OpenCPN Version 2.1.0 Beta

[cagney]

Testing 518.
Viking Sailor

Zooming out BSB charts hangs 518.

Cannot reproduce!

With CM93 and BSB quilting enabled zoom out chrashs and exits cleanly.

Cannot reproduce!

manimaul

zooming in and out of s57 causes segfault

Cannot reproduce!

ptizef #506 scenario 2, confirm crash. First pic below.


The DW arrow looks fine.
More on arrows. The second picture below looks just fine, but zooming in leads to the last pic, with just to many arrows marking the south going traffic flow. What about all those small pink/magenta circles? They disappears if zooming out, but appears when zooming in, in seemingly changing patterns??

Thomas

[cagney]

Dave

As you know, I sometimes use Phiggins aisgps simulator for testing releases. This simulator is available here.
With build 514 and 518 the simulator no longer works correctly on OpenCPN. "Own ship" is no longer receiving gps signals, and don't appear on the chart. The ais for other vessels works as before.

Thomas

[bdbcat]

Thomas....

1. Funny circular artifacts are a bug. See #501

2. Extra TSS arrows are in the dataset. Who has ever scrubbed cm93 for this kind of thing?

3. Have you got latest CVS? Late last night change in sentence.cpp, affecting NMEA RMC sentence.... Manimaul helped to track it down.

Thanks
Dave

[cagney]

Dave

I downloaded and compiled ~1500Z today.
Thomas

[Ptizef]

To conclude on zoomto crash on my small notebook XP , it's focused on very small route/track , of a length near zero Nm . It happens too when trying zoometo on active track with only one track point (boat not moving).
Adding something like
if (rw<1)
rw=1;
in routemanagerdialog.cpp after line 383 to limit zooming solves the problem

Jean Pierre

[bdbcat]

Cagney:

Thomas, I just tried the phiggins simulator on Ubuntu with unicode build. Works fine.

Please check sentence.cpp, line 54

NMEA0183_BOOLEAN SENTENCE::Boolean( int field_number ) const
{
//   ASSERT_VALID( this );

   wxString field_data;

   field_data = Field( field_number );

   if ( field_data.StartsWith(_T("A")) )
   {
      return( NTrue );
   }
   else if ( field_data.StartsWith(_T("V")) )
   {
      return( NFalse );
   }
   else
   {
      return( Unknown0183 );
   }
}

if you set

DebugNMEA=1000

in the config file, you can get logfile monitoring of NMEA status.

Thanks
Dave

[Netsurfer]

When quilting is enabled the 'Enable CM93 Detail Slider' entry disappears from the context menu (even if only CM93 charts are installed/ available).
I would love to see the slider being integrated in the GUI anyway (see Feature Requests) .

Gunther

(Win 7, 32 bit, build 518, only CM93 charts installed)

[Ptizef]

Hi Dave

Just a detail ... but
Enabling track create a track with one track point , but if the boat is not moving and I disable track and then re-enable it , the new created track as no track point .
In this case , track will really start at the second track point and the first leg will be lost.
Furthermore, zoomto this track (with no track point) leads to big problem ...

regards
Jean Pierre

[cagney]

Originally Posted by bdbcat

Cagney:

Thomas, I just tried the phiggins simulator on Ubuntu with unicode build. Works fine.

Please check sentence.cpp, line 54

NMEA0183_BOOLEAN SENTENCE::Boolean( int field_number ) const
{
//   ASSERT_VALID( this );

   wxString field_data;

   field_data = Field( field_number );

   if ( field_data.StartsWith(_T("A")) )
   {
      return( NTrue );
   }
   else if ( field_data.StartsWith(_T("V")) )
   {
      return( NFalse );
   }
   else
   {
      return( Unknown0183 );
   }
}

if you set

DebugNMEA=1000

in the config file, you can get logfile monitoring of NMEA status.

Thanks
Dave

Dave

OK, simulator working again! Thank You!
Some unrecognized NMEA sentences in the log...

Thomas

[bdbcat]

Ptizef....

JP. I am confused about your latest post.

We should not be able to create a track with 0 or 1 points. See the code, chart1.cpp line 3287

void MyFrame::TrackOff(void)
{
      if(g_pActiveTrack)
      {
            g_pActiveTrack->Stop();

            if ( g_pActiveTrack->GetnPoints() < 2 )
                  g_pRouteMan->DeleteRoute ( g_pActiveTrack );
      }

      g_pActiveTrack = NULL;

      g_bTrackActive = false;

      if(m_toolBar)
            m_toolBar->ToggleTool(ID_TRACK, g_bTrackActive);
}

Please try your explanation again, so that I can reproduce the problem.

Also, for cm93 underzoom limit: Please, what is the actual pixel size of your computer screen. What we try to do is limit the underzoom to about 340 degrees of longitude covering the entire width of the screen.

You could try experimentally changing the factor at cm93.cpp line 4363:

      return 3.0 * (WGS84_semimajor_axis_meters / (canvas_width / canvas_scale_factor));

a value of 3 produces (3/pi) * 360 degrees, or about 340 degrees. Try something less, and see how it works....

Thanks
Dave

[cagney]

Dave

Course up rendering improved. OK when not using quilting.
This is the result of quilting ......

Thomas

[Ptizef]

Dave

Thanks for your answer

1)Track with no track point
Sorry , I wasn't so clear . May be the terme "create" is not correct .
When you click on the track icon to enable track , a track is "created" and if you open route manager , you can see it named automatically "Active Track" and if you open route properties , you can see that there is 1 track point . If the boat don't move and If you click one more time on the track icon to stop track , off course this track will disappear and that's the result of the code in your post .
But , then if you click once more on the track icon to re-enable track, a new track is "created" and if you open route properties , you can see there is no track point . If the boat move now , the first track point will be created according to the parameters (time or distance ) and the first leg will be lost .
This is due , I think to the code (line 1973)

void Track::AddPointNow()
1969 {
1970
1971 wxDateTime now = wxDateTime::Now();
1972
1973 if((m_prev_glat == gLat) && (m_prev_glon == gLon)) // avoid zero length segs
1974 return;
1975
1976 if(m_prev_time == now) // avoid zero time segs
1977 return;

I tried
if(((m_prev_glat == gLat) && (m_prev_glon == gLon)) && m_bRunning)

It works but may be there is another impact and I've no GPS at the moment to test more .
2° maximum under zoom CM93
I have the problem only on my 1366x768 screen and vista and in quilt mode .
On my small 1024x600 and XP , maximum under zoom is OK
Problem with the screen or vista ?
I tried several values of the factor from 0.5 to 100.0. In fact I saw a big impact in no quilt mode , bur very few impact in quilt mode .
Alway the left part of the screen seams to be "away" from the rest . If I zoom in for only one step , the problem disappear

best regards
Jean Pierre

[bdbcat]

Cagney: Course-up in quilting is broken, on my list. Thanks.

Ptizef:
1. zero point tracks: Thanks for the clear explanation. I get it now, and the solution is simple. We need to set Track::m_prev_glat, m_prev_glon to an unreachable lat/lon when we start a new track, so that the first point is always added.

2. Underzoom. Sorry, I spoke before I truly understood how my code actually works. Underzoom bounding is now OK for both modes. He should now bound to 180 degrees maximum longitude. This will avoid fully any wrap-around longitude problems. Is 180 enough for practical use?

Thanks, guys
Dave

[bdbcat]

Hello testers!

OpenCPN Version 2.1.0 Beat Build 520 is ready.

http://southbaynetwork.com/opencpn/o..._210_setup.exe

Changelog:
107. Correct logic for quilt zooming in a location with no charts available.
108. Correct RouteManagerDialog route-delete logic.
109. Correct Track and Grib toolbar icon selected state graphics.
110. Correct Compass Rose and GPS status icon locations on startup and resize.
111. Implement Route Leg course/distance display on right click of route segment.
112. Correct fault on DeleteAllTracks if track is active
113. Improve formatting of RouteProperties dialog.
114. Correct some DeleteAllRoutes logic with RouteProp and RouteManager open together
115. Correct logic on highlighting of marks in RouteProp dialog
116. Display tweak: If the only chart available on the chart bar is cm93, then do not highlite on rollover.
117. Implement chart depth unit display in quilt mode, if possible (i.e. all charts in quilt are same units)
118. Limit (again) the maximum underzoom of cm93 charts to accomodate large wide screens.
119. Correct fault found with missing config and chart database on initial startup.
120. Added indication of poor georef solution to chart info box.
121. Correct faults caused by excessive overzoom on some cm93 cells.
122. In Course-Up mode, implement one-shot timer for low definition rotation rendering on mouse pans.
123. Implement "OVERZOOM" indicator in quilt mode.
124. Correct cm93 symbolization of 2 way deep water routes.
125. Disallow quilting mode for all TMerc charts.
126. On sectored lights, allow cursor pick for object query only near light location, not on light sector.
127. Correct fault in closing RouteProperties with route externally deleted.
128. Correct Windows hang with multiple RouteProperties dialogs accessed.
129. Improve chart selection logic on quilt pan.
130. Correct Routemanagerdialog->ZoomTo for excessive overzoom.
131. Correct logic error concerning repeated zero length tracks.

---Release Beta 520---

Notes:

125. Sigh...We just can't make TMerc quilting work without re-projection to a common meridian, I think.

Known problems:

1. Course-up in quilting mode. Screen clips oddly....
2. Course-up performance in general is not great. Needs work...
3. Euro IENC overlayed ENC cells. Still researching this one. Problem: There is no (clear to me) indication
in the ENC file that this file is to be treated as an overlay.
Of course, we can think of ALL ENC cells as overlays, and bring them onto the screen by a transparent masked BLIT operation. This is very slow, and would affect every S57 ENC, IENC overlay or not.. Probably not a good idea. I resist the idea of an "IENC Mode Switch". Still thinking on this.....
4. New Icons. In process.....

CVS Updated

I'm off to sea for a few days. Headed for Block Island, I think.

Have fun, Good Luck.
Dave

[champ8242]

Dave,

Once more...impressive!!!
Thanx and have a good trip

Bart