ckgt / nematode Goto Github PK

For the parser a unit test framework would be useful. There are some open source solutions like Google Test, CppUnit or Boost.Test.

I am using a modem Cypress CTM-200 and processing the GPS coordinates with Nematode.

I am getting the following sentences from the modem. Unfortunately a plus sign is included before the altitude field in the GPGGA sentence. Plus sign is not supported and the sentence cannot be processed.

$PMID,353968095226388,35396809522638810
$GPGGA,194045.00,4531.6094,N,07331.1947,W,1,11,01.1,+00053,M,,M,,00006B
$GPRMC,194045.00,A,4531.6094,N,07331.1947,W,000.2,000.0,160519,,*21

First off, thank you for this great tool, it works really well.

The only problem I have occurs when I use a gps that sends checksums. Sometimes I receive a valid NMEA checksum of 0, and the parser treats that as an invalid checksum and throws an error. An example sentence from my GPS:
$GPRMC,173138.000,V,3145.5214,N,09704.5057,W,000.0,000.0,170318,000.0,E,N*00

I get the error:
GPS Data Bad Format [$GPRMC] :: Checksum is invalid

The month of year field in $GPRMC sentences is indexed 1-12. This is also how it it is stored in the month class member of GPSTimestamp. In getTime(), however mktime() is called, which expects months be 0-11.

This results in the timestamp returned by getTime being off by one month.

The mktime function used to convert the tm struct to time_t (seconds since Jan 1, 1970) assumes that the tm struct time is in local time. Hence, in the instance that the system time is not equal to UTC, the GPS time returned is shifted by the timezone in the opposite direction.

Hey, would you mind putting the actual text of the zlib license in your License file, along with your copyright? I'd like to use the library but don't want to violate zlib by not properly attributing you.

When not including this header ::strtoll and friends are missing (see http://www.cplusplus.com/reference/cstdlib/strtoll/). Tested with arm-none-eabi-g++. Thanks in advance

Hello,

Even though this is not really a big issue, I would like to point out that the code is not suitable for GN*** or GL*** sentences.

I tried to adapt it, and make it work for all prefixes of NMEA sentences but I did not succeed.

Has anyone made an attempt at this before ?

Thank you.

According to https://anavs.com/knowledgebase/nmea-format/ there are additional message formats:

• GP for GPS only solutions
• GL for GLONASS only solutions
• GA for GALILEO only solutions
• GN for multi GNSS solutions

The GPS messages are supported, while the others are not, limiting the usability of the library.

I'll do a fork and add the other formats (as long as I have test data for them) as described in #4 .

And do a pull request later.

otherwise NMEACommand::addChecksum() will add wrong checksum string if checksum>=0x80

#include <string>
#include <iostream>
#include <iomanip>
#include <sstream>

int main()
{
    char checksum = 0x80;
    unsigned char uchecksum = 0x80;
    std::stringstream ss1, ss2;
    ss1  << std::hex << std::uppercase << std::internal << std::setfill('0') << std::setw(2) << (int)checksum;
    ss2  << std::hex << std::uppercase << std::internal << std::setfill('0') << std::setw(2) << (int)uchecksum;
    std::cout<< ss1.str() << ", "<< ss2.str()<<std::endl;
}

result in

FFFFFF80, 80

NMEASentence and NMEAParser constructors cause warnings on member variable initialisation order; I suggest changing from:

NMEASentence::NMEASentence() 
: isvalid(false)
, checksumIsCalculated(false)
, calculatedChecksum(0)
, parsedChecksum(0)
{ }

to

NMEASentence::NMEASentence() 
: isvalid(false)
, checksumIsCalculated(false)
, parsedChecksum(0)
, calculatedChecksum(0)
{ }

and from

NMEAParser::NMEAParser() 
: log(false)
, maxbuffersize(NMEA_PARSER_MAX_BUFFER_SIZE)
, fillingbuffer(false)
{ }

to

NMEAParser::NMEAParser()
: fillingbuffer(false)
, maxbuffersize(NMEA_PARSER_MAX_BUFFER_SIZE)
, log(false)
{ }

or use non-static member initialisation where they're declared.