Re: b32ff1a86 Breaks OSX and FreeBSD

[Bernd Zeimetz]

On 12/21/19 3:58 AM, Gary E. Miller wrote:
> bzed: But I actually think I found a bug in fake.py
> bzed: send() on a socket does not care if the data was actually send. it 
> returns the number of bytes that were sent, the appication has to take care 
> of it
> bzed: it should be sendall()
> bzed: I'll give that a try in my fork
> 
> Fingers crossed.

Done that with my own sendall implementation with some extra logging.
Didn't fix the issue, but I think it would still be good to improve
the code here, but not the for upcoming release.

https://gitlab.com/bzed/gpsd/commit/c747b78d30963ffe943dcdf5eb22de89123c5e3f
if anybody is interested.

So the next task will be to figure out if the data is changes/lost
during transport, or in gpsd, or if the parser is having issues.

gps/fake.py gives hints that we are facing a timing related issue
here. But tcp communication and timing issues sounds completely
wrong to me, so my guess is a threading, locking, Python GIL or
whatever else issue.... will be interesting to find it as it is
not easy reproducible.


-- 
 Bernd Zeimetz                            Debian GNU/Linux Developer
 http://bzed.de                                http://www.debian.org
 GPG Fingerprint: ECA1 E3F2 8E11 2432 D485  DD95 EB36 171A 6FF9 435F