Ocpn-Draw and Boundary

[Hakan]

Jon
I didn't read all your text carefully.
I've now set to allow 2 navobj. and here are the two from one stop/start of O.
Håkan

[jongough]

Håkan,
It seems to be a windows related issue. I have tried moving my boat to your location using the CM93 charts and followed what you did and it all worked OK. I am fighting windows at the moment as the new version of OCPN seems to have changed the way the application is built and I cannot get it to work.

Do any other ocpn_draw_pi objects display the same error, i.e. boundaries?

Jon

[Hakan]

Jon
Boundaries are fine and no obstacles at O stop/start.
Anyone else running W and can make a EBL test, Rick?

If you don't find anything I could set up for debug of O-Draw but I'll be to busy for that a couple days from now.
To be continued..............

Thanks
Håkan

[rgleason]

Hakan, I'll be back tomorrow and will try it too. Have to move to new version and catch up with git.

[Hakan]

Jon, Rick
Pls read http://www.cruisersforum.com/forums/...ml#post1959175
It could be a trace.

Håkan

[jongough]

Hi,
I have just put up patch 530. This one provides three types of boundaries:

• Exclusion boundaries with the interior filled

• Inclusion boundaries with a boarder round the outside (non-opengl only at the moment)
• Neither boundaries which only have the lines

This will now give a visual (non-opengl) indication of the type of boundary. The boundary properties and the ocpn_draw_pi properties have been updated to allow the specification of the type of boundary and the width of the inclusion boundary boarder.

Inclusion boundaries will look weird if you have any of the boundary lines crossing each other as the fill pattern goes in the wrong place. I am not sure I can do much about that, sorry. BUT, should any boundary really have self crossing lines? They can be drawn, but......

I am working on the opengl version now and hopefully will get it working soon.

Jon

[jongough]

Hi,
Patch 535 now includes opengl changes to show inclusion boundaries as well as some other fixes for these types of boundaries.

Jon

[jongough]

Hi,
Patch 540 is the OD side to enable the watchdog plugin to use inclusion boundaries as an anchor watch. I have submitted a pull request to the watchdog plugin with the needed changes to make this work. In the watchdog, under boundary alarms, is the ability to select an inclusion boundary, and get the GUID if the boat is within the boundary, and then to raise an alarm if the boat moves outside of the boundary.

Jon

[Hakan]

Jon.. for info...
The OCPN patch to fix mixed config files didn't made any difference for the "extended" EBL issue, as probably expected. Neither the last O-draw fetch, also as expected I suppose.
You don't have any new ideas have you? And is it still only me seeing this? - Any more Windows users here?

Håkan

[jongough]

Håkan,
I found the problem and fixed it with patch 544. It was to do with persistent EBL's with non-fixed end points, such that it moved with the boat. At the time the first boat position update occurred there was no previous location information to work with, so you had a doubling up of the lat/lon of the end point, hence the LONG EBL.

Jon

[Hakan]

Jon..
Yes you did! Very good!

One observation:
When a persistent EBL is active you continuously append to "ODnavobj.xml.changes".
My file growth to 1 325 kB in twenty minutes with the same single EBL as on my previous pictures. If I by accident sail with that for a day or two the file will be very big? (OK - not very realistic but, you know, everything can happen.)

Thanks
Håkan

[jongough]

Håkan,
The issue here is that the EBL is not really designed to be persistent. I put that in more to help with crashes and for completeness. I will look and see if I can stop the updates going to the xml files whilst still allowing a restart to give you a correct EBL.

Jon

[bcn]

Jon,

as you said: EBL shall not be persistent (but for a crash..)
Being an active bearing.

The one you implemented is much more.

Hubert

[Hakan]

OK - I can agree to that and doesn't see any particular reason to have it persistent. If I need an EBL it's mostly to check if a another ships bearing is changing or not. There is no need to get it back if I close OCPN. But why was the persistent option introduced?
Maybe you can take that option out and save some work?

I do appreciate this plugin and the EBL feature. Not less for the radar plugins.

Thanks
Håkan

[rgleason]

Jon

It was to do with persistent EBL's with non-fixed end points, such that it moved with the boat. At the time the first boat position update occurred there was no previous location information to work with, so you had a doubling up of the lat/lon of the end point, hence the LONG EBL.

Hakan

When a persistent EBL is active you continuously append to "ODnavobj.xml.changes".
My file growth to 1 325 kB in twenty minutes with the same single EBL as on my previous pictures. If I by accident sail with that for a day or two the file will be very big? (OK - not very realistic but, you know, everything can happen.)

Jon

The issue here is that the EBL is not really designed to be persistent. I put that in more to help with crashes and for completeness. I will look and see if I can stop the updates going to the xml files whilst still allowing a restart to give you a correct EBL.

Hakan

OK - I can agree to that and doesn't see any particular reason to have it persistent. If I need an EBL it's mostly to check if a another ships bearing is changing or not. There is no need to get it back if I close OCPN. But why was the persistent option introduced?
Maybe you can take that option out and save some work?
I do appreciate this plugin and the EBL feature. Not less for the radar plugins.

Bcn
as you said: EBL shall not be persistent (but for a crash..)
Being an active bearing.
The one you implemented is much more.

============
This is the easiest way out of course, but why does this have to happen?

• "When a persistent EBL is active you continuously append to "ODnavobj.xml.changes" ...file growth to 1 325 kB in.."

If an EBL is attached to the boat, it must append to ODnavobj, but can't it do it with some pause, and work similar to tracking, but you certainly do not need to save the old locations, which you probably don't. So it is probably some memory problem that makes it grow. Maybe Carcode could advise in more detail what the problem is. Then perhaps the persistence issue could be resolved.

If an EBL is not attached to the boat, its points are simply stored there. These EBL's certainly would be more useful with an option for persistence.


I think that in order to keep persistent EBL when not attached to the boat, the persistence EBL attached to the boat would also needs to work.

I had always considered persistence was a good feature, despite the reasons Jon made it, and I do think it is useful when it the EBL is not attached to the boat for example, ..for marking Dead Re-conning locations, details, etc. -- Just my view, it is supposed to be a tool, and if it can be done, I think people would find it useful.

----
I did compile and uploade Ov4.1.1108 and Ocpn_Draw, Watchdog and Weatherfax to Opencpn Beta File Thingie and have been trying out the boundary features for inclusion and exclusion. First tested with Opengl off.

1. EBL lines do not match up with the VRM, they come up a little short.

2. EBL lines seem to have the inclusion, exclusion, neither selection, is this right?

3. When I select neither, the boundary outline disappears on windows machines, although it is there still but transparent. I can still select it with a right click and change properties, for example.

4. I tried to get the watchdog alarm feature working, with an inclusion boundary.
A. Set a boundary around the boat, and picked inclusion,
B. Went into Watchdog and got the guid from the Boundary > Properties
C. Went back into Draw_pi and pasted it into the "Boundary" Guid slot.
D. Right clicked outside the boundary and selected Navigate to here.
E. Then clicked on the Red boat to set the new route to active.
F. Then I tried moving the boat by right click and select Move boat here and getting the alarm.
G. Then tried moving Boundary points closer to the boat to reduce the distance from the boundary and get an alarm.
H. I suppose I should find a brief Nmea file and run that with VDR, since I am trying to simulate. What do you do Jon?
- I should be able to get this going I think.
- Also should I try it with Opengl tomorrow?

- I think there needs to be better screenshots in the documentation and it end part needs to be finished and edited, which I will try to get done.