OpenCPN Beta Version 3.3.1931 Released

[CarCode]

See:
http://www.cruisersforum.com/forums/...ml#post1612685

Gerhard

[PjotrC]

Complex tracks seem to be rendered differently now... this probably concerns the (non-contiguous) track segments.

This is fairly important, at least until providing a proper drawing tool. The segments are part of GPX standard, and usually are handled correctly (sample from EasyGPS attached).

[nohal]

Pjotr...
Can you attach the GPX file as well? It will save us a while creating one for testing...

Pavel

[bdbcat]

Folks....

Whoever just now sent a crash report for V3.3.2028....

2028 is in the sandbox, unreleased. It is two hours old...

I cannot use a crash report for builds which you do locally. The PDB symbol files don't match, so sending a crash report is a waste of bandwidth.

Whoever you are, please tell us here what you have found with github master, whenever you find it.

Thanks
Dave

[PjotrC]

Originally Posted by nohal

Pjotr...
Can you attach the GPX file as well? It will save us a while creating one for testing...

Pavel

This example can be found in the Layers Library

http://opencpn.org/ocpn/sites/defaul...d/TSS_zones.7z

The tracks shown above are: pl_akwen6, pl_akwen6a and TSS_Lawica_Slupska. The most complex there is probably TSS_Canarias.

Thanks for looking at this

P.

[tom_BigSpeedy]

I compiled latest sources and had a short try ...

It's worse now, than it EVER was. Took me 30 secs to crash O !
Crashreport sent.
Started O with OpenGL off and CM93 : OK
Turned OpenGL on : OK
Switched to raster KAPs to test zooming : OK
Switched back to CM93 : BOOOOM !

Restarted O : Black screen, no more CM93
Turned OpenGL off : CM93 back again, switching to KAP and back works
Turned OpenGL on with CM93: Black Screen, but no crash.

Rebuilt complete Chart DB : nothing changed ...
As soon as I turn OpenGL on, by screen turns black with CM93 !

I attached the part of the logfile that ended with the crash.

Annoyed,

Thomas

It's my Win7 Laptop; Intel grahpics drivers are updated to the latest version ...

[bcn]

Originally Posted by bdbcat

Folks....

Whoever just now sent a crash report for V3.3.2028....

2028 is in the sandbox, unreleased. It is two hours old...

I cannot use a crash report for builds which you do locally. The PDB symbol files don't match, so sending a crash report is a waste of bandwidth.

Whoever you are, please tell us here what you have found with github master, whenever you find it.

Thanks
Dave

The wolfes on your heels...
But as you have a very experienced dog to defend you

[tom_BigSpeedy]

Dave,

it was me, sending the crash report. See 1 post above yours.

OK, will stop sending crash reports ...

See crash description above.

Currently I cannot use OpenGL anymore.
Which inhibits me in testing weather_routing, as it requires OpenGL. That's what I originally intended to do ...

But believing in Santa Claus, I thought, "aah, let's compile the latest sources first ..."



System was set up on the latest available 3.31931 directory structure.
So there's nothing missing ...

Thomas

[CarCode]

Dave,

Commit: 1ba30b8a2f03df5a0be88c5a57566857a8b015d6 [1ba30b8]
Correct fault on waypoint HTML link delete by context menu.

doesn't work. OpenCPN still crashes but I couldn't find why. Probably has something to do with MouseEventHandler.

Gerhard

[cagney]

Dave

There are a lot of info missing in CM93 using the current git version.
Compare these two screen dumps.



Both versions running at the same time, using the same conf file.

Thomas

[bdbcat]

tom_BigSpeedy....

Ah, nice to hear from you on 3.3.2028.
Glad to hear also that you can build and test for Windows. Rare and valuable skill, that...

I suspect continuing problems with Intel OpenGL drivers. Not your fault, they are what they are. In 3.3.2028, I tried to re-enable "simple" stencil buffer operations for Intel, but maybe this is still asking too much. It does work for my current stable...

Or maybe it is something completely different.

Anyway, if you can build, I would appreciate it if you could try the following patch and see if OpenGL behaves for you.

in current github, glChartCanvas.cpp:line 881

    s_b_useStencil = false;
//    if( stencil && ( sb == 8 ) )
//        s_b_useStencil = true;

This patch will force disable the OpenGL stencil buffer, and resort to the somewhat slower "depth buffer" clipping algorithm.

Let me know, please.

Thanks
Dave

[bdbcat]

cagney...

Regarding missing Depth Contours on Staffans Bank near Landskrona...

Strangely, it only affected that one cm93 E-level cell. Probably related to object order in the database. Anyway, definitely related to display priority calculation for line features.


A fix in github now.
Thanks
Dave

[tom_BigSpeedy]

Originally Posted by bdbcat

tom_BigSpeedy....


Anyway, if you can build, I would appreciate it if you could try the following patch and see if OpenGL behaves for you.

in current github, glChartCanvas.cpp:line 881

    s_b_useStencil = false;
//    if( stencil && ( sb == 8 ) )
//        s_b_useStencil = true;

This patch will force disable the OpenGL stencil buffer, and resort to the somewhat slower "depth buffer" clipping algorithm.

Let me know, please.

Thanks
Dave

I'm at work now, will do tonight, and let you know.
Thomas

[bdbcat]

Speedy Thomas...

OK, thanks. I made some more changes in githib and definitely disabled the buggy stencil buffer for Intel graphics platforms.

Probably best for you to now pull the latest and try a fresh build, since the little patch above will not be the same in the current github.

Thanks
Dave

[tom_BigSpeedy]

Hi Dave,

nope, doesn't work.
As soon as I set the tick at OpenGL, my screen turns black ...
I also removed the ini file, to have a new one created ...

And now it doesn't matter, if I switch from BSB to CM93 and back.

It stays black, see attachment.

Thomas