OpenCPN Beta Version 3.3.1731 Released

[bdbcat]

Gilletarom...

on this problem:

2° I have 3 groups of charts: CM93, S63 and NV-Verlag....And a hole appear on the chart

Please try this:
1. Remove cm93
2. Remove S63 charts/plugin.
3. Load only NV-Verlag.
4. Make a group for NV-Varlg

Is the hole still present when you show NV-Verlag group?
Is the hole present when you show all charts (should be same, since there is only one group.)

Thanks
Dave

[bdbcat]

sailorF54/jp...

regarding Dialog Font problems:

Please verify the setting of options->"Enable responsive graphics interface".
It should be OFF (unchecked).

I think there is a problem here. It seems that for dialog font sizes to take effect, the responsive interface should be disabled.

Also, I note the duplicate French/English items in the font element selector menu. This is fixed for next Beta.

Thanks for the feedback.
Dave

[sailorF54]

Originally Posted by bdbcat

sailorF54/jp...

I think there is a problem here. It seems that for dialog font sizes to take effect, the responsive interface should be disabled.

Thanks for the feedback.
Dave

It was disabled when I tested...
(I don't know exactly what it is for, so I didn't enable it... )

[Ptizef]

Originally Posted by bdbcat

sailorF54/jp...

regarding Dialog Font problems:

Please verify the setting of options->"Enable responsive graphics interface".
It should be OFF (unchecked).

I think there is a problem here. It seems that for dialog font sizes to take effect, the responsive interface should be disabled.

Yes it was unchecked when I tested.

Originally Posted by bdbcat

Also, I note the duplicate French/English items in the font element selector menu. This is fixed for next Beta.

It's Ok for me with the last Git

Jean Pierre

[Ptizef]

I got a curious phenomenon in the display on my 10' screen W8.1 net book :
You can see this strange behavior in the three shots attached (two with CM93 and one with rasters) ( both with original 1731 and last Git )

I don't know how it happens, but it happened only in openGL mode (I was unable to reproduce in no GL mode) For the moment, I was also unable to reproduce it on My other Vista notebook
When it happens, the only way to stop it is to restart O ( Or deactivate GL mode )
Anyone else has this problem ?

I precise that it's relatively new as I have more than 150 hours navigation with that machine, mainly with 3.3.1606 and 1710 without that problem

Thanks
Jean Pierre

[Gilletarom]

Originally Posted by bdbcat

Gilletarom...
....
[/CODE]Please try this:
1. Remove cm93
2. Remove S63 charts/plugin.
3. Load only NV-Verlag.
4. Make a group for NV-Varlg

Is the hole still present when you show NV-Verlag group?
Is the hole present when you show all charts (should be same, since there is only one group.)

Thanks
Dave

Hello Dave,

1 This morning when I prepared the post, I started writing "I can easily easily reproduce" the anomaly. And oddly, in that moment, I could not reproduce the phenomenon. So I wrote only "I can reproduce the bug."

2 Tonight, repeating my tests, I easily got to reproduce. I even had several holes in the charts. See screenshot.

I noticed that it was important to click zoom - before returning to the NV-Verlag charts, if I wanted to reproduce the anomaly.

I also noticed that the holes disappear, immediately, if I clicked Zoom -.

3 I removed all charts except the NV-Verlag. And I deleted all groups except the NV-Verlag group.
From that moment, I could not reproduce the phenomenon. O I stopped, I raised O. Without reproduce the phenomenon.

B.R. Gilletarom.

[tom_BigSpeedy]

I had 2 more crashes yesterday.
The 1st one with the latest sources (compiled with wxWidgets 2.8.12, VC2010, Win7), the second one with the official 3.3.1731.
I had openGL on, on both sessions. They ran unattendedly on the laptop under deck while we were sailing, so there was definitely no user interaction. I only used the dashboard plugin.
Both crashes happened after about 3-4 hrs.
Strange : I had track logging on, and after the crashes the complete tracks (from start of O, recorded till then) were gone.
This is new to me, at least the track was recorded and saved OK in the past ...

Both sessions ran with -p option.

The O windows had completely disappeared and only the standard Windows popup was there ("O does not react anymore...")

After that experience I switched back to 1606, which did it's job as expected.


Thomas

[Ptizef]

Related to the display problem (#125)
I've found when it happens :
Displaying a colored overlay map in grib-pi immediately start the problem and it continues even I close grib-pi. The only way to stop the phenomenon is to restart O. I don't understand, because I did dozen and dozen times and I've never seen this before. I suppose it started with 1731
And it seems that the more I display overlays, the worth it is !
There must be some incompatibility between overlay display, 1731 display and my graphic card as I am unable to reproduce on other device
Jean Pierre

[Gilletarom]

Hello Dave,

Originally Posted by Gilletarom

hello Dave,

Here is the result of a test with the NV-Verlag charts.
...
2° I have 3 groups of charts: CM93, S63 and NV-Verlag.

First, using chart group, I display 63 S chart goal I am on Guernsey. So, nothing. Normal. Then I display NV-Verlag, using chart group. And a hole appear on the chart. It is necessary to click on the bar chart to resolve this problem. See screen copy A, B and C.
...
B.R. Gilletarom.

We have some S57 charts that SHOM provide for test.

So this morning I load these S57. I create a group SHOM and I test from S57 group to NV-Verlag group.

And I have the same problem with hole as with S63.

B.R. Gilletarom

[seandepagnier]

Dave
The setting dialog font problem appends on my two devises . A HPCQ60-Vista and a ASUS-W8.1 10'

But it's another very annoying problem I faced two times very recently : a windows crash and the lost of the entire tracks.

Pitzef,

I am with you here. I lost all my tracks many times. Not always from crashes (although I do tend to crash frequently because of developing) but also from the battery running out of power, and also from buggy drivers locking the computer.

I suggest we fix this problem forever by appending each new track point to the file on disk as it is added. A single point should be a fast operation (a rewrite of the entire track every time is too slow)

Originally Posted by bcn

Ptizef,

you might check if there is a memory leak - memory usage is creeping up slowly and finally crashes O. Exact time until crash depends on charts and usage.
We have two PCs with faulty drivers for OpenGL which show this behavior.

Resource Monitor or Task Manager.

Hubert

I found with valgrind that the portable ocpnregion class has a small memory leak. There are a lot of small leaks, but it is difficult to narrow them down (I had a leak report which was over 100 pages long) because mostly they are caused by libraries opencpn uses (intel graphics driver, wxwidgets etc...) and also a lot of memory is not freed upon exit.

[jonasaberg]

Originally Posted by boat_alexandra

There are a lot of small leaks, but it is difficult to narrow them down (I had a leak report which was over 100 pages long) because mostly they are caused by libraries opencpn uses (intel graphics driver, wxwidgets etc...) and also a lot of memory is not freed upon exit.

What is the best way to handle this as an end user? Reboot for breakfast? ;-)

[Gilletarom]

Hello Dave,

Originally Posted by Gilletarom

Hello Dave,
We have some S57 charts that SHOM provide for test.
...
And I have the same problem with hole as with S63.
B.R. Gilletarom

Hole also with before, raster Argentina, then NV-Verlag.

[bcn]

Originally Posted by jonasaberg

What is the best way to handle this as an end user? Reboot for breakfast? ;-)

Well, in our case we disabled OpenGL.... at the system on the boat.

Not the solution to optimize raster charts perhaps. And not for low-end systems.
A difficult situation for the OpenCPN hardware zoo. Especially under Windows systems and there the OpenGL drivers.

It might show up that there are (unpredictable) HW/SW combinations that are stable and others not.
So when you find that your system and your configuration is solid rock, be happy.

Heavy testing for the moment to get a bigger picture would be my recommendation.

For the rest: stay at stable (3.2.2) in parallel and sail

Hubert

[bdbcat]

Gilletarom....

May I please see your OCPN logfile immediately after showing the NV-Veralg chart quilt with "holes" in it?

Thanks
Dave

[bdbcat]

jp...

re:

Related to the display problem (#125)
I've found when it happens :
Displaying a colored overlay map in grib-pi immediately....

To save me some time, could you please attach a colored GRIB overlay that causes the referenced problem?

Thanks
Dave