OpenCPN Version 1.3.5 Beta Technical

[Ptizef]

HI Dave
1)With 1.3.4 , building the charts directory with BSB and CM93 took about 10 secondes . Now , with 135 1210 , the same operation with exactly the same set of maps takes at least 2.5 - 3 minutes ! And during this time , if I click on the pup up window , I get a crash
(vista home)

2) My CM93ed2_2009 stops at 80° north . If I click outside , I get a crash

many thanks
Jean Pierre

[cagney]

FWIW
Testing Ptizef's points on Kubuntu 9.10.
1 Yes it is slower rebuilding the chart database, but while this is going on, no pop-up i visible.
2 No crash, works as expected.

Thomas

[ChuckSK]

BSB chart 530 marks the GPS data as invalid. That is the only chart found so far with this problem.

Chuck

[Ptizef]

Hi

you can see below screen shots related to the maps directory built :
the first one after about ten seconds
the second after a couple of minutes
the third is a crash .
each time , for the BSB directory works fine and fast but it slow down (or crash ) on CM93 directory building (in 134 it was practically unvisible)

Jean Pierre

[ChuckSK]

Re #198 above:

Changing chartimg.cpp lines 3755 and 3756 to initialize lonmin and lonmax to 180.0 and -180.0 instead of 179 and -179 fixed the problem.

Chuck

[TerjeRogne]

Thanks a lot Dave for a beautifully simple program!
Keep up the good work - avoid too many nice to have features.
OpenCPN is now my first choice - in addition to a bunch of good old paper charts!

The 1.3.4 release has some annoying limitations with the depht symbols in shallow waters - which are close to cured in the 1.3.5 1210 release - however the behaviour is not stable - sometimes the old red symbols in the CM93 charts return, instead of the 6 edged symbol showing awash rocks. Turning off OpenCPN and starting anew gives a fresh screen - see the two pictures below.

I don't know what is the failure - but as you see the area has extremely many soundings, which is often the case in Norwegian waters - maybe some kind of overflow?

I'm running the program on a powerful Win7 machine with 8GB RAM, but I experience the same problem with a weaker XP machine.

Best regards!

[ChuckSK]

The wvschart seems to have disappeared.

[bdbcat]

ChuckSK....

Good catch, there, Chuck...

Odd chart, that one. Anyway, it surfaced a latent bug in the georef calculations around 180E/W, again.....

Fixed it in a slightly different way, and will be in the next Beta.

Thanks again, and keep up the sleuthing!

Dave

[bdbcat]

TerjeRogne....

I have been unable to reproduce the unstable condition you have indicated with respect to awash rocks. They look fine to me here.

Technically, opencpn is evaluating this rock as a underwater obstruction whose least depth(awash) is less than the safety depth (looks like 10m in your case), and is located within a depth area which is otherwise everywhere deeper than the safety depth. It is clearly the last test which is the problem. Sometime OK, sometimes not.

Is there any reproducible technique which you have found which brings on the effect? Any hints on keystrokes, menu actions, etc?


Thanks
Dave

[TerjeRogne]

Thanks for the fast answer Dave.

With more systematic testing I find that it seems like the problem is related to certain areas in the CM93 maps.
In the screenshot to the left there is no problem. The only action performed here is to open 1.3.5 1210 (safety depth is 8m). Then the chart was pulled about 10% towards left and OpenCPN was closed.

The screenshot to the right is the screen as OpenCPN was opened again. It opens like this always with the map in this position.
Pulling the map 10% to the right before closing gives a correct map again when opening.

This behaviour is reproducable - the area is critical. There are also several other areas where I get the same behaviour. These areas have a lot of shallow soundings close to each others.

Do you have access to the same maps? If so you might test the same areas - the position should be possible to read in the two pictures.

Thank you

Terje

[Ptizef]

Hi Dave
If I open the tool box , click on the charts tag without changing any thing and click "OK" or go to any other tag and changing or not the other tag and click "OK" launch a new chart directory re-building (with all the problems described on #196 and #199 )

regards
Jean Pierre

[crienoloog]

Sorry to intrude here...
The beta version is an .exe and thus for windows only. Is there a Ubuntu/linux version as well?
Thanks, C.

[idpnd]

Originally Posted by crienoloog

Sorry to intrude here...
The beta version is an .exe and thus for windows only. Is there a Ubuntu/linux version as well?
Thanks, C.

You will be rather pleased to hear that you can easily compile a bleeding-edge version yourself!

Compile from CVS isntructions here.. Easily done.

There is an unresolved bug with the wxwidgets in the repos however, so you will have to make a small modification on the newer ubuntus (9.04 and 9.10)

[crienoloog]

Thanks, but that is way to complicated for this noob...
Regards, Eric

[bdbcat]

Hi folks...

Here is Version 1.3.5 Beta 1216.

Changelog:

63. Correct GRIB SIGHW tracking logic causing fault.
64. Implement transparency for some GRIB overlays.
65. Correct Database update logic for currently open or cached charts.
66. Improve HiDef GRIB Isobar rendering with splines.
67. Improve Mercator distance/direction algorithms.
68 Improve performance of chart database update cycle.
69. Correct UWRTOC symbolization rules
70. Correct cm93 crash above Latitude 80N/S
----Release Beta 1216----

Notes. Changes are minor and localized, with minimal regression risk. Famous last words....

We are ready to try some internationalization of embedded text strings.
Please see the new i18n thread on the forum for information.

We are coming down to the wire on test time for 1.3.6 Release candidate.
I plan to be substantially offline for a 3 month cruise starting in early Jan, 2010.

opencpn is, after all, all about cruising on real boats.....

So get your last minute "gotta be fixed" bugs/features in now, and I'll re-prioritize for the sprint.

Thanks again for all your test effort and comments.

Dave