OpenCPN Beta 5.5.1114 (RC-2) Released

[bdbcat]

Hur450502...
OpenCPN 5.5.x is not officially supported on RPi Bullseye (aka Debian 11).
The core program opencpn probably runs reasonably on Bullseye, but no plugins for that platform will be available for some time.


I suggest you stay with (the much more stable) Buster/32 for actual navigation with OpenCPN.


Thanks
Dave

[bdbcat]

CarCode, and other parties truly interested in enc object encoding/rendering...


re:

For people without reading glasses I enlarged the screenshot.
Shown is the question mark (?) as a top mark in this example.

In this screenshot, if you will inspect the attributes of the TopMark fully, you will see that the information encoded for that TOPMAR object in the ENC cell is:

Color: red
Inform:  triangle, point up.

This object, like many others on cm93, is encoded incorrectly according to S57 specifications. What is missing is any "TOPSHP" attribute. TOPSHP is the only reliable method to determine rendering of a TOPMAR object. Arbitrary parsing of INFORM attributes to "guess" a rendering shape is not supported nor required by S52, as it would not be considered reliable.
Thus, OpenCPN renders a "?" for this object.



If you can find a TOPMAR with a TOPSHP attribute defined that is rendered incorrectly, that would be a legitimate bug. And I would be happy to inspect and correct any such errors found in any enc, including cm93.


Thanks for the report
Dave

[CarCode]

Originally Posted by bdbcat

This object, like many others on cm93, is encoded incorrectly according to S57 specifications. What is missing is any "TOPSHP" attribute. TOPSHP is the only reliable method to determine rendering of a TOPMAR object. Arbitrary parsing of INFORM attributes to "guess" a rendering shape is not supported nor required by S52, as it would not be considered reliable.
Thus, OpenCPN renders a "?" for this object.

You should explain also why the o-charts are only shown correct when OpenGL is disabled:

Originally Posted by Aldebaran17

Problem solved: deactivating OpenGL makes all nice again. I believe I had to turn it off with the current official version (5.2.4) as well. So, seems to me that the RC2 version has OpenGL turned on by default, which may not be good for Macs.
Attachment 248429

[Hakan]

Originally Posted by AedanC

Hakan,
It has failed now in the same manner with OpenGL off.
Aedan

Since it happens for you on a specific PC and not on another it may be related to that machine rather than OCPN itself? Neither can I reproduce this nor do we have similar reports from others.
So what could we test more?
1. Check the opencpn.log (and opencpn.log.log) file for oddities. That's something like extensive entries flooding the file. Is it big? (That's more than 1 MB) (One example is the plugin radar_pi set to a verbose debug level.)

2. I have once seen a config file oddity giving similar effects. (One row was a million chars long)
To test that I suggest:
Note your settings like charts and connections. (And/Or use a template)
Stop OCPN.
Rename the config file so you have to "start from scratch". C:\ProgramData\opencpn\opencpn.ini to e.g. opencpn.ini.org
Start OCPN, and add new chart locations, connections and install/activate used plugins (After a master catalog update).
A saved template can also make your life easier. On the other hand a clean config(ini) file is a good training to handle OCPN.

[stelian]

Hi Dave,

Originally Posted by bdbcat

OpenCPN 5.5x is officially supported on Ubuntu LTS releases, in this case 18 and 20.

We will address bugs found in later Ubuntu releases after the Release of OCPN 5.6.0.

Ok, sure, I can understand this. But since I have seen a 21.10 build in the PPA, I thought it was supported (or at least usable...)

I'll report my issues again once 5.6.0 is out.

Stelian.

[HUR450502]

Originally Posted by bdbcat

Hur450502...
OpenCPN 5.5.x is not officially supported on RPi Bullseye (aka Debian 11).
The core program opencpn probably runs reasonably on Bullseye, but no plugins for that platform will be available for some time.


I suggest you stay with (the much more stable) Buster/32 for actual navigation with OpenCPN.


Thanks
Dave

Dear Dave, thanks a lot for your fast answer. Currently I am back on Buster. However, can you gve me a hint how to get back to pre 5.5.1114?


I didn't intend to install a beta on my nav computer, but - being in a harbour - I wanted to update the system just for actiality and security and ended up in numerous error messages "assertion failed", ... etc.



Or is there a way to get around these problems with 5.5.1114 some way?



The problem her is in my opinion, that you automatically get a beta release installed - unintentionally - when you just want to update e.g. the plugins.



Any help would be most appreciated and, sorry, I didn't want to "conquer" this thread with foolish Linux illiterate questions .



Best regards, Uwe

[Ptizef]

Hi

surface GO windows 10
Enable touch screen on
test touch screen without mouse
I was testing a self compiled current master version and I found that the zoom in/out with two fingers is partially broken.
1) when "smooth panning/zooming is set
When I try to zoom ( I put my two fingers on the screen and move its closer or further) nothing appends . The zoom append effectively when I remove my fingers from the screen. This make zooming practically unusable.
1) when "smooth panning/zooming" not set
In this case the situation is better and zooming seams to work correctly. In addition it is possible to zoom and drag in the same movement and this is good. Nevertheless, there are two problems
The first is that there is something like a tempo between the first movement and the response.
2) the second and worst is if the two fingers are moved first laterally, nothing will happen then. I have to remove my fingers and retry. This is difficult in rough condition

Reverting to the "official" beta 1114 is the same
I don't know when this appended, but reverting to 5.2.4 and zoom is back to the old behavior
Sorry for my bad explanation and I hope you understand what I mean
JP

[CarCode]

Originally Posted by bdbcat
This object, like many others on cm93, is encoded incorrectly according to S57 specifications. What is missing is any "TOPSHP" attribute. TOPSHP is the only reliable method to determine rendering of a TOPMAR object. Arbitrary parsing of INFORM attributes to "guess" a rendering shape is not supported nor required by S52, as it would not be considered reliable.
Thus, OpenCPN renders a "?" for this object.

It well known CM93 does not follow the S57 specifications in any case. Therefore the OCPN code uses e.g. a file called cm93.cpp
If you don't want to support CM93 anymore you should announce that and delete its special functions.

[Ptizef]

Hi

The zoom behavior described in my former message concerned GL mode
In no GL mode:
If "smooth panning/zooming" is set, the same as GL mode
If "smooth panning/zooming" is not set, I can zoom and drag but as soon as I remove my fingers, zoom level and drag position are back to the former value
JP

[bdbcat]

HUR450502...


For rPI Buster/32 bits, the current standard repo offers version 5.5.1116. This should be OK for you.


Thanks, and sorry for the run-around.
Dave

[bdbcat]

CarCode.


re:
"You should explain also why the o-charts are only shown correct when OpenGL is disabled:"


This is an acknowledged bug in the oeSENC plugin, under investigation.
FYI, it only occurs in MacOS, on MacBooks with Retina displays.


Dave

[AedanC]

Originally Posted by Hakan

Since it happens for you on a specific PC and not on another it may be related to that machine rather than OCPN itself? Neither can I reproduce this nor do we have similar reports from others.
So what could we test more?
1. Check the opencpn.log (and opencpn.log.log) file for oddities. That's something like extensive entries flooding the file. Is it big? (That's more than 1 MB) (One example is the plugin radar_pi set to a verbose debug level.)

2. I have once seen a config file oddity giving similar effects. (One row was a million chars long)
To test that I suggest:
Note your settings like charts and connections. (And/Or use a template)
Stop OCPN.
Rename the config file so you have to "start from scratch". C:\ProgramData\opencpn\opencpn.ini to e.g. opencpn.ini.org
Start OCPN, and add new chart locations, connections and install/activate used plugins (After a master catalog update).
A saved template can also make your life easier. On the other hand a clean config(ini) file is a good training to handle OCPN.

Indeed it could be something to do with the machine, that had occurred to me.

I have renamed the .ini and the navobs files and restarted. No charts, no AIS feed. After a few hours it had failed again.

I will be away now until Tuesday, any suggestions for what to try next would be welcome. Would it be worth looking for graphics driver updates?

Best regards,

Aedan

[Hakan]

Aedan..
Please post your log file here. (Rename it like opencpn.log.pdf to fit the Forum rules)

[HUR450502]

Originally Posted by bdbcat

HUR450502...


For rPI Buster/32 bits, the current standard repo offers version 5.5.1116. This should be OK for you.


Thanks, and sorry for the run-around.
Dave

Dear Dave,

I installed again 5.5.1116. However, I am getting still these error messages (see attached screenshot), which led me to completely do a new installation (which then got me "Bullseye") Now I am back on my "old system", ("Buster/Bionic").

I can overcome this situation however, by pressing 30!! times "continue", or un-checking "Show this dialog next time".

This way I can use 5.5.1116, however, I can no longer update any plugins. I get to the "update Catalog (Master), I get a list of updates (which are important to me), however, when I want to do the update(s), nothing happens.

Do you have any hint for me how to get around this? Do you need any more information? If yes, which?
Thanks again for your help and understanding.

Best regards, Uwe

[bdbcat]

Uwe...
Please, what is your basic OS?


Please post results of:

$ uname -a

and

$ cat /etc/os-release

Thanks
Dave