Weather Routing

[rgleason]

Kevinh Your enthusiasm is good. If you don't have much programming experience this will be a big learning experience! But go ahead and try.
https://opencpn.org/wiki/dokuwiki/do...piling_mac_osx

[rgleason]

For those using Weather_routing there is an excellent discussion available on one of the Plugins github Issues which focuses on Polar Files.

@Kevinh, it would be great if you could compile for MacOS!

[rgleason]

New Windows version with many improvements!
https://github.com/rgleason/weather_routing_pi/releases
Many thanks to Theeko74 and Sean.

I've done about 20 routings with no problem.
Please advise if there are any problems.

[wholybee]

Originally Posted by rgleason

KevinH,
I am sorry but I do not know how to compile for MacOS. MacOS is often compiled by Pavel [nohal] and Peter Tulp [ptulp]. Pavel is quite busy most of the time and Peter not really available right now. Sorry, perhaps a mac user will help. We would really like to have mac user testing these plugins too, if we can get the PI compiled without too much trouble. Right now it is changing fast and it might be better to wait a little while.

Sean
I downloaded the Win file Bump Version from here https://ci.appveyor.com/project/sean...0.51/artifacts

It runs nicely. with the Optimal Route Feathers! The example files and configurations were not changed yet, but they did run properly. See screenshot

I might be able to help with this. I have recently successfully compiled for the first time, though it was the windows version, in a VM running on my Mac. I do have previous Mac development experience though its dated I could probably get the environment running fairly easily.

My challenge has been that all of my programming experience is from a long time ago;Git is new to me, wxwidgets is new to me, vc++ is new to me. But I have used both c++ and objective c long ago so the code is at least readable to me.

A suggestion, the developer documentation should include the steps in Git to check out 4.8.2. The instructions to compile 4.8.2 are all but useless without that step, and even people with otherwise valuable programming expertise might not be that familiar with version control, or Git.

[nohal]

Originally Posted by wholybee

I might be able to help with this. I have recently successfully compiled for the first time, though it was the windows version, in a VM running on my Mac. I do have previous Mac development experience though its dated I could probably get the environment running fairly easily.

My challenge has been that all of my programming experience is from a long time ago;Git is new to me, wxwidgets is new to me, vc++ is new to me. But I have used both c++ and objective c long ago so the code is at least readable to me.

A suggestion, the developer documentation should include the steps in Git to check out 4.8.2. The instructions to compile 4.8.2 are all but useless without that step, and even people with otherwise valuable programming expertise might not be that familiar with version control, or Git.

I'm afraid providing the instructions to checkout a tag (= "Run git checkout v4.8.2") will not actually change anything but show you that you really should have looked at the first chapter of https://git-scm.com/book/en/v2, just 5 minutes later.
We unfortunately don't have manpower to teach people programming and duplicate every piece of documentation out there. We (=I) also do not have energy, time, resources, whatever, to maintain the outdated toolchain, dependencies and know-how. Said that, all the build instructions are updated for the current master and that is what developers should use. 4.8.2 is history.

Pavel

[seandepagnier]

Originally Posted by wholybee

I might be able to help with this. I have recently successfully compiled for the first time, though it was the windows version, in a VM running on my Mac. I do have previous Mac development experience though its dated I could probably get the environment running fairly easily.

My challenge has been that all of my programming experience is from a long time ago;Git is new to me, wxwidgets is new to me, vc++ is new to me. But I have used both c++ and objective c long ago so the code is at least readable to me.

A suggestion, the developer documentation should include the steps in Git to check out 4.8.2. The instructions to compile 4.8.2 are all but useless without that step, and even people with otherwise valuable programming expertise might not be that familiar with version control, or Git.

I would say, try to follow the instructions in the wiki. If they don't work or you get stuck, just post a question, then fix the wiki.

Several people have compiled my plugins on mac, and I don't think it's difficult.

For extra credit, figure out how to make travis autobuild for mac.

[rgleason]

Pavel, yes there is a newer Development Environment using wxWidgets 3.1.1, but as you said, we are still working on updates to the v4.8.2 plugins. I am trying to understand why are you suggesting using this newer version when the results do not seem to work on v4.8.2? (From my limited experience with Jon's compiled versions of Ocpn-Draw.)

[kevinvh]

Originally Posted by rgleason

Pavel, yes there is a newer Development Environment using wxWidgets 3.1.1, but as you said, we are still working on updates to the v4.8.2 plugins. I am trying to understand why are you suggesting using this newer version when the results do not seem to work on v4.8.2? (From my limited experience with Jon's compiled versions of Ocpn-Draw.)

Hi everyone,

I have spent a couple of hours reading the Developer Manual - Compiling on macOS, Homebrew website and Xcode on the Apple App Store. Xcode is a 5.5 GB download and that is not possible for me using a mobile phone (no WIFI). I think that compiling for Mac is a little beyond me unless I had an example to follow.

Think I will respectfully back-away and wait for the compiled version of WR PI for Mac to be released.

[Gilletarom]

Hello Sean d'Epagnier, hello Did-g,

I made a clean re-installation of O4.82 evacuating all the parallel installations present on my Lenovo under W10.
And I use the latest version of WR 1.11.1 that RGleason put online.

I chose to search a weather-route from Cherbourg-inside to Cowes.

Yes, I'm testing a difficult situation.
- On the one hand, I place the starting point in the foreshore of Cherbourg, so, WR must get the boat out of this area.
- On the other hand, I ask WR to bring the boat to Cowes, entering the Solent, either from the west or from the east.

With a 15-minutes timestamp, WR is unable to get out of Cherbourg's foreshore. With a time stamp of 5 minutes, WR leaves the front of Cherbourg.

With a time stamp of 5 minutes, the duration of the calculations is too long. More than 11 minutes and WR is unable to finish.

I placed boundary across the Channel, east and west of the area. This reduces the amount of computation to trace the isochrones.
In this case, it takes almost 7 minutes for WR to arrive about 2 miles from Cowes, passing west of the Solent.
But at this point, WR still hangs.
Looking at the statistics, we see that the calculations are not complete.

If I take the trace of the weather-road, we see that this trace goes to the finish while the weather-road has not arrived at Cowes.

There must be a bug at this level.

If you want to test this situation, I put you in a zip file:
- the xml file of WR
- the weather file used (with winds and currents)
- the boundaries used.
To recover this zip file, follow this link: DL.FREE.FR

[skipperearly]

Tried to enter bug on fly but somehow failed to upload my files, so I record it here and hope it helps to get WR-pi improved.

Used same polar, grib and start/end positions and wanted to calculated options for different start dates.
1st calculation finished without troubble
2nd calc failed with land-error
3rd calc finishes but without giving end date/time

see attached screenshot and related files.

grib file was too large for uploading

[Gilletarom]

Hello Sean D'Epagnier,

A few days ago, in this thread:
http://www.cruisersforum.com/forums/...ml#post2594019
I asked what should be understood in the definition of the polar, of:
True Wind Speed
True Wind Direction
Apparent Wind Speed
Apparent Wind Direction

Thanks again to RGleason for telling me.

I did today, two tests with the figaro2 boat. In the first test, I used "True wind", and in the second test, I used "Apparent wind".
It seems that the two weather-roads calculated by WR are the same. At one detail, the end is slightly different. But, this detail is probably related to my previous post.

I made a third choice again. I took "True wind direction". The weather-road remains the same. WR also hangs without arriving at Cowes.

Several questions come to my mind:
- If I change my choice, does this have an impact on the configuration?
- What consequence does this have on the "boat" used?
- Is it necessary to offer users the four choices in WR?

Best regards ?

[Gilletarom]

Hello Skipperearly

Originally Posted by skipperearly

Tried to enter bug on fly but somehow failed to upload my files, so I record it here and hope it helps to get WR-pi improved.
....
see attached screenshot and related files.
grib file was too large for uploading

You can :
- Place all your files in a zip file
- Use this service : Free - Envoyez vos documents
- Place here the link that you receive from dl.free.fr

So we can test with all your configuration.

Best regards. Gilletarom

[rgleason]

Several questions come to my mind:
- If I change my choice, does this have an impact on the configuration? -No, I would not expect so.
- What consequence does this have on the "boat" used? None.
- Is it necessary to offer users the four choices in WR?
Gilletarom this is a good question. I always thought it was to show the polar graph in a different way, but the graph does not change significantly for me, It is impossible to determine any difference... so that leads to your question.
What are these choices for? - They certainly are not settings that affect the Computation.
I would also like to be able to see the entire name in the dropdown!!! - for sure.

Gilletarom, are these examples using your original setup with boundaries and a grib file? Because my computer SurfacePro3 i5 was able to complete all of your routes no problem. I can try again later if I know it is the same examples.

[rgleason]

Originally Posted by boat_alexandra

I would say, try to follow the instructions in the wiki. If they don't work or you get stuck, just post a question, then fix the wiki.

Several people have compiled my plugins on mac, and I don't think it's difficult.

For extra credit, figure out how to make travis autobuild for mac.

I hope Wollybee can do it! I like your 'extra credit' challenge. I think Kevinh is going to be able to do it once he gets connected to the internet.

[Gilletarom]

Hello RGleason,

Originally Posted by rgleason

.....
Gilletarom, are these examples using your original setup with boundaries and a grib file? Because my computer SurfacePro3 i5 was able to complete all of your routes no problem. I can try again later if I know it is the same examples.

To day I uninstall all and re-install all. So re-create the start point and the end-point.

I used these files :
20-01-2018_Boundary_Cotes_Francaises_29.gpx
22-01-2018_Boundary_Cotes_Anglaises_30.gpx
arome_2017-01-11-C.grb
Boat_Figaro2.xml
WeatherRouting_Cherbourg_Cowes.XML

and I see this file in my PC :
C__ProgramData_opencpn_plugins_weather_routing_pol ars_figaro_2.pol.contours

You can download all the files here : DL.FREE.FR