OpenCPN Beta Version 3.3.2107 Released

[Baikal]

Originally Posted by bcn

Baikal,

nice video - are the charts from Lake Baikal public and available (for free)?

The grey zones and the problems with tiling (clipped objects and texts) we have described as well for the S-63 charts earlier in the thread and Dave is aware of it.
Blocked overzoom is acknowledged as well.

What else have we missed?

Hubert

charts Baikal as well as all the other charts Russian inland waterways are not publicly available
for end-user products are only available in Jeppesen Marine.
Another identified several challenges
1 No opportunity to move on the map in Win8 with standard high contrast theme.
2 After a while disappears recommended track (in the video at the end)
3. raster KAP also how strange works. charts are not displayed automatically.

[PjotrC]

Displaying GRIB across the IDL...

Not new, same effect also in 2.5.x. Strange I missed it before...

[PjotrC]

Originally Posted by PjotrC

Strange I missed it before...

Now I know why.
When the N edge of the area is in the N hemisphere, all is working fine. Now my GRIB area is completely in the S hemisphere...

[Ptizef]

Dave
I use two machines
1) A quite old Compaq Presario CQ60 16' Intel Pentium dual T3400 @ 17mhz _ VISTA
log GL lines
[CODE]
3:49:22 PM: OpenGL-> Renderer String: Mobile Intel(R) 4 Series Express Chipset Family
3:49:22 PM: OpenGL-> Detected Intel renderer, disabling stencil buffer
3:49:22 PM: OpenGL-> Texture rectangle format: de1
3:49:22 PM: OpenGL-> Detected Windows Intel renderer, disabling Vertexbuffer Objects
3:49:22 PM: OpenGL-> Vertexbuffer Objects unavailable
3:49:22 PM: OpenGL-> Detected Windows Intel Mobile renderer, disabling Frame Buffer Objects
3:49:22 PM: OpenGL-> Framebuffer Objects unavailable
3:49:22 PM: OpenGL-> Using Depth buffer clipping
3:49:22 PM: OpenGL-> Using s3tc dxt1 compression
3:49:22 PM: OpenGL-> Compressed tile size: 128kb (6:1)
3:49:22 PM: OpenGL-> Minimum cartographic line width: 1.0
3:49:22 PM: OpenGL-> Minimum symbol line width: 1.0
[CODE]
I used this one for the tests described in #87,95,101
I compiled and tested the last Git
In release mode it seems to work fine now. I didn't get any crash, switching CM93 <-> Rasters and GL ON <-> OFF, panning and zooming.
In debug mode I still can't use CM93. Sometime I can open it but after a few seconds dragging and zooming I get the attached dump assert failure

2) A low powered Asus x102B 10' AMD A4 1200 APU
log GL lines
[CODE]
4:24:59 PM: OpenGL-> Renderer String: AMD Radeon HD 8180
4:24:59 PM: OpenGL-> Texture rectangle format: de1
4:24:59 PM: OpenGL-> Using Vertexbuffer Objects
4:24:59 PM: OpenGL-> Using Framebuffer Objects
4:24:59 PM: OpenGL-> Using FBO Stencil buffer
4:24:59 PM: OpenGL-> Using Stencil buffer clipping
4:24:59 PM: OpenGL-> Using Scissor Clipping
4:24:59 PM: OpenGL-> Using s3tc dxt1 compression
4:24:59 PM: OpenGL-> Compressed tile size: 128kb (6:1)
4:24:59 PM: OpenGL-> Minimum cartographic line width: 1.0
4:24:59 PM: OpenGL-> Minimum symbol line width: 1.0
[CODE]
On this one I compiled and tested also new Git and I got the same result
In release mode OK ( I was good yet with the official 3.3.2107 )
In debug mode I get exactly the same crash assert interruption
So this problem seems to be graphic card independent

Jean Pierre

[Sail323]

I also can't overzoom in 3.3.2107.

[Ptizef]

Originally Posted by PjotrC

Now I know why.
When the N edge of the area is in the N hemisphere, all is working fine. Now my GRIB area is completely in the S hemisphere...

I've tried different combinations South/North/west/est and I can't reproduce on Windows
You use Linux ? Is this particular file or every file in the south ?
Attached example in the nearly same region

Jean Pierre

[PjotrC]

Originally Posted by Ptizef

Is this particular file or every file in the south ?

Windows. Right, for now I have just this one file...

Displays OK in Expedition LT.

I will try for more samples. Unfortunately this particular grib is .5 MB. A smaller subset indeed shows OK both sides of IDL.

Edit: I re-loaded the grib (Zygrib) for the same area, but a later model run, and all is well...

[bdbcat]

Jean Pierre, and other source builders....

Please checkout and build from current github. A couple of important points:

1. Corrected fault the JP has on windows debug release.
2. Re-enable overzoom on all chart types.
3. Probably solved the leftover process shutdown problem on windows.

Thanks, and good luck
Dave

[bdbcat]

Baikal...

Regarding the trouble with Win8 Hi Contrast theme:

This is a real problem, affects OpenGL only.

Some research done. Has nothing to do with OCPN. OCPN does not know or care about the current screen theme.

Problem seems to be related to Windows screen compositor (WDM). Other OpenGL users have reported this problem with other applications.

Probably nothing can be done about this problem, except suggest:
a. Don't use this theme, or if you must...
b. Don't use OpenGL.

Sorry, but that's it.
Dave

[nohal]

Dave...
The basechart kick-in on zooming is solved for me, overzoom works.
I noticed a glitch while a chart should be displayed, but just a small piece of it is displayed in the upper left corner unless it is panned to cover the whole screen (First screenshot and a video at https://vimeo.com/106397820)
I also believe it should probably already kick in in the situation shown on the second screenshot (it does while zooming in a bit further).
All this involves US81004_1

Pavel

[bdbcat]

Pavel...
Glitch on upper left corner fixed in github now. Simple logic error on screen region calculation for a particular edge case.

The other question is a judgement call: What is the minimum scale at which we want to display the smallest scale chart available? I chose to set the limit so that the chart is "barely" readable, in order to preserve performance on lo-spec machines.

Alt-key + wheel zoom to play with this...

Thanks
Dave

[nohal]

Dave...
Yes, understood. Actually OpenGL is so much faster now after those "useless optimizations" that I moved my definition of barely readable a bit

Pavel

[cagney]

Originally Posted by bdbcat

Jean Pierre, and other source builders....

...
2. Re-enable overzoom on all chart types.
.....

Thanks, and good luck
Dave

Dave

The setting

AllowExtremeOverzoom=0

in the conf file is no longer honored. It has no effect, and on close down O resets it to

AllowExtremeOverzoom=1

Thomas

[Ptizef]

Originally Posted by bdbcat

Jean Pierre, and other source builders....
Please checkout and build from current github. A couple of important points:

Dave,

Done on my two devices, Vista and W8.1 (last commit : "Turn OFF texture cache debug printf. " )

On both assert interruption has disappeared and I got no crash.
For shutdown problem, correction seems to work but needs more tests.

many Thanks
Jean Pierre

[bdbcat]

Cagney...

I have relented philosophically on the issue of overzoom. I cannot be captain on all the boats using OCPN.

So, the config setting "AllowExtremeOverzoom" is deprecated, and does nothing any more. Overzoom is always enabled.

I'll remove it for next Beta.

Thanks
Dave