OpenCPN Version 2.4 Beta Build 506

[bdbcat]

Ahoy OpenCPN'ers...

OpenCPN Version 2.4 Beta Build 506 will soon be available for download from opencpn.org.

Monitor the opencpn.org Downloads page to find and download the latest Beta version.....

Source is available now on sourceforge git repo.


Changelog
69. Correct AIS BlueFlag decoding (FS#433)
70. Correct AIS decoding of targets with invalid parameters (FS#424)
71. Correct some errors in AIS target reporting (FS#425)
72. Correct AIS target list sorting and Range display. (FS#430)
73. Shift ScaleBar to right when grid is displayed. (FS#409)
74. Correct rendering of rotated/greyed user defined ownship icon. (FS#422 FS#423)
75. Correct some README file names/locations.
76. Correct raster chart rendering on extreme overzoom. (FS#428)
77. Correct text buffer overrun problem on vector charts seen on some AMD-64 systems.
78. Correct cm93 quilted line object "smearing" at some locations and scales.
79. Correct auto-pan on route create.
80. Correct active track rendering at very large scale.
81. Correct logic for chart stacks larger than 30 charts. (FS#316)
82. Implement new floating transparent toolbar. Thanks to Will (Manimaul) for the ideas.
83. Redesign right click context menu layout (FS#256)
84. Various CMake and PlugIn patches, thanks to Jean-Eudes.
85. Add shift key direction flip control to F6 dimmer.
86. Correct positioning of long rollover strings (FS#299)
87. Correct AIS target timeout logic (FS#338)
88. Add Traditional Chinese language support

Release Beta Version 2.4.506

Notes:

82. This is new. Note the "grabber" symbol at the far right side of the toolbar. Use this handle to drag the toolbar wherever you want it.
It "snaps" to the edges. Right click on the grabber, and the toolbar shifts to vertical format. On Windows (at least), the toolbar dims down transparently until you roll over it. Have fun....

85. Shift key with F6 reverses dimming direction, allowing easier control. Find just the level you want.

Other notes:
No good results or feedback have been obtained from intermittent crashes on cm93 quilted pans.
Please test this aggressively, but in simple and easily reproduced scenarios if possible.

There are some language file updates for this Beta version.
With luck, we will get all the .po files updates for the next Beta. Keep them coming...

Thanks, and Good Luck
Dave

[nohal]

Dave...
Works pretty well so far, just a minor observation of behavior I would not expect - the toolbar does not snap to the very edge of the screen at the right and bottom. Also that it ignores the GPS status bar and covers it feels a little unfriendly. I personally could live without the toolbar turning transparent, it gets a little lost.
While you have a docked plugin window and resize it, the toolbar does not stay glued to the edge of the chart viewport and overlaps the plugin.

It looks like we have a new memory leak somewhere, will do some Valgrinding later on Linux to investigate it more.

Pavel

[cagney]

Thank you Dave & Team for the new beta.
Some very interesting features in there.
Some initial observations from compiling and very light testing.
For compiling on Ubuntu 11.04, this line still needs to be added in CMakeLists.txt after line 84

  ADD_DEFINITIONS("`pkg-config --cflags --libs gtk+-2.0`")

To compile a package on amd64 one change is needed. line 75 -77 in CMakeLists.txt must be changed to this.

IF (NOT ARCH)
SET(ARCH "amd64")
ENDIF (NOT ARCH)

Which I think means that Cmake is not correctly identifying amd64, actually don't find an architecture at all. The default, above was i386.

Nice floating toolbar, Will.
I found one problem immediately.
The WMM plugin button doesn't work, this plugin instead comes on when clicking the "Change Color Scheme" button.
The WMM plugin also triggers a repeatable crash.
If the plugin is on. Go to the toolbox and disable it. Click OK.
Go back to the toolbox and enable the WMM plugin.
Instant crash.


The Ubuntu packages are now available on opencpn.org.

Thomas

[nohal]

Thomas,
did you build the WMM plugin from source or did you use the precompiled library? Can't reproduce on Win, will have a look on Ubuntu later.

Pavel

[cagney]

Originally Posted by nohal

Thomas,
did you build the WMM plugin from source or did you use the precompiled library? Can't reproduce on Win, will have a look on Ubuntu later.

Pavel

Pavel
I think I compiled it my self. I just downloaded the 64 bit version from opencpn.org. Same problem.
I could also repeat this on my 32 bit Ubuntu 10.10, with an "old" downloaded version.
Any tests I can do?

Thomas

[balp]

Originally Posted by bdbcat

Please test this aggressively, but in simple and easily reproduced scenarios if possible.

Basic testing on Ubuntu 11.04, looking good, with Raster, bsb,, cm93, and S57 charts. the new toolbar, the unity window manager wuil make for a great onboard experience. Not sure i have time for a real try at the boat this weekend thou.

[cagney]

Interference between tide & current stations.

Pos ~ 048 32.9913 N 122 59.9171 W.
To get the tides in the picture below, currents must be turned off.


Which leads to this picture. From testing last beta I drew the conclusion that the blue was water, and the hight of the blue column represented the state of the tide. If that was the case the blue would be much taller i the picture below as it's nearly high tide. What does the relation between yellow and blue mean?


Thomas

[bdbcat]

Folks....

There is some problem with the WMM PlugIn, or its connection to OCPN.

For testing of 506 now, please remove(or temporarily rename) the WMM PlugIn from the shared directory.

We'll have more to say on this after analysis.(Pavel...)

Thanks
Dave

[nohal]

All...
Looks like we are not ABI compatible with previous versions - all plugins built against them seem broken - on Win causing immediate crash on plugin init, on Linux weird behavior activating all the "old" plugins on color scheme change.
Rebuilding the plugins against .506 seems to solve the problems. Until we rebuild them, get rid of all plugins but GRIB and Dashboard which are part of the distribution, thus rebuilt already.
WMM plugin segfault while loading the magnetic data confirmed as well.

Pavel

[nohal]

Testers...
I've rebuilt the WMM plugin against .506 and corrected the crash while repeatedly re-enabling the plugin. Download links at PlugIns | Official OpenCPN Homepage updated.
Note that the filename has changed for i386 Linux, so you have to delete the old libwmm_pi.so manually not to have both the old and new plugins.
Let me know if you observe any other weird behavior with this new version.
The rest of the plugins will still not work as expected and I will post the updated versions as I'm ready to release the codebase.

Pavel

[cagney]

Thank you Pavel.

Things back to normal with WMM

Thomas

[ChuckSK]

There is still an active debug code break in cm93.cpp line 1604.

Chuck

[Roger Hutchinson]

Love the software but I have a problem with the GRIB plug-in. I click on it but nothing happens. Normally it would open the GRIB file, allowing you to select the file you want to overlay. Can any on tell me what I have done that prevents the use of this app?

[nohal]

Roger,
hard to say, everything looks normal here. Try to have a look in the logfile and post it's contents if they seem to contain something related. Also your ini file and info about which version of OpenCPN and on which operating system are you running could be pretty helpful.

Pavel

[bdbcat]

Pavel....

Thanks for looking into this PlugIn problem.

Any idea how this happened? We bumped the PlugIn API to 1.5, added a new method, and implemented a stub function in pluginmanager.cpp.

Apparently, that changed the base program exports enough to break things.

And, as I sit here thinking, about it, I think I see.
In fact, that's probably exactly what happened.

We added another public method to the base class opencpn_plugin, namely

virtual void SetAISSentence(wxString &sentence);

And we put its definition in the middle of the existing method definitions, which seems to have scrambled the base class vtable.

I wonder if it would have worked if we had added the new method definition to the end of the existing defns?

This is an important point. We really need to preserve downward ABI compatibility as we evolve the PlugIn API.
Said simply, new OCPN versions should not break existing PlugIns.

I'll experiment some, and report back....

Thanks again
Dave