OpenCPN Version 2.6 Beta Build 1705

[bdbcat]

Sandeman....

The NMEA source GPSD probably does not work with most recent gpsd revisions. It is scheduled to be dropped soon.

Try Network LIBGPS instead....

Thanks
Dave

[sandeman]

Originally Posted by bdbcat

Sandeman....

The NMEA source GPSD probably does not work with most recent gpsd revisions. It is scheduled to be dropped soon.

Try Network LIBGPS instead....

Thanks
Dave

Thank you again Dave.

I re-installed GPSD and switched to libgps in the opencpn configuration. The position is now displayed properly.

The "Show GPS/NMEA data stream window" displays no output? Although the Dashboard and XGPS do display the position correctly.

Sandeman

[nohal]

Sandeman...
gpsd does not output NMEA sentences, but it's own data structures - That's why there's nothing to be displayed in the NMEA stream window.

Pavel

[cagney]

Originally Posted by bdbcat

Thomas....

Crash on 64bits:

As I understand it, you can reliably crash on 64 bits with one specific config file, which you have previously attached.

I tried that config file locally on 32 bits, no problems.

If you can reliably crash, then the thing to do is systematically remove questionable items from the config file, one, by one, until you find the one item that causes the crash.

You ddd dump is not conclusive, I'm afraid.

Thanks
Dave

Dave

It's a gpsd problem. Whenever the NMEA data source is set to "Network LIBGPS" when closing down, opencpn crashes on the next start. Deleting (or changing) the line "Source=LIBGPS:localhost" in the conf file, is all that is needed to start the application again.

Thomas

[bdbcat]

Hello all...

Anyone else having trouble with F5 colorscheme change on cm93? See post #50, this thread.....

Works fine for me.

Thanks
Dave

[bdbcat]

Thomas....

64bit libgps problem:

Does xgps work on this system with gpsd started?

Thanks
Dave

[sailorF54]

Originally Posted by bdbcat

Hello all...

Anyone else having trouble with F5 colorscheme change on cm93? See post #50, this thread.....

Works fine for me.

Thanks
Dave

Vista
First issue: duplicated, but only in non GL model:
You can't get out of the dusk or night mode with F5 (it only affects routes, ownship etc... see pic with route and ownship highlighted in night mode) until you change scale or pan but then F5 only affects toolbar, status bar...
You then end up with weird things onscreen after panning (see pic, note quilting is on)

[sailorF54]

Originally Posted by bdbcat

Thomas....

64bit libgps problem:

Does xgps work on this system with gpsd started?

Thanks
Dave

Second issue

Duplicated, in non GL mode (Vista)
in GL mode, dimming does affect the whole computer screen: when you minimize OpenCpn, the screen remains dimmed.

[rooiedirk]

Hoi Dave,
Also duplicated under linux.
only non OpenGL. Hitting F5 dims only window and toolbar, but not the chart area. when you start panning, the 'new' visible chart areas are dimmed as they should be.
Hitting F5 again has no affect on chart area, only on window and toolbar, giving weird affect on the chart area.
Dirk

[nohal]

Dave...
Also can reproduce the dimming problem - This is a product of panning the chart to the SW and then NE while using F5... Have looked at the code for a little while, but can't find a way to invalidate and redraw the chart canvas properly

Pavel

[cagney]

Originally Posted by bdbcat

Thomas....

64bit libgps problem:

Does xgps work on this system with gpsd started?

Thanks
Dave

Dave
Once OpenCPN is up and running, gps through Network LIBGPS , and xgps both works.

Thomas

[JesperWe]

Originally Posted by bdbcat

Anyone else having trouble with F5 colorscheme change on cm93? See post #50, this thread.....

Works fine for me.

Dave, I see it. It's the s57chart.cpp:2236 that doesn't fire, so reallyNew is false and there is no update.

/j

[bdbcat]

All...

Thanks for the input on F5 colorscheme change problem...

I must be dense. I cannot reproduce it.

To confirm....

Procedure:
linux...
1705...
cm93, quilted....
non-opengl...

Hit key F5, and observe that chart does not change color, while screen decorations do change color.
Pan some, and see the new sections in proper color.

I cannot reproduce. Must be some race condition?

I will look at the code. Thanks Jesper for the hint.

Dave

[JesperWe]

Originally Posted by bdbcat

I must be dense. I cannot reproduce it.

Very strange. Looks like it's my fault, but what I cannot understand is why I had it working before, and you have it working now.

This is clearly missing:

diff --git a/src/s52plib.cpp b/src/s52plib.cpp
index 3e5b619..dafb652 100644
--- a/src/s52plib.cpp
+++ b/src/s52plib.cpp
@@ -996,6 +996,8 @@
 
       if( !useLegacyRaster ) ChartSymbols::LoadRasterFileForColorTable(
                   m_colortable_index );
+
+      m_ColorScheme = scheme;
 }
 
 S52color* s52plib::getColor( const char *colorName ) {

[rooiedirk]

Above patch is indeed doing the trick for the weird F5 behavior for me
thanks.
Dirk