Re: [Gpsbabel-code] GUI idea (was Re: gpsbabel lowranceusr.c bug report+fix)
Brought to you by:
robertl
From: Robert L. <rob...@gp...> - 2013-07-18 15:16:46
|
We replaced guibabel years ago with the Qt code. That's where any improvements should be focused. On Jul 18, 2013 10:09 AM, "Jerzy Tarasiuk" <Jer...@fu...> wrote: > Hello, thanks for your reply. > > I found a file "guibabel" distributed with the gpsbabel > - it is quite antiquated (2006-03-21) and does not handle > most of gpsbabel formats, so I have idea to make better. > > My idea is that GUI gets info what the gpsbabel can do > from the program and allows to specify all possible > formats, filters and options - without need to change > the GUI script when new ones are added. > > I planned using gdb to extract format and filter lists > and their options, but it is slow (takes minutes to get > all data) and I still do not have full information what > I need to get - how to "decode" internal formats, which > are defined in the file "internal_styles.c". > > Options -^3 and -%1 (about the first I found info in > the "gpsbabel-code" archive, second I found looking into > main.c) give the info quickly, although something is > missed in it: option type flags, there are e.g. ranges > of mutually exclusive options, seems -^3 does not show > that some options form such a range. > > Questions: Is currently someone working on GUI in Tcl/Tk? > What such a GUI should provide to be useful? I assume > the most important is ability to find the program usage > info without need to look into documentation, and apply > it in a convenient way (without typing names); with near > 200 formats, a way to search within them is a must. > > With kind regards, Jerzy Tarasiuk > > > Some clarifications about my previous mail: > > > On Fri, Jul 5, 2013 at 12:06 PM, Jerzy Tarasiuk > > <Jer...@fu...>wrote: > >> Symptoms: I used an Lowrance USR (from Silva Atlas GPS) file > >> and converted it to GPX (-i lowranceusr -o gpx), then back > >> to Lowrance USR (-i gpx -o lowranceusr), and again to GPX. > >> The second GPX file was significantly different from the first > >> - many (probably all, I did not verify it) coordinates were > >> shifted, they became slightly smaller than they were. > > Where "significantly" presumably means lower than the precision of your > > GPS, right? > > Right. The "significance" meant amount of differences, > not severity of them - the shift was around 0.6 m. > > > This is C++; we always have a well-defined > > lround().<http://www.cplusplus.com/reference/cmath/lround/> > > The program can be compiled as C program, and the compiler > needs -std=c99 option to recognize then lround(). I don't > know if there is anyone using older compiler w/o the option. > > >> With the fix, the only difference becomes line 8, containing > >> time where the file was written. I have an idea to fix this, > > I don't know what line 8 you're speaking of, but this change rather > > predictably blew up the test suite. I've regenerated your reference > > files and have committed the whole change in r4401. > > I meant line 8 of GPX file - it contains time when the file > was written, so obviously if can be different; setting the > GPSBABEL_FREEZE_TIME stores beginning of Unix time there, > so the time is the same each time - useful for testing. > > > The timestamp in GPX is working as intended. If we were a simple > > 1:1 converter, some cleverness in preserving source file creation might > > be in order, but since we can merge and filter files, we regard the > output > > OK. Maybe it would be useful to timestamp the output file > according to data in it - to newest time found in the data > - maybe as a filter option? Especially for those who have > many data files collected, so they could use "ls -lrt"... > > I understand the file may be newly created work, I just > did not use gpsbabel in other way than format change. > > > |