I don't know how to drive git.  If you give me a pointer to the recipe to get 
older versions, I'll try some of them to see if I can get it back to a solid 
failure.




-- 
These are my opinions.  I hate spam.