S-63 plug-in for OpenCPN

[yachtman]

Excuse my ignorance but how do you change the update file that is here a few pages back as a dpf file into an exe one as it persists in remaining a pdf file (win 7)

have tried rename and via proprieties, but wont run as an application.

Thanks

[bcn]

In W7 you might have to change the "hide know file extensions" (or similar - don't have an English version of W7 at hand) option in the Windows Explorer settings in order to get rid of the .pdf extension.

The resulting OCPNsenc.exe copied into in S63-pi folder runs then perfectly.
You don't have to execute it - this is done by the program itself.

[yachtman]

Thanks for that, yes works now, but I think my acer note book is not man enough to load half the coast as it has paused part loaded.

[bcn]

Can you give a bit more information where you think the poor PC stopped (our XP is about 5 years old and takes its time as well when importing S-63 files):

Checking the userpermit ok?
Importing cell permits?
Then panning to the Bristol Channel (8a set)?

[yachtman]

Hi

It struggled on loading south and east coast, bristol bit went at a better speed.

permissions seemed quite quick, qbout the same as loading the dutch free encs. Not a great detail though, feels as if an inner layer is missing.

Once loaded, on a restart they re load very quickly. with no problem panning etc

[bdbcat]

Gilletarom...

Regarding your new security warning for windows:

I am not familiar with this message, and what Windows internal code condition could produce it.

1. Is this new with OCPN 3.3.1303?
2. Does this warning come up with no PlugIns activated?
3. Which PlugIns cause the warning?

The base OCPN does never communicate with the outside internet, unless a particular network connection is established to do so, as for remote AIS monitoring.

As far as I know, no PlugIns directly access the internet. Certainly none I am responsible for.

For sure, s63_pi does not. S63_pi establishes a localhost socket connection between OCPNsenc.exe and the PlugIn to communicate status. This is called "backchannel" in s63_pi the debug screen. There is no other TCP/IP or UDP connection established by this PlugIn. OCPNsenc can function as a SENC server on a network, but this mode is not used in s63_pi.

Perhaps there is Windows guru who can enlighten us on this firewall behavior?

Thanks
Dave

[bcn]

Originally Posted by bdbcat

For sure, s63_pi does not. S63_pi establishes a localhost socket connection between OCPNsenc.exe and the PlugIn to communicate status. This is called "backchannel" in s63_pi the debug screen. There is no other TCP/IP or UDP connection established by this PlugIn. OCPNsenc can function as a SENC server on a network, but this mode is not used in s63_pi.

Perhaps there is Windows guru who can enlighten us on this firewall behavior?

Thanks
Dave

Dave,

searching Internet I found a statement, that just opening a port for listening will trigger the firewall warning. No need to try to connect to the Internet.

And I do remember that the warning popped up at my W7 machine as well.
Don't remember when exactly, though... Has been during last 10 days for sure.
And when you allow it once it will not appear further (firewall rule)

Hubert

[bdbcat]

Sean, et-al...

Regarding security of S63 chart cells in the s63_pi:

As noted, all the cell permit, decryption, and decompressing occurs in the OCPNsenc "helper" application, which is closed source. There are a number of mechanisms employed in OCPNsenc and the PlugIn to prevent chart "leakage", none of which I will describe here

I the final Release, we will ensure that SENCs and S57 cells never appear in clear-text as disk files, and that eSENCs will not be portable to machines other than the one on which they were created.

As part of the Alpha and Beta testing, it would be useful for those with a "hacker's" inclination to take a look around at the s63_pi source, and try to find simple vulnerabilities. Communicate with me privately if any are found, please.

Finally, I should note that no DRM scheme is immune to dedicated attack, and so it is with S63. But it is a pretty effective scheme for what it does.

Thanks
Dave

[bdbcat]

Hello testers....

On performance:

1. eSENC file creation is a bottleneck. This is not due to S63 decryption/decompression. Those are performed in memory for security reasons, and are quite fast. The bottleneck is converting lat/lon descriptions of AREA features into renderable primitives. In this area, the s63_pi should have performance comparable to OCPN with standard S57 cells.

2. There is complex logic required to determine if a previously built eSENC is still valid. If invalid, it is rebuilt. We would like to know of cases where you think that an eSENC has been rebuilt in error, i.e. when no change was made to the chart configuration. This would be a bug.

3. Rendering bugs: There are several, including depth contour rendering, selection of SAFETY contour, and obstructions such as wrecks. Also maybe trouble with SCAMIN interpretation. All of these issues will require a new OCPN core Beta. The next OCPN Beta is proceeding in parallel, and will be out in a week or so, along with a new s63_pi.

The next s63_pi will have some changes in the GUI, based upon feedback and real use scenarios encountered so far.

Please remember, this is Alpha testing. We don't even know how many wheels this car will have, or whether it steers by wheel, footpedals, or tiller. The paint colour and cupholder discussions will come later, after we move to Beta testing.

Thanks for your patience and input

Dave

[seandepagnier]

Originally Posted by bdbcat

Sean, et-al...

Regarding security of S63 chart cells in the s63_pi:

As noted, all the cell permit, decryption, and decompressing occurs in the OCPNsenc "helper" application, which is closed source. There are a number of mechanisms employed in OCPNsenc and the PlugIn to prevent chart "leakage", none of which I will describe here

So you must provide binaries for every supported platform. This explains a lot, I thought this plugin was completely open source.

As part of the Alpha and Beta testing, it would be useful for those with a "hacker's" inclination to take a look around at the s63_pi source, and try to find simple vulnerabilities. Communicate with me privately if any are found, please.

This one is so obvious it might as well be public, and will always exist:

a plugin or other program could which automatically moves the viewport and creates raster charts from screen grabs of the rendered vector charts. I was considering implement this for performance improvements (so caching tiles of vector charts as raster) but it would defeat security as well.

Of course the resulting files would be huge and much more difficult to redistribute, but probably well within the storage capacity of modern drives.
Also would also not have as good quality for all scales and the vector chart specific features would be lost etc...

Maybe that's enough reason to not worry about it?

Originally Posted by bdbcat

Hello testers....

On performance:

1. eSENC file creation is a bottleneck. This is not due to S63 decryption/decompression. Those are performed in memory for security reasons, and are quite fast. The bottleneck is converting lat/lon descriptions of AREA features into renderable primitives. In this area, the s63_pi should have performance comparable to OCPN with standard S57 cells.

Working on a vertex shader program to convert lat/lon into mercator coordinates using the GPU. Of course we would need to pack the coordinates in vbo's or vertex textures or something.. the current scheme is not very efficient.

2. There is complex logic required to determine if a previously built eSENC is still valid. If invalid, it is rebuilt. We would like to know of cases where you think that an eSENC has been rebuilt in error, i.e. when no change was made to the chart configuration. This would be a bug.

3. Rendering bugs: There are several, including depth contour rendering, selection of SAFETY contour, and obstructions such as wrecks. Also maybe trouble with SCAMIN interpretation. All of these issues will require a new OCPN core Beta. The next OCPN Beta is proceeding in parallel, and will be out in a week or so, along with a new s63_pi.

Maybe my optimization support be merged into the next core beta as well? I grealy optimized text rendering as well as icons for vector charts. From what I can see vector charts are still very inefficiently rendered.

Really all the triangles's coordinates need to be put into an array and uploaded, then all the lines, then all the points etc... and arranged in blocks based on the color or texture or mode they need to be rendered. These arrays can be in tiles, but the coordinates should stay in graphics memory from frame to frame and only uploading needed tiles, not recalculated every frame as it currently is. In other words, either display lists (which probably use vbo's) or vbo's, not immediate mode. I think the frame rate could be ten times higher.

[Gilletarom]

Hello Dave,

Thank you for your reply to my post # 122.

I was naturally anxious to see this screen appear.

This is a screen that appears whenever I try to connect via wifi to a network in a hazardous environment. For example, if I am in a train station, an internet cafe, etc ...
In principle, without my knowing it really brings better, it reduces the risk of piracy of my communications.

But now I'm at my home and I do not use the web via wifi connection. I use a cable that connects me directly to the modem via ethernet link. Over the wifi card in my PC is disabled.

He had never appeared with OpenCPN. It appeared for the first time when I installed the plugin S63

This "abnormal" screen appeared for the second time when I add and I enabled WMM plugin on the last installation I did after cleaning the machine.

When this screen appears, there are three possibilities:
a) To say that the wireless network is a private network "on" and then "Ok"
b) To say that the wifi network is a "public network" and then "Ok". So no security for the transfer of information but it seem that some securities are enable by windows.
c) Cancel

The first time this screen appeared, I said "private network". The S63 is installed and run correctly. I do not inactivate it.

The second time I "canceled". And WMM plugin was not activated.

Yesterday I started to enable WMM plugin. This famous window firewall is not returned. The plugin WMM is activated normally.

I do not know to tell you more, my knowledge being that of a basic user.

Just a doubt .... Re-installation, on my machine, after the great cleaning was done by downloading new O322 and O331303. And if the files had been tampered? Changed from your original on the server?

Best regards, Gilletarom.

[rgleason]

Is there any way to get mutt nick ( dam smart keypad) to look at this?

Sent from my SCH-I545 using Cruisers Sailing Forum mobile app

[bcn]

Originally Posted by Gilletarom

He had never appeared with OpenCPN. It appeared for the first time when I installed the plugin S63

This "abnormal" screen appeared for the second time when I add and I enabled WMM plugin on the last installation I did after cleaning the machine.

When this screen appears, there are three possibilities:
a) To say that the wireless network is a private network "on" and then "Ok"
b) To say that the wifi network is a "public network" and then "Ok". So no security for the transfer of information but it seem that some securities are enable by windows.
c) Cancel

The first time this screen appeared, I said "private network". The S63 is installed and run correctly. I do not inactivate it.

The second time I "canceled". And WMM plugin was not activated.

Yesterday I started to enable WMM plugin. This famous window firewall is not returned. The plugin WMM is activated normally.

I do not know to tell you more, my knowledge being that of a basic user.

Just a doubt .... Re-installation, on my machine, after the great cleaning was done by downloading new O322 and O331303. And if the files had been tampered? Changed from your original on the server?

Best regards, Gilletarom.

Gilletarom,

in the case of the S-63 PI the case should be clear:
OpenCPN 3.3.1303 opens a port in order to communicate with the plug-in.
Windows firewall alerts about this new port and asks if you want to permit this comunication - this does not mean that O is communicating with the Internet. (Once you allow O might do - you can see the rules of WinFirewall when you enter into the details at the Windows Control Panel).

Once allowed, the firewall will not ask again. If the plug-in changes the port, the alert will be raised again.

WMM should not provoke any alert. Pavel is the person to confirm this.

Hubert

[Gilletarom]

Hello Hubert,

Originally Posted by bcn

Gilletarom,

in the case of the S-63 PI the case should be clear:
OpenCPN 3.3.1303 opens a port in order to communicate with the plug-in.
Windows firewall alerts about this new port and asks if you want to permit this comunication - this does not mean that O is communicating with the Internet. .....

WMM should not provoke any alert. .....

Hubert

Thank you Hubert. This explanation of the internal communication to the PC through a port can be good.

About WMM I do not forget that S63 was enabled. It is perhaps S63 that triggered the opening of the window when I enabled WMM


B.R. Gilletarom.

[PjotrC]

Some hours looking at s63_pi...

S63_pi, Windows Vista, No OpenGL

Installation/permits issues:

1. Parallel install created "OpenCPN" directory, not "S63_pi" (as mentioned by Gilletarom)

2. Crash on panning the chart probably after cancelling of importing permits around Test 8c. Report sent. Cannot reproduce.

3. Test 7c fails on DS1: 2 errors "cannot find ENC cell base or update in specified exchange set...GB283100, 283200"

4. A message "done with loading cells" in the OPENSenc log would help

5. Win7 clean install with all default values - no problems.

- not very seriously, but I gave the next 2 cases a try:

6. Testing Portable Mode with the same User Permit -
Permit Ok, Cell Permits imported from portable device for Test 8a (3 charts), works OK.
Close Portable OCPN.
Call "regular" OCPN - hangs because of "other OCPNSenc.exe running", reinstallation needed, after reinstallation - s63_pi comes enabled, nothing works anymore
manual removal of opencpn.ini, and all s63 related data directories in ProgramData necessary, reinstall again

7. Foreign Userpermit added as New: clears the old permit, rejects the new one, program hangs, reinstallation with manual removal of opencpn.ini, and all s63 related data directories in ProgramData necessary.

Rendering issues:

8. ENC texts on cell boundaries are only partially shown, so not readable

9. Light descriptions (texts) positioned nearby do not appear/disappear in concert as scale is zoomed in/out. E.g. one buoy and a lighthouse goes away, another buoy nearby stays visible...

10. Lights specifications are never included in "Object Query" on lighthouses, buoys etc.

11. Object Query for the lighthouse at Riding Point, Grand Bahama (near 26 38N, 78 14W) yields the Pilot Boarding Place, which is 3 miles away, if scale not sufficiently large.

12. Strange lights rendering: the light circle is drawn in color, but often nothing inside, at central position. On s57 charts there is always a beacon or buoy or lightouse or some other symbol inside.

I understand not all cases in the IHO test set are relevant, since ECDIS compliance is not an objective here.

Fast work indeed, the software develops faster than I can look at it...

Cheers,

Piotr