New Grib Presentation

[rgleason]

Ptizef I recieved this by email, but it does not appear here. Can you add the screenshot?

Here is the message that has just been posted: *************** Need help please I have tried to pull the "official " master into my repository . I made first fetch upstream and then merge upstream/master I have now a new branch named DETACHED HEAD with unsynced commits (which doesn't appear in my remote git) and I can do nothing (see attached a screen shot) How can I escape from this trap ? many thanks Jean Pierre ***************

Ptizef,

I am very sorry you are having trouble, and I hope I am not the cause!!

Detached HEADS are not a problem, I've had those and there are several ways to fix that, but first lets try to work out what GIT commands you did in sequence. Then maybe Pavel can help. There are ways of "undoing" too.
Did you issue these commands?

git fetch --all ?
git merge upstream/master ?

then using
git branch -a
"shows a new branch named DETACHED HEAD with unsynced commits
(which does not appear in my remote git)" <-- please explain further.
"I can do nothing" <-- it it frozen? No screen shot available.

I might suggest not doing anything more until someone who knows GIT better can provide some guidance, like Thomas or Pavel, or Sean.

[seandepagnier]

Originally Posted by rgleason

There are some small anomalies in reading grib data from different formats, that will need to be fixed in the grib_pi.dll, but I believe those will be remedied eventually when there is some time. I have done a little study on this, but did not want to inject it into this process.

Could you post details so these can get fixed soon.

Originally Posted by Ptizef

I am very sorry to not understand all what have been written above due to my bad English and my weak skill but I think the change I've done in NEW_GRIB are completely incompatible with master branch coming from NEW_GRIB and grib_enhancements merge
So when I work on master , what exactly I have to do to apply my change ?(a lot of things doesn't work and have to be fixed)
Jean Pierre

I realize you have more improvements, great. Could you please merge them with the master branch? I did merge all of the new grib stuff with my changes before, but it is really difficult to have to repeat this process considering we are moving forward and away from the old grib stuff. If anything is unclear just ask.

If you could give specifics of what doesn't work that may help as well.

As for grib version 2; not implemented.

[rgleason]

I had an idea about how to move the improvements Jean made over, after he gets GIT sorted out. Perhaps someone does a diff on the two codes? so that the differences show, then you might be able to pick up those code bits and move them over? Then get Jean Pierre set up with a fresh copy and he can continue his work on that? -Just thought.

I think I can provide a list of the features he has improved.

Sean I will attach a document for the anomalies next, they were in an email you may not have received.

[rgleason]

SEan here is the document for the anomalies. If you want me to do some more testing on this, please let me know. You have moved us forward tremendously (same with Jean Pierre) and I would like to thank you once again.

[rgleason]

I just wanted to say regarding Climatology, in addition to COGOW see Thomas Post #128 above, there are other scat data available see Post #132 now, even if the new Quickscat is not put up on the space station until Jan 2014, see links in post #132 to #134, also ERDDAP for data seems to be very flexible and complete with various formats, xml, graphic, etc.

[rgleason]

Remote Sensing Systems Windsat website is a good example of the kind of Climate information that could be useful. This is a monthly view in July, 2012, weekly can also be shown week of July 14, 2012 for example. Useful for Route Planning purposes. If wind direction is also provided (which I think it is) it would be very useful. Perhaps 5-10 years of weekly data would not take up too much space?

I think Sean has already identified a "Climatology" plugin that would allow us to view and perhaps use it for routing, that he is interested in.

RSS / WindSat Data Images / Monthly

Here is Hurricane Sandy Oct 29, 2012
3-day average just before landfall NYC

[Ptizef]

Hi all

My Git problems are solved . Sorry for disturbing you

I fetch OpenCPN/OpenCPN/master and worked on it . I had some minors bugs : a crash with single date file , a problem with drop down list display an some boring warning when compiling .
these are fixed at https://github.com/Ptizef/OpenCPN/commits/master
what I have to do to push it in master ?

There is a fundamental problem I would like to speak about :
The video has very small steps : That means for example that in a file where the original steps are each 6 hours , data are shown each 1 hour ( or even 30mn !)
This is extremely dangerous ! And nothing is said to the user.
The default setting should be video step = file step
The smaller steps option could be authorized but with a warning when setting it and this option should not be persistent
What do you think about that ?
Jean Pierre

[Gilletarom]

hello,

I updated the french translation of the grib language file on I18n :

http://www.cruisersforum.com/forums/...ml#post1217459

Gilletarom

[rgleason]

Ptizef,

Congratulations on the GIT success.Thanks for your fixes. I think you have to ask Pavel or perhaps Sean for a Pull Request, and I don't know precisely how that is done, so you need to get that direct from them, not me, but I would be very interested in what the commands are. Can you explain to us a little more how you solved the git problem and what commands you used, for learning purposes? Thanks.

Yes there was a single date frame problem, I noticed that when you load a file, the previous one stayed displayed, then when you select forward or reselect the date the actual file loaded would show properly. Earlier, sometimes single frame date/time failed and I thought Sean fixed it.

I think your observation about small tweening steps setting is a consideration.
CPU overwhelming may occcur? Is that the problem? Or will it just play slowly each step and user can stop it playing and go change the setting? I am trying to figure out the danger.

I think it might be best if the initial setting is video step = file step as Jean suggests, but after I change it to interpolate, say 1 hour, I would prefer that it stay at that setting, unless I change it.
It there is a popup box or note it should be appropriate to the problem. Jean Pierre, perhaps you could explain a little further what the danger is because it has not been a problem for me. Using 1yr old laptop with i7-2720 2.20ghz chip.

[rgleason]

Ptizef,

Once you get comfortable using GIT remotes, doing pull requests and conceptually with grib_pi code, do you think you could implement some of the refinements I have seen in your earlier grib such as:

1. Pressure Isobar Text and boxes: Slightly smaller text and rounded corners on boxes. Reads better. Maybe with a size adjustment for eyesight.
2. Expand on your convenient email requests to saildocs.com by adding the "Modify" button and easy checkboxes for grib file data types and the resolution etc.
3. Perhaps if there is no objection, move the slider bar up to the first line and make it a little shorter as you have done earlier, to save real estate?
4. I am sure there are other things you have done that I did not notice, but it would also be very helpful to add some more grib sources if possible. We could help figure out which ones if you would like...?
5. For windows users, because the overlay for seat emp and currents, particularly are so dark and mask the chart and other data, it would be helpful if we could have a transparency slider to adjust that overlay setting parameter. This transparency slider would probably have to be in the setting area for each grib data type and a new parameter exposed for user control. I believe the setting is saved in the ini file but it may be an internal array, but Sean would be able to tell you. It should mimic or follow the way he did it to keep code clean.

Thanks, so much for all your efforts and team play!

Best Regards,
Rick

[nohal]

J-P...
Here comes what makes github a better place https://help.github.com/articles/using-pull-requests or it's stripped to the bone version: https://help.github.com/articles/cre...a-pull-request

Pavel

[seandepagnier]

Originally Posted by Ptizef

Hi all

My Git problems are solved . Sorry for disturbing you

I fetch OpenCPN/OpenCPN/master and worked on it . I had some minors bugs : a crash with single date file , a problem with drop down list display an some boring warning when compiling .
these are fixed at https://github.com/Ptizef/OpenCPN/commits/master
what I have to do to push it in master ?

I have already done a few small changes since the merge to master as well.

I think commenting out the line that sets the combo text is wrong, because it will not correctly display time values which do not fall on a grib record.

The bit setting the selection is good.. unfortunately wxwidgets works differently on different systems so I'm guessing that's why it was needed.

There is a fundamental problem I would like to speak about :
The video has very small steps : That means for example that in a file where the original steps are each 6 hours , data are shown each 1 hour ( or even 30mn !)

It is possible to configure the step size to different values, but I think the problem is the minimum is 1 hour. It would be nice to be able to have step sizes wider to 2 or 3 hours or more. Maybe it should be a text value and you could enter a decimal, or hour:minute or some format.. Any suggestions for entering timespans?

Maybe also a nearby button to set the step size to the steps in the file.. but I don't think the file is guaranteed to have all steps at an equal spacing, so I don't know how to determine exactly what this is.

This is extremely dangerous ! And nothing is said to the user.
The default setting should be video step = file step
The smaller steps option could be authorized but with a warning when setting it and this option should not be persistent
What do you think about that ?
Jean Pierre

Why is it dangerous?
Why would there need to be a warning?

Maybe instead of a warning, a button which brings up a dialog and completely explains what the options do would be more useful.

There really isn't much difference between having a step size of 5 minutes or 3 hours anyway, except that you can get a smoother interpolation.

Maybe on a slow machine when adjusting the slider it cannot update the isobars fast enough. If you want to skip ahead 3 hours every time, you can still use an hour divider of 12 (every 5 minutes) or whatever value you want, and set slices per update to 36, and it will skip ahead 3 hours each frame in playback.

Maybe the units in these options could be changed to make it easier to work with.

[seandepagnier]

Another thought: maybe precipitation and cloud cover maps could optionally be transparent as part of the map. This would allow them to work with other overlays better, and also give a better display.

An option for setting the transparency of the overlay is a good plan, but I think it can only work with opengl.

[CarCode]

There are several issues regarding a proper po/mo file for translations.
First it is called "Fahrenheit" and not "Farenheight", a typo only.
Second and more annoying there are several strings in the code starting with _T("... or wxT("... and therefore not in the pot file. Should be changed to _("... where a translation is desirable.

[seandepagnier]

I never was much good with spelling.

I fixed in git. As for the other strings.. they are in the generated code and I forget to enable internationalization. that is something I just fixed 2 days ago in my repository, but needs to get pushed to master I'll send a pull request. In the meantime you can use:
https://github.com/seandepagnier/OpenCPN