OpenCPN Beta Version 3.3.1731 Released

[seandepagnier]

Originally Posted by bdbcat

Sean...

1. wxASSERT fixes the assert crash. Dunno why, exactly.

Good to know, but I also wonder why this wasn't changed before.

2. I have fixed the Windows linkage problems in gshhs, and will be in my next batch of commits. So no need for a pull request. Actually, the problem was a bad diff in the git commit. The entire method GshhsPolyReader::crossing1Init() went missing. I have no idea how linux actually linked the full binary....

The reason is, it was not being used (didn't handle crossing the date line) but the dateline-safe version was, so I deleted it to avoid confusion. I forgot to delete the prototypes. In linux, you can make prototypes, but at link time since the functions were not ever called, they got eliminated and there were no link problems.

3. I also would like to see specific GRIB files that cause trouble, so I can add them to my test pool.

I believe it is the ones with negative index values which I didn't have any to test (downloading now)

4. I am close to a Beta release, tomorrow probably. So you may wish to hold off on any tweaks for a couple of days, please.

Thanks
Dave

I really want to fix the negative gribs first, (should be today) and I made another small fix to the gshhs (which will only affect weather routing users, but would be nice) If you can't accept changes, then just don't merge the fork request.. it can wait

Originally Posted by Ptizef

Hi Dave, Sean

Ok I think I have understood
It work ( at least the first test ) with Saildocs's files
But it do not work with Zygrib's files
I faced that problem in the past and found they have a difference. I think they have a negative DJ or something like that don't remember exactly and no time this morning to search. I joint a file as example

I confirm : No display and crash when changing setting and interpolate is selected

Jean Pierre

Jean,

Thank you.. I was wondering why it appeared to handle negative index values. I guess for gribs, the data is allowed to be in a mirror image, why this is allowed I don't know, it seems to complicate source code, but I will make a fix today.

[seandepagnier]

I have made a small change and now zygrib files work as well as the moving gribs from before with interpolation.

I don't know if moving gribs with negative Dj values will work (I don't know where to get one)

I also know, that moving gribs, or gribs in general which have different scales in space resolution of (.5 degrees for 3 days, then 2.5 degrees after that) will not work, but I don't know where to get one.

I also never had a crash happen, maybe only on windows? It should not ever crash, because if it did, then it will crash still in the cases of gribs it can't handle like the ones mentioned above.

[Ptizef]

Originally Posted by boat_alexandra

I don't know if moving gribs with negative Dj values will work (I don't know where to get one)

As far as I know, they don't exist . Moving gribs are a function of Saildocs, no other gribs providers (for the moment ! )

Originally Posted by boat_alexandra

I also know, that moving gribs, or gribs in general which have different scales in space resolution of (.5 degrees for 3 days, then 2.5 degrees after that) will not work, but I don't know where to get one.

You can get one easily by request function and selecting 0.5 resolution and more than 8 days horizon .
first days will have 0.5x0.5 and beyond these first 8 days, 2.5x2.5 resolution
You will see that between the last record with 0.5 and the first with 2.5 interpolation don't work (display nothing)
If you can't I'll post one
[/QUOTE]

Originally Posted by boat_alexandra

I also never had a crash happen, maybe only on windows? It should not ever crash, because if it did, then it will crash still in the cases of gribs it can't handle like the ones mentioned above.

Seem to happen with zygrib files and with "altitude" data, but I am testing the new git and will come back with my tests result

Jean Pierre

[Ptizef]

Test last git (3.31824)
- compiled fine OK
-Grib-pi ( do not work as it is for ZyGrib)
After applying manually Sean's pull request (not yet merged) "fix interpolation for gribs with negative D"
- OK for Saildocs files with or without "altitude" data
- OK for Zygrib files without "altitude" data
- Crash with Zygrib with "altitude" data

Jean Pierre

[seandepagnier]

Originally Posted by Ptizef

As far as I know, they don't exist . Moving gribs are a function of Saildocs, no other gribs providers (for the moment ! )

Ok. I think it may work if such a file did exist but wanted to test.

You can get one easily by request function and selecting 0.5 resolution and more than 8 days horizon .
first days will have 0.5x0.5 and beyond these first 8 days, 2.5x2.5 resolution
You will see that between the last record with 0.5 and the first with 2.5 interpolation don't work (display nothing)
If you can't I'll post one

I will try, but I don't understand how to use the email request. I can copy the contents of the text to send an email, but I don't know which address to use. I will try to take a look at the source to find out. It would be nice to display the address so users without a supported email client can copy and paste.

Seem to happen with zygrib files and with "altitude" data, but I am testing the new git and will come back with my tests result

Jean Pierre

The altitude is a new feature? I downloaded one with zygrib to try. Nice work! I get a combo box to select which level, then I see grib data at 500 hPa, with Altitude, Temperature and Rel. Humidity. All the boxes including wind turned yellow with a yellow warning on the chart canvas also displaying "WIND data at 500 hPA !". It appears to be working fine. Maybe you can post the grib file and the backtrace of where it crashed.

Originally Posted by Ptizef

Test last git (3.31824)
- compiled fine OK
-Grib-pi ( do not work as it is for ZyGrib)
After applying manually Sean's pull request (not yet merged) "fix interpolation for gribs with negative D"
- OK for Saildocs files with or without "altitude" data
- OK for Zygrib files without "altitude" data
- Crash with Zygrib with "altitude" data

Jean Pierre

[rgleason]

GFS:42N,38N,070W,76W|0.5,0.5|6,9,12,15,18,24,30,36 ,42,48,54,60,66,72,84,96,108,120,132,144,156,168,1 80,192,204,216,228,240,252,264,276,288,300,312,324 ,336,348,360,372,384|WIND|6.0,079

=======
GFS:42N,38N,070W,076W|0.5,0.5|6,9,12,15,18..84,96. .384|WIND|6.0,079
(same as above, just a shorter line)
To: query-grib@saildocs.com
Subject: gribauto
Body:
Area: 42N to 38N and 070W to 076W
Resolution: Starts at 0.5 x 0.5 and changes to 2 x 2 at 192 hours or 8 days
Intervals:
6 - 18 hours 3 hr intervals 0.5 x 0.5
18 - 84 hours 6 hr intervals 0.5 x 0.5
84 - 192 hours 12 hr intervals 0.5 x 0.5
192 - 384 hours 12 hr intervals 2.0 x 2.0
From
http://www.cruisersforum.com/forums/...ml#post1453674

It would be nice to shorten the request and reduce size by adjusting the intervals and frequency, but that would be the next step.

GFS20140129205703845.grb
File size is 14kb. 16 days.
Needs to be somewhat smaller I think for over SSB which can be achieved several ways.

======
GFS:42N,38N,070W,076W|0.5,0.5|6,9..18,24,30..84,96 ,108..192,216,240..384|WIND|6.0,079
(Changed hours 192-384 to 24 hr interval)

6 - 18 hours 3 hr intervals 0.5 x 0.5
18 - 84 hours 6 hr intervals 0.5 x 0.5 (84 hrs = 7 days)
84 - 192 hours 12 hr intervals 0.5 x 0.5 (192 hrs = 8 days)
192 - 384 hours 24 hr intervals 2.0 x 2.0 (384 hrs = 16 days)

GFS20140129210203856.grb
File is 13kb, 16 days

======
GFS:42N,38N,070W,076W|0.5,0.5|6,9..18,24..84,96,10 8..192|WIND|6.0,079
(Dropped 192 - 384 hours)

6 - 18 hours 3 hr intervals 0.5 x 0.5
18 - 84 hours 6 hr intervals 0.5 x 0.5 (84 hrs = 7 days)
84 - 192 hours 12 hr intervals 0.5 x 0.5 (192 hrs = 8 days)

GFS20140129210203856.grb
File is 12kb, 8 days

[rgleason]

Some more work on walking gribs and explanation here:
http://www.cruisersforum.com/forums/...ml#post1523689

Explanation about the intervals must be integers and not fractions - "Note that if you have a request without whole integers the change in increment will be ignored and your grib will likely be bigger with more frames."
http://www.cruisersforum.com/forums/...ml#post1527669

This is about concatenating gribs, various tests that work.
http://www.cruisersforum.com/forums/...ml#post1525612

[rgleason]

This may be a little late, but I hope it helps. I don't know what a negative what-ever is.
Thanks Sean and Jean Pierre for working this out!

Here is an example request to query@saildocs.com

GFS:42N,38N,070W,076W|0.5,0.5|6,9..18,24..72,84..1 68|WIND|4.5,079

I've also added some other examples of moving gribs below the first one.


---Later. Sorry now I understand you are working on Zygrib! Sorry about that.

[seandepagnier]

I got the email query to work. I had to copy the text in the box, then paste it into the email client, I assumed this would be automatic. The email address and subject were automatic.

Anyway, I got an 8 day grib and it has 0.5 degree resolution all the time. So I tried a moving grib and same result, even on day 8 I have .5 degree resolution.

I noticed a glitch, when zoomed out so that only every other arrow is shown, when switching to the next moving grib, the interpolation is not smooth because it switches from odd to even... we could "fix" this with interpolation when zoomed out, but not sure this is better or even a problem.

[Ptizef]

Originally Posted by boat_alexandra

I will try, but I don't understand how to use the email request. I can copy the contents of the text to send an email, but I don't know which address to use. I will try to take a look at the source to find out. It would be nice to display the address so users without a supported email client can copy and paste.

I don't understand why you can't use eamil request. It has been tested on windows (XP,Vista,W8.1) and Linux at least Debian
Anyway, the request address is in the config file, no need to explore sources
[/QUOTE]
The altitude is a new feature? I downloaded one with zygrib to try. Nice work! I get a combo box to select which level, then I see grib data at 500 hPa, with Altitude, Temperature and Rel. Humidity. All the boxes including wind turned yellow with a yellow warning on the chart canvas also displaying "WIND data at 500 hPA !". It appears to be working fine. Maybe you can post the grib file and the backtrace of where it crashed.[/QUOTE]

Yes it's a new feature since a while now.
I will try in debug mode and post an example. Coming back soon ...
Jean Pierre

[rgleason]

The first grib file I uploaded above is not a moving grib, so I made another request direct from the plugin request to query@saildocs.com
send GFS:45N,38N,79W,63W|0.5,0.5|0,3,6..192|=
WIND,PRESS,APCP,TCDC,AIRTMP,SEATMP,CAPE,WIND500,HG T500|5,80

Have compiled and cpack opencpn 3.3.1824 then used exe to install.
Then used last nights compile and cpack for weather_routing, watchdog, etc to install into the new beta version.

Observation during Install with exe: When doing the advanced setup, you need to delete the Opencpn in the path after selecting the advanced setup path, in this case ...Program Files (x86)/Opencpn3.3.1824

With Smooth Zoom and NO Opengl after Forced Full Rebuild of charts
BSB charts pan and zoom very nicely
CM93 seems to work fine too.
Vector ENC Noaa charts set on "Standard" also work nicely, with a slight hesitation
and sometimes a senc is updated (normal).

With smooth zoom off and Opengl enabled with just Texture Map checked
Have Forced a full rebuild of Senc
The charts seem stretched out and out of scale for some reason.
Pan and Zoom seem a bit jumpy.

Before this I tried Weather routing with a current walking grib and it failed with WxRte not opening and Opencpn freeze.

[Ptizef]

Originally Posted by boat_alexandra

Anyway, I got an 8 day grib and it has 0.5 degree resolution all the time. So I tried a moving grib and same result, even on day 8 I have .5 degree resolution.

If you choose "Saildocs" and you select 16 days, you will get 16 days, 8 days with 0.5 and 8 days with 2.5 each 12 hours
anyway here attached a file who have two resolutions as described

Jean Pierre

[Ptizef]

Grib_pi crash with zygrib files and "altitude" data

You can find attached a file who create these crashes ( in fact any file of this type uploaded from Zygrib's tool or got by mail from zygrib create a crash)
just play with it with playback or by dragging the timeline, or with alternatively "next" and "prev" buttons

Also attached two screen shot of the crash in debug mod

I have also sent a crash report

Jean Pierre

[rgleason]

Jean Pierre,
How would I determine a saildoc's grib resolution in a given timeframe?
1. By inspection of the file?
2. By looking at some aspect of the the grib, with interpolate unchecked?

Also are you saying that Saildocs does not recognize our requests for resolution changes? (I had outlined how to do this above based on another sailor's experience.) So it just does what it wants with regard to resolution?

Quote:
Originally Posted by boat_alexandra
Anyway, I got an 8 day grib and it has 0.5 degree resolution all the time. So I tried a moving grib and same result, even on day 8 I have .5 degree resolution.

If you choose "Saildocs" and you select 16 days, you will get 16 days, 8 days with 0.5 and 8 days with 2.5 each 12 hours
anyway here attached a file who have two resolutions as described

[seandepagnier]

I finished fixing the crash as well as fully supporting interpolation for moving grib even with different timescales. Unfortunately I missed the beta.

Pitzef,

The first file you posted works, but it seems every odd grib record set has all the records while the even ones only have a some records. This looks funny even without interpolation. Maybe it's ok, or we need some more logic to skip past this entree if it is lacking the needed record so interpolation works for all datatypes at all times, not sure.