Hi, because this is my first post in this list and I also start with an issue (I have a fix) and questions, some words about me.
I am Torsten from Paderborn/Germany and I am a computer scientist and a mathematician. Besides this, my hobby is sport, especially running and using the bicycle or better 2-wheels. I am using qmapshack regularly, mostly for planning routes/tracks and for loading, inspecting and modifying gpx tracks. Because QMS can read fit files, I also take a look at tracks in fit files without converting them to gps tracks first. Short story: Since I am using the Forerunner 935 together with the Stryd I cannot load my fit files from the FR 935 with QMS. Long story: I rarely load fit files from my FR 935 with QMS. Last Friday I did this, loading a new one with QMS and it failed. But I knew that some time before this was possible. Thinking about the changes in my environment I recognized that I have added the Stryd to the FR 935. And yes this was the reason for the issue. Fit files without the Stryd could be loaded, fit files with the Stryd were failing. So it is time to contribute to the wonderful QMS fixing this issue. I have analyzed the root cause and I have fixed it at home in version 1.15.2 Intro for the questions: Oliver I knew https://github.com/Maproom/qmapshack/wiki/DeveloperCommitCode So I will follow it, e.g. creating a ticket and inside the ticket I will describe the issue and also the root cause and the idea for the fix. After this I will do a pull request. Because everything happens inside the gis/fit folder I used the same programming style and I tried to introduce the fix with only small changes in the existing infrastructure. My questions: Is there something else I should do? E.g. should I run some tests or something similar? I programmed under Ubuntu 20.04, so I have tested it only there. Should I try to compile it on other systems? For Windows this should be possible. For the Mac it is difficult, because I don't have an appropriate system. My experience, doing cross development for a very long time, tells me that it should be fine if it works under Linux. I did not use OS specific things, just normal QT and C++ code. Currently I did everything with Sublime and ddd/gdb, not using Qt Creator. I can/will add an original fit file which had this issue and/or a very short version of it. I used the FitCSVTool(.jar) for creating the short version. Should I put it under test/unittest/input/fit? I will also add the csv file from which I created the fit file together with a short README if this is OK. Additionally I will add a second short fit file which contains a test for a code change. I think, currently the code of this change will not be hit by any created fit file. But according to the specifications it is possible to run into such a situation. The manually created fit file shows this. It is similar to the situation I got now using the FR935 and the Stryd together. In the past it was not expected that this will happen. ;-) A quote from a comment inside the current source: "....... So far no fit file with more than one developer data ID has been created." Greetings, Torsten
_______________________________________________ Qlandkartegt-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/qlandkartegt-users
