Version 1.3.5 Beta Build 1117 is now available.
1. Correct UWTROC symbolization for covered/uncovered rocks.
2. Correct BSB
Chart PLY values for DTM on non-WGS84 datum charts
3. Correct track last point logic, which was erroneously creating null track segs.
4. Correct symbolization of CM93 CTRPNT object. Requires new s57data files.
5. Experimentally correct Route->SendToGPS, new default routename in xxRTE sentence is '00'. Needs test...
6. Correct logic on startup with no chart database nor config file, in case user skips chart selection.
7. Improve cm93 symbolization, anchors and chains.
8. Correct COM ports
left open on termination and after SendToGPS.
9. Add additional timeout between NMEA
messages on SendToGPS.
10. Experimentally change NMEA
xxRTE message to use uppercase "C". May break some existing GPS
interfaces. Needs test...
11. Correct some compound cm93/S57 light description strings, such as VQ(6)+LFL W 10s 8m 5M
12. Correct RNC outline logic for non-rectangular charts
crossing 180 E/W.
13. Improve georeferencing algorithms for RNC's with small number of REF points.
14. Correct memory leaks
and generally improve performance in tracking function.
15. Remove yellow chart accuracy highlighting hints on tracks due to high performance penalty.
16. Allow only one "?" dialog.
17. Allow extreme overzoom, with on-screen notification.
18. Correct loss of wheel
zoom and hotkeys after minimization on MSW.
19. Correct occasional hang/crash on chart window resize.
20. Add support for $xxGLL and $xxVTG NMEA sentences.
21. Correct "MAX Detail Here" logic
22. Implement preliminary GRIB file support.
23. Force uni-core operation for multi-core processors under MSW.
: Do not un-install your existing Version 1.3.4. Simply run the new opencpn_beta_setup.exe. Your config and database will be preserved.
Most changes are self evident.
I hope you can recognize your own special bug
13. We now handle scanned/converted charts spanning large latitude range much better. Technically, we found that we need to consider the eccentricity of the prolate spheroid earth model for very small scale Mercator charts Eccentricity is not as critical for large scale charts.
As a side effect, we now handle charts with only 4 REF points much more accurately. It may be reasonable now to use only 4 REF points on charts of scale smaller than about 1:1MM. Please try this out if you are working in the chart sub-group.
17. Overzoom. You all by now know the real issues involved with navigating using extreme overzoom, especially on ENCs. That said, due to popular request, here it is...
22. GRIB: This is an initial cut at GRIB file support. The main purpose here is to ensure that the largest number of GRIB file types and sources can be supported, and to start the inevitable discussion surrounding the UI and display rendering.
This Beta decodes and displays wind
barbs and pressure isobars at sea level.
The code is based on the zyGrib project
a. Enable GRIB icon in Toolbox->GRIB. (Thanks manimaul!)
b. Select GRIB icon.
c. In the GRIB dialog, select the "folder" icon and specify the directory in which your GRIBs are located.
d. In the file tree, navigate to a GRIB file of interest, and drill down to the forecast
date you wish displayed.
e. If a GRIB file is shown as grey-ed out, it means that this GRIB file contains no wind
or pressure data, or is unreadable.
As I said, this is a first stab at GRIB support for opencpn
. Comments on UI and display are welcome.
Please keep in mind that opencpn is an ECS, not a GRIB viewer. GRIBs on opencpn are intended to be a convenience for passage
planning. Other programs will be better choices for analysis of oddball GRIB data models and slick graphics.
23. Multi-core processors. We have reached the un-verified conclusion that using multi-core processors with some (maybe all?) virtual port emulator programs causes corruption of the C run-time memory heap. This bug has been, and still is, tough to pin down.
While technically not a problem within opencpn, a crash is a crash. So we make this change.
Performance is not an issue. Multi-cores are only used in opencpn for the threads handling the raw NMEA and AIS
data sources. The mainline code is single-threaded, so will only ever run on one core
Those users with multi-core processors who have had unexplained crashes with AIS
or NMEA data coming in through a virtual port are encouraged to try to break this Beta by any means you can think of.
Well, there it is.
Good luck with the testing. Please confine 1.3.5 Beta feedback issues to this (new) thread, to help keep things as organized as we can.
Thanks in advance for all your feedback and insights