Re: [gpsd-dev] Hal's sirf1 bug fixed

[Eric S. Raymond]

Hal Murray <address@hidden>:
> 
> address@hidden said:
> > Any time you see a test failure that consists of a solid stretch of omitted
> > lines, that's the address@hidden timing problem.  You should be able to 
> > make it go
> > away by increasing the WRITE_PAD and CLOSE_DELAY values in gps/fake.py. 
> 
> Since all my recent problem (on Linux) were with tcp-test.log, I added 100 
> passes through that to my check script.
> 
> I changed the CLOSE_DELAY on the linux branch from 0.1 to 0.2
> That solved all my problems on Linux boxes.
> 
> I haven't seen any problems on FreeBSD.
> 
> It works on a Raspberry Pi.
> 
> I'm still getting lots of errors on a 64 bit NetBSD box, even with slow=yes.
> It works on a 32 bit system (without slow mode).
> 
> I tried 
> WRITE_PAD_SLOWDOWN = 0.1
> CLOSE_DELAY_SLOWDOWN = 4.0
> That mostly worked.  Yes it was slow.
> 
> Here is the only glitch I noticed:
> Processing test/daemon/tcp-test.log 
> gpsd:ERROR: TCP device open error can't connect to host/port pair.
> gpsd:ERROR: tcp://127.0.0.1:49194: device activation failed.
> Processing test/daemon/tcp-torture.log
> 
> The summary didn't complain about any errors.
>   Regression test successful: no errors in 89 tests (0 not found).
> My extra 100 tries of tcp-test.log worked OK.

I think we're good to go.  This is all race conditions in the tests -
irritating but (unlike the hour-offset bug) not a real problem.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>