OpenCPN Version 4.2 Released

[Ivanm]

Originally Posted by Ivanm

O 4.2 win7 the menu bar disappears and on startup freeze an icon spinning for about 3 minutes then back to normal

Sent from my SM-G920F using Cruisers Sailing Forum mobile app

I had 12000 waypoints , made it not visuable now O4.2 run without any problems when I make waypoints visable again it freeze and very slow

Sent from my SM-G920F using Cruisers Sailing Forum mobile app

[tiersp]

Originally Posted by bdbcat

tiersp...

I downloaded 7130.kap, and tested with 4.1.1301 and 4.2.0.

No difference, and no problems noted.

Perhaps you can post a screenshot of your situation, so we can compare and analyse.

Thanks
Dave

bdcat,
Thank you for your answer,
here are two screenshot
one with cm93 and the other with 71.30.kap
I have put three mark and you can see the differences.

[rgleason]

Bdbcat, Pavel, or Thomas,

Both Gilletarom and I would like to know how to upload a file to the opencpn.org website and make a download link to that file. There are several examples of this in the User Documentation. Gille would like to upload an xml file (not an image file) that defines a new Vector Style, and we can find no place to do this. It is not a feature of the User "File Browser" where upload are only for images, and I can find no other tool that enables us to upload a file so that we can make a download. It appears that there is a "download" directory however. (See the link please).

Do any of you know how these download links are created? They can be very useful and we should know how in order to maintain the User Manual.

Thank you.

See http://www.cruisersforum.com/forums/...ml#post2038072

[bdbcat]

tiersp...

Interesting. For me, at the same spot, I see no problem. See screenshots attached.

Anyone else tried this?

Dave

[houlejm]

Originally Posted by bdbcat

tiersp...

Interesting. For me, at the same spot, I see no problem. See screenshots attached.

Anyone else tried this?

Dave

Dave, tiersp

Like Dave I do not see any great difference but I would have problems sailing there with a GPS.

Just for a test I put a mark at the intersection of LAT/LON lines. It's coordinates should be even. I see quite a large error on the latitude.

See screenshot

Jean-Marie

[CarCode]

Jean-Marie,
it seems all to be quite different. The wonderful world of OpenCPN...
Gerhard

[houlejm]

Originally Posted by houlejm

Dave, tiersp

Like Dave I do not see any great difference but I would have problems sailing there with a GPS.

Just for a test I put a mark at the intersection of LAT/LON lines. It's coordinates should be even. I see quite a large error on the latitude.

See screenshot

Jean-Marie

I did more tests:

In my GPS I created a Waypoint with EU1950 datum, then changed the datum to WGS84.

I entered the converted Waypoint into O and bingo, it is at the right place.

The error is easily seen if you activate the grid in O.


JM

[HappySeagull]

...the drawn lines of latitude and longitude are commonly incorrect on old "Kaps". When old charts were digitised to a new datum (one hopes it was wgs 84! ),nobody bothered to blot out the lines drawn per the original datum.

[bcn]

Just for comparing:
- CM93 with marker on full integers (no decimals hanging around, using a circle for the marker as the anchor point is clear then)
- raster chart 7130 from VisitMyHarbor set - in this case the lat/lons are spot on with the lat/lon overlay from OpenCPN

[Gilletarom]

Dave, All,

Originally Posted by tiersp

bdcat,
Thank you for your answer,
here are two screenshot
one with cm93 and the other with 71.30.kap
I have put three mark and you can see the differences.

I feel that this is an old problem resurfaces. It had been settled long ago. May be in the 3.2 or 3.2.2 O version. And besides, it's already this 7130 chart that had served to show this defect ...

[HappySeagull]

Originally Posted by bcn

Just for comparing:
- CM93 with marker on full integers (no decimals hanging around, using a circle for the marker as the anchor point is clear then)
- raster chart 7130 from VisitMyHarbor set - in this case the lat/lons are spot on with the lat/lon overlay from OpenCPN

well, there you go! So! anybody got an up-to-date lightlist? Harping on it, but a layer with lightlist positions is pretty handy when various charts are being compared.

[.Paul.]

Originally Posted by bdbcat

tiersp...

Interesting. For me, at the same spot, I see no problem. See screenshots attached.

Anyone else tried this?

Dave

Dave and tiersp,

I have also looked at this chart. I set the three waypoints on the locations of the marks as shown on CM93, and the positions on the 7030.kap chart are within one or two chart pixels. The 7030.kap has "KNP/...GD=European 1950" and "DTM/-3.420000, -4.92" offsets in the KAP header.

I see that there is a space before the longitude offset in the KAP header. The screenshots tiersp posted show an error in the longitude and no latitude error. The longitude error seems to be the same for each mark. The space in the DTM/ line and the longitude error might be a coincidence.

I recently corrected a small skew in a scanned chart. OpenCPN 4.1.x in Windows correctly displayed the revised chart after the usual "Scan Charts and Update Database", but updating the chart database was not sufficient on my Android. I had to Clear the Texture Cache. After the clearing the cache, the GeoRef was correct. I don't use the Texture Compression with Caching on my Windows machine. So, I did not have anything in the cache to clear. Anyway, tiersp's computer might have texture cache images of 7030.kap (and other charts that have datum shifts) made by a version of OpenCPN earlier than the recent corrections to the GeoRef code, and "Clear Texture Cache" might be needed.

Paul

[bdbcat]

Folks...

Thanks for the focus on this particular chart. I think we understand it:

1. As Happy Seagull notes, old charts converted from arbitrary datum to another datum, say WGS84, will show their printed lat/lon grid incorrectly. The printed lat/lon grid would be correct in the original datum, but not the new one.

2. Since most everyone shows that this chart georefs consistently with cm93, we can concluded there is no fundamental regression here on O4.2

3. I suspect .Paul. is right, and there is a mismatch on the OpenGL cache texture, somehow. We await a report from tiersp with OpenGL disabled.

Thanks again
Dave

[.Paul.]

The dropdown Icon list in the Waypoint Properties dialog is too tall as shown in this Windows screenshot.



This makes it difficult to select icons near the bottom of the list with the mouse (or touchpad) pointer. A workaround is to use the tab key to highlight the Icon box. Then use the down-arrow or up-arrow key to select an icon.

I reported the problem in July, and it seemed to be fixed in the OpenCPN 4.1.925 beta. The problem is back in 4.2.0. Going back through some builds, I see that a build with commits through the evening of Jan 20 EST has an Icon list that fits as it did in 4.1.925. The icons in the list are smaller than the ones on the chart canvas.



The build with commits through the evening of Jan 21 EST has an Icon list that is too tall. The same as in 4.2.0.



Paul

[tiersp]

Originally Posted by bdbcat

Folks...

Thanks for the focus on this particular chart. I think we understand it:

1. As Happy Seagull notes, old charts converted from arbitrary datum to another datum, say WGS84, will show their printed lat/lon grid incorrectly. The printed lat/lon grid would be correct in the original datum, but not the new one.

2. Since most everyone shows that this chart georefs consistently with cm93, we can concluded there is no fundamental regression here on O4.2

3. I suspect .Paul. is right, and there is a mismatch on the OpenGL cache texture, somehow. We await a report from tiersp with OpenGL disabled.

Thanks again
Dave

Hello Dave,
I do not use OpenGL.
The problem is the same with all the raster charts in European 1950.
The problem occured with O4.2 only and not with all the previous versions.
Thanks again
Patrick