OpenCPN Release Candidate 3.3.2503 Released

[bdbcat]

Thomas...

Good morning.

One last fix from Pavel on Options->UI->fonts for GTK. In github now. Please test on all your linux platforms, if possible.

Thanks
Dave

[seandepagnier]

I noticed that enabling chart outlines on cm93 mode is very slow. Can anyone else notice this?

I commented out the two lines that delete the display list and set it to zero, and the performance was once again very good. Is there some reason for why this can't be?

I also noticed using the measure tool if it crosses the dateline, or too many meridians messes up (try measuring a route around the world with it)

[Gilletarom]

Originally Posted by cagney

Gilletarom
...
Unfortunately I don't have the rights to edit at that level.

Thomas

ça alors ! Je n'en crois pas mes yeux. astonishing !

[CarCode]

Originally Posted by bdbcat

Mac users...

I am seeing a persistent crash on Mac Lion with 3.3.2503.

The situation is with the Dashboard PlugIn, Preferences dialog. When the layout of the Dashboard is changed, (add or remove an instrument) the system will crash on leaving the preferences dialog by the "OK" button. But strangely, it only crashes if the Dashboard is floating, not docked. If docked, all is well, no crash.

I stepped back to Betas from earlier last year, say 3.3.16xx, and still have the trouble there. So this issue seems to have been around a long time.

Anyone else seeing this?

Thanks
Dave

I cannot replicate this, neither with 3.3.2503 nor any older version nor with my own compiled version.
MacBook Pro 17", Yosemite 10.10.1
Please file your crash log from ~/Library/Logs/DiagnosticReports/ to have a look.
Gerhard

[CarCode]

It is actually reported that more professionals switch from GCC to LLVM/Clang:
LLVM Project Blog: Using clang for Chrome production builds on Linux
Is this a way to facilitate the OpenCPN project?
Gerhard

[bdbcat]

Sean...

re:

I noticed that enabling chart outlines on cm93 mode is very slow. Can anyone else notice this?

   I commented out the two lines that delete the display list and set it  to zero, and the performance was once again very good.  Is there some  reason for why this can't be?

The issue is panning at small scale. We need to make sure we pick up the new cell MCOVR structures. Dropping the list is is a crude solution, but safe.

See FS#1601 : Charts CM93 missing chart outlines - a "Vestas" check.

This issue needs more thought: how to be sure and show just the right larger scale outlines, without excessive clutter and performance hit.

Dave

[cagney]

Originally Posted by bdbcat

Thomas...

Good morning.

One last fix from Pavel on Options->UI->fonts for GTK. In github now. Please test on all your linux platforms, if possible.

Thanks
Dave

Dave

My early appearance was a bit of a red herring......

I have made a short test of the font selection dialog on Kubuntu 14.04 64 bit and Xubuntu 14.04 32 bit and found no problems.

Thomas

[transmitterdan]

I think I have been able to bumfuzzle build 2503 on Windows 7. I have a lot of ENC and BSB as well as old CM93 files. I have tried running this:

opencpn -unit_test_1

It took many hours to go through the charts which didn't seem right. At first I thought it was the SENC generation but I ran it again with all the SENC files already made so it didn't make them again. But it was still running after about 12 hours and I could not close O using the close window button. So I "killed" it with task manager and looked at the log file. There are things like this in the log file:

9:00:45 AM: Initializing Chart C:\Users\Dan\Documents\ENC_ROOT\US4TX52M\US4TX52M. 000
9:00:48 AM: RenderHPGL: width 1235 height 313
9:00:49 AM: RenderHPGL: width 405 height 452
9:00:49 AM: RenderHPGL: width 405 height 452
9:00:51 AM: RenderHPGL: width 404 height 452
9:00:51 AM: RenderHPGL: width 404 height 452
9:00:51 AM: RenderHPGL: width 404 height 452
9:00:51 AM: RenderHPGL: width 115 height 154
9:01:12 AM: RenderHPGL: width 405 height 452
9:01:12 AM: RenderHPGL: width 405 height 452
9:01:14 AM: Initializing Chart C:\Users\Dan\Documents\ENC_ROOT\US4TX58M\US4TX58M. 000

Notice that there is a gap of about 20 seconds in the log file where O just seems to hang for a while doing nothing.

Then there's this:

9:20:44 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:20:49 AM: RenderHPGL: width 1235 height 313
9:20:49 AM: RenderHPGL: width 1235 height 313
9:20:54 AM: RenderHPGL: width 115 height 154
9:20:54 AM: RenderHPGL: width 1235 height 313
9:21:07 AM: RenderHPGL: width 1235 height 313
9:21:25 AM: RenderHPGL: width 115 height 154
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.
9:21:48 AM: DashboardInstrument::OnPaint() fatal: wxBufferedPaintDC.IsOk() false.

It seems that something has gone off the rails with large gaps of time "lost" and that's why it takes all day for -unit_test_1 to finish. Does anyone think this is a real problem or just something local to me? I have never been able to leave build 2503 (or 2430) running all day without it going into some "weirdness" either locking up with almost 2G of ram allocated (normally it is 800MB) or long periods of stalls.

[nkiesel]

Originally Posted by bdbcat

Thomas...

Good morning.

One last fix from Pavel on Options->UI->fonts for GTK. In github now. Please test on all your linux platforms, if possible.

Thanks
Dave

Looks good on Ubuntu 14.10 amd64 opengl self-compiled against wx3.0. I _think_ the fonts are a big bigger (or bolder) than before, which is nice.

However, I noticed one thing (which I think is not really new): the object-query for a light does not show the light properties if you have a mark / waypoint at the same place. See attached screenshots (close to the "San Mateo area" from my previous bug reports; though I think this does not matter). I simply right-clicked on the buoy and selected "drop mark" and object-query lost it's light description.

I know that the light description only shows up if light display is enabled, which I also think is not nice (reported a long time ago and back then it was deemed "too much hassle" to fix).

Same also happens if there is a layer waypoint at the mark (of which I have many because my club uses lots of these marks as racing marks).

[nkiesel]

Originally Posted by bdbcat

Thomas...

Good morning.

One last fix from Pavel on Options->UI->fonts for GTK. In github now. Please test on all your linux platforms, if possible.

Thanks
Dave

Another problem with pop-ups: when there is a layer waypoint at a buoy and you double-click on the buoy, it shows the layer waypoint properties (instead of the buoy properties) even if the layer is disabled. Right-click and "object query" correctly shows the buoy properties.

[nkiesel]

Just created a pull request that fixes this

[HotRod]

Hey... Has OCPN 4.0.0 been released today? On my nix pc it reaads 4.0.0 on the top of the screen...

[bcn]

Today 12 UTC.....

See here:
http://www.cruisersforum.com/forums/...st-139227.html

[wdkester]

[QUOTE=wdkester;1715926]Can someone else confirm this oddity?

I notice that all the charts are NAD83. It occurs at many zoom levels. I've run CCleaner.

The following was subimitted against the previous release.

"I wouldn't call this a bug, but the quilting S/W sort of hiccups while panning down the NY state barge canal.

I'm at chart 14786_59, zoomed to scale= 15800, the next chart (_60) appears in the chart bar but isn't quilted. (EC1)

If you left click the chart bar (2nd chart from the left) to see its contents, chart _60 will be displayed quilted with _59 properly.

If you don't click the chart bar, and drag it a little further west, it will display a strange quilt. (EC2)

Drag it a little bit further and it will _59 disappears and _60 is dispalyed.


This behavior is caused by having "Full Screen Quilting" turned off.