Re: [gpsd-dev] Regression test failed on fortrex-201.log.chk

[address@hidden]

Hello,

for some reason, gpsd do not switch  the m8 to u-blox mode.

This is the reason, why this problem occur.

I do not have a nmea0183 v4.1 spec at the hand now, but the u-blox spec says, 
that  NMEA0183 v4.1 added a signal id field at the end of the xxGSV  sentence.

For the u-blox m8, you can switch to nmea0183 v4.0 compatibility with the 
windows u-center software.

Until gpsd is able to switch a m8 receiver to u-blox mode, we will get a lot of 
support requests.

And on the long term, i think, that we can not ignore the new revision 4.1 of 
the standard.. 

Reinhard

-----Original-Nachricht-----
Betreff: Re: Regression test failed on fortrex-201.log.chk
Datum: Mon, 01 Jun 2015 11:18:01 +0200
Von: "Eric S. Raymond" <address@hidden>
An: "address@hidden" <address@hidden>

address@hidden <address@hidden>:
> The line 154 is truncated in the check file and in the log file.

I just fixed that.

> 1.) Why does the truncated line pass the nmea0183 parser.

Because there's no check in the parser for the total number of sat records -
and probably cannot be, it varies too much across receivers.

> 2.) Why do you think, that the patch is not nessary, u-blox m8 receiver, i 
> use a navilock NL-8222MP do not report satellite positions without.

Probably because the M8 switches to ublox binary mode.

> 3.) Is it more important to support this broken checkfile then the output of 
> devices claiming to comform to nmea0183 version 4.1? 

No.  We should have a short-line test, but one isolated from the tests of
well-formed output on any particular device.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>