Beta Test / Technical

[idpnd]

Originally Posted by HamishB

beyond the Gsocket bug, what shape are the packages in? do they pass the lintian tests? ie are they in good enough shape for me to add to the main DebianGIS package source SVN repo? where are the package sources? I only see the .deb download (yeah, I know I can unzip that) and see no sign of them in the opencpn SourceForge CVS repo.

Expertise overload! hehe. I had to look up lintian test first. The precompiled packages are mostly provided by the lead dev Dave who is currently on holiday. Second in command is "Psyches" on this forum and at sf.net. Implementation of the patch sounds like a great idea.

does the Ubuntu download at Download OpenCPN | work on Karmic as-is?

Yes, it does.

Btw just had a look around the GRASS project, looks exciting. There's been discussion regarding a universal overlay format for adding points of interest/satellite imagery/etc in a google earth KML/KMZ type manner. Perhaps you have any ideas on an advantageous format.

[cagney]

Thanks HamishB.
I went back and reviewed a few mails from Dave after the release of 1.3.4, when the Gsocket bug first appeared.
We were so close.....and then had to do with an ugly workaround.
Basically it was a typo that Dave and I missed!!
So here we go. I have attached an updated nmea.h (strip the .doc).
Use this to replace opencpn/include/nmea.h.
I have a clean compile with this change.
idpnd and others ...please test.

Thomas

[louzou56]

hello,
I use the latest version.
I compared the reception of AIS target with 2 programs: opencpn and seaclear.
Opencpn not find the "fixed target" such as semaphores to earth, while seaclear yes!

why?

Christophe POTHIER

[Psyches]

Originally Posted by cagney

Thanks HamishB.
I went back and reviewed a few mails from Dave after the release of 1.3.4, when the Gsocket bug first appeared.
We were so close.....and then had to do with an ugly workaround.
Basically it was a typo that Dave and I missed!!
So here we go. I have attached an updated nmea.h (strip the .doc).
Use this to replace opencpn/include/nmea.h.
I have a clean compile with this change.
idpnd and others ...please test.

Thomas

Thomas, thanks. The fix is now checked in on the sf.net cvs mainline (post-1.3.6). I'll watch for test issues and rollback if needed.

Mark

[HamishB]

Originally Posted by louzou56

I compared the reception of AIS target with 2 programs: opencpn and seaclear.

I am curious to know if anyone has compared AIS support against the new AIS decoder from Gpsd?

Hamish

[sredna]

Originally Posted by Psyches

Thomas, thanks. The fix is now checked in on the sf.net cvs mainline (post-1.3.6). I'll watch for test issues and rollback if needed.

Mark

I can't build at all now, no matter rerunning the configure script, no matter if i disable GSocket def in giotypes.h or not.

/usr/include/glib-2.0/gio/giotypes.h:120: fejl: ‘GSocket’ has a previous declaration here
In file included from /usr/include/wx-2.8/wx/gsocket.h:179,
                 from /usr/include/wx-2.8/wx/sckaddr.h:20,
                 from /usr/include/wx-2.8/wx/socket.h:24,
                 from ././include/nmea.h:119,
                 from src/chart1.cpp:255:
/usr/include/wx-2.8/wx/unix/gsockunx.h:40: fejl: using typedef-name ‘GSocket’ after ‘class’
/usr/include/glib-2.0/gio/giotypes.h:120: fejl: ‘GSocket’ has a previous declaration here
In file included from ././include/nmea.h:119,
                 from src/chart1.cpp:255:
/usr/include/wx-2.8/wx/socket.h: In member function ‘wxSocketError wxSocketBase::LastError() const’:
/usr/include/wx-2.8/wx/socket.h:118: fejl: ‘struct _GSocket’ has no member named ‘GetError’
In file included from src/chart1.cpp:265:
././include/grib.h: At global scope:
././include/grib.h:83: advarsel: ‘typedef’ was ignored in this declaration
make: *** [chart1.o] Fejl 1

[sredna]

Well, I checked out nmea.h r 1.24 and i can build again. But your workaround appears to not work.

I did remove config.log, do autoreconf and rerun configure, to no avail, with nmea.h r 1.25.

Anders

[idpnd]

Originally Posted by HamishB

I am curious to know if anyone has compared AIS support against the new AIS decoder from Gpsd?

Hamish

Interesting idea, AFAIK AIS is decoded internally with OpenCPN?

[HamishB]

Originally Posted by idpnd

I had to look up lintian test first.

(it's just the .deb package policy auto-checker, apt-get install lintian && lintian package.deb)

for the long version see the Debian New Maintainers' Guide

> > does the Ubuntu download at opencpn.org work on Karmic as-is?
> Yes, it does.

ok, in that case I can include it on the OSGeo LiveDVD & virtual machine without delay:
https://trac.osgeo.org/osgeo/browser...all_opencpn.sh

is it possible to get a PDF or Doxygen dump of the help-wiki? (It is likely the user will be running this on their laptop a long way from any reliable internet connection.) Or is the HTML from the Help menu good enough for all practical porpoises?

Btw just had a look around the GRASS project, looks exciting.

loads of low level power in there, especially when combined with GDAL tools. Unless OpenCPN plans to get GeoTIFF support (yes please!) I hope have a look at making the (any GDAL/GIS format) to BSB operation a bit more streamlined & automated. Right now it is rather.. awkward. gdal_translate and gdalwarp can do most of the heavy lifting already.

There's been discussion regarding a universal overlay format for adding points of interest/satellite imagery/etc in a google earth KML/KMZ type manner. Perhaps you have any ideas on an advantageous format.

my first reaction would be to go with GPX (via libexpat?) and be done with it.
Over in GpsDrive we have 2-3 main waypoint/POI systems:

1. the old: "way_*.txt" a simple tab delimited file with lat,lon,label.
1.5) IIUC there's a similar route.txt thing as a list of waypoints.
2) the new: waypoints are held in a small SQLite database
2.5) Points of Interest (POI) from the OpenStreetMap project are alse stored in a SQLite DB.
3) GPX waypoint import (in developement, GPX tracks and routes are already working)

(I'm not explaining it very well, take ver. 2.10pre7 or 2.10pre8svn for a spin for a better idea)

If possible, I'd recomment something already compatible with GpsBabel and/or GDAL's OGR. GPX fits that bill and is both extensible and easy to work with.

I don't think any of us would be working with 100,000s of waypoints, so file size should not be a problem. It it is you can always zip them and call it a KMZ file.....

Originally Posted by Psyches

Thomas, thanks. The fix is now checked in on the sf.net cvs mainline (post-1.3.6). I'll watch for test issues and rollback if needed.

as this is not a bug of your making per se, I wonder if it is not more appropriate to just patch the offending Ubuntu versions as needed? (the debian packaging system makes applying custom patches quite simple & automatic)

(if the patch can be made to work at all!)

regards,
Hamish

[cagney]

sredna & Psyches

Today....I have the same problem as sredna! Yesterday I managed to compile with the patch and without the normal workaround, don't know ...yet, what's up.

Psyches, it think you better revert to the old code on source forge.

Thomas

[sredna]

Regarding the GSocket problem - what about fixing it internally, by renaming the offending object in openCPN? Do you know of any external code using that header?

Regarding data exchange - I am all for using GPX! support is there, it is simple and mostly suitable for openCPN needs, and it is extendable if you need additional features.

[PeteCross]

Hello,

Is there anyone else out there compiling the Sourceforge CVS for MSW?
If so, have you seen different performance/characteristics, mainly in panning and zooming between the ready compiled version and the self-compiled?

I am beginning to think that the sources are not identical.

I am using the free VC2008 Express Edition, plus some ATL and MFC headers for the resource file!

/Pete

[Psyches]

Originally Posted by cagney

sredna & Psyches

Today....I have the same problem as sredna! Yesterday I managed to compile with the patch and without the normal workaround, don't know ...yet, what's up.

Psyches, it think you better revert to the old code on source forge.

Thomas

Done.

[Psyches]

sredna, Hamish, re: the ongoing GSocket problem - any easy solution that works across all OpenCPN platforms sounds good to me. Ideally that solution should be constrained to where the problem occurs, in linux. I'm linux-challenged: I don't currently have an env and my dev is all on Win, so I rely on forum folks to propose and test. It'd be great if the causal dependencies had their own permanent fix rather than tweaking OpenCPN. I assume this will happen at some point since this is a general issue between wxWidgets and other linux framework. But again if it's easy, I don't see a problem with a specific and safe OpenCPN change.

Mark

[Psyches]

Originally Posted by PeteCross

Hello,

Is there anyone else out there compiling the Sourceforge CVS for MSW?
If so, have you seen different performance/characteristics, mainly in panning and zooming between the ready compiled version and the self-compiled?

I am beginning to think that the sources are not identical.

I am using the free VC2008 Express Edition, plus some ATL and MFC headers for the resource file!

/Pete

Hi Pete, yes I do this. I haven't noticed any big perf problem when I compile, and I generally compile and run the debug version on VC2008EE as well. I haven't done a side-by-side comparison though. Feel free to post specifics on problems you see.

Mark