I think I did that right. Not a fan of GitHub. Not yet anyway.
There are many conflicts with the pull request. Am unsure what to suggest for now. Not a fan of github either, but winforms is exceptionally hard to merge code. I realize this doesn’t help at all 
I think the problem is that as we make changes down the timeline. They are based on the latest that we have access to. Meanwhile changes are being made at the head of the project that were not anticipated. Our errors on this end are from changes being made on both sides of the files. The only way to do this differently is to stop the progress at the head while we get around to making changes at the tail. That is not good for the project progression. Therefor, it comes back to code blocks that can be inserted at the head if so desired by the head management. If someone has a great idea that is useful for a large group of the users then it may be worth inserting. Otherwise???
Now the problem becomes, if a user has an idea or work around that is needed for their specific application then how to make it easy for them to utilized the open source portion of the project to adapt and keep the software “up to date.” One piece of software that I use and need to update yearly has a “Legacy” file that contains local modifications that can be imported upon each update. So what I was thinking is a “Legacy.cs” that contains just the modifications that can be called from the main GUI. It would need sections broke out into individual groups. Just starting the thought process on this. But thinking about the add-on’s that are available for various programs. We are in the hay business and may not need some of the features that someone in the carrot business or tree planting may need. Our tool offset is extreme to the right. Some few using the software may benefit from a far right modification. Flag saving would be nice from time to time, especially first of the year to get the trash out of the field. So again a flag modification might benefit someone else. Just thinking out loud. Might should have been in a new features request.
Github went all sideways here recently, an update was made that unfortunately deleted a lot of files nad was tricky to get stable once again. The current master is quite stable and i am hoping to leave that one as a stable version. I agree, coding is constantly done and usually involves many files, the oldest one loses.
I also always feel bad when someone adds code like yourself, but then has to add it 3 more times because everything changes. 2 Weeks is like 3 years for normal software here. Many updates very quickly.
Is the code mainly in openGL.cs? I had to nuke the flags branch - but will make it again. I will not touch that one and then we can use it to compare against the master and even manually fill in the code. Is the code complete? and tested?
I’ll try to get it all in the new code again. Fairly few changes just scattered. The hardest part for me is the panel. I have not found a good way to copy a panel with its buttons into a new code. This basically turns a panel visible if a flag is picked. Then cycles through the flags. If you type in the box and hit save it creates a file. If you delete the flag it deletes the file and renumbers and renames the other files. Once a flag is picked, along with the Dimond at the base it calls a white draw 3d text with a size factor of 3. Fairly simple. I did need to start the flags at 1 and not 0.
Also, just so you know. The latest couple of versions does not save the flag file. If you have flags and delete any of them and close the field, when you re open all the flags are back. It is missing a SaveFlag () in the close field routine.
You have a separate file for every flag? Perhaps just load a single file and manipulate a list of classes of the data required. I made a flags branch. I won’t touch it.
Thinking about this, probably don’t need the complexity of touching the flag on the screen, perhaps just using a pulldown and selecting the number directly would be easier.
The right click on the opengl screen was missing a file save. The other 2 deletes were ok. Flags haven’t had a code change in probably 2 years, time for some updates.
Yes it creates a file for each flag that is edited and saved. Other that that it only displays what is in the flagpts[]. If there is a file display it and tell if it came from a file. If there is no file, then just display the default information.
You are correct on touching the screen. It just gives a starting point.
I could not get the visual basic and github to do what they were supposed to do. It kept giving the error that your flag tree was not found. So I uploaded it to a new repository in here:
Edit: Alright that did it I think. There were no conflicts found.
Edit2: Forgot to add the if routine around the keyboard shortcuts so you can edit the text box when there is a flag picked.
Edit3: The Keyboard Shortcut Fix has been added to the pull request.
@BrianTee_Admin @wilberttalen
Big changes to the flag notes and save. No more individual files. Changed the flags.text reader and writer to accept strings. Also added same feature to the kml writers. Seams to work much better.
It is here if the pull request does not work.
3 Likes
Very cool. I’ll have to play around with this. Nice work Kent.
Now just to get that crazy Google earth in the background! That seems to be a tough nut to crack
I merged the first pull request into flags, but you also have a pull request directly into master. It has to be done into the flags one so closed it. I have no idea how to merge the headlands and flags versions. This is the part that frustrates me to no end, great code and unable to merge them.
No one can merge or pull request into master without review by one other admin.
I must have clicked the wrong branch in github. Late night troubles!!!
As a question, because I don’t know github, do each of my commits to my files end up in the pull request?
I’ll probably have to build it again, but I’m getting better and faster at it. I need to build an add on function that others can add also. Just thinking. At certain locations throughout the main code call for the add on, like what is drawn, what the tractor looks like, flag code. I have no idea how the software companies do it. But just thinking.
Tough nut…
In version 3.08, I reloaded the necessary parts of Vrate to get a background. Simple button to if off show load screen, if on turn off. Took most of the Vrate stuff out and called it background. You know how all of that works. My thoughts, and very preliminary, draw a kml bigger than the field, say twice as big, load Google Earth to do the copy. Then bring back into AOG to crop like you moved headland. Texture it to center of the field, hanging over the field boundaries, by twice as big, if that is what was used to draw kml. Could use paint or something else to crop, but why not AOG. The bitmap edits built into c# seam fairly effective. I don’t understand any of them yet, but starting to look at them.
Your thoughts? ???
I think the tricky part would be making it exactly twice as big as the boundary - the boundary is the only reference we really have from the google earth image, so its easy to crop exactly to that boundary line then it fits exactly.
You are right. So I used the Maxfield size and added or subtracted it accordingly. Then I drew the cut line in the Google earth in the kml. It could be drawn separate just as well. Inching closer.
Edit: Here it is in Google Earth. Note the thin red line. So technically it is 3 times largest field dimension in one direction and 1 short dimension plus 2 largest in the other.
Wow, this is exciting stuff! The field to the right of the one you’re working on (alongside the river) looks pretty wild btw. must be around Brian’s stomping ground somewhere.
Yes, I was just working from the simulator start point. I think Brian may be familiar with the area. See other post, I actually saved the picture with the other field boundary still own.
Edit: That wild looking field might be why Brian got us all into this stuff in the first place. LOL!
Actually that’s our farm in the pic… The thing with 2D is you can’t see the 15 degree slopes in places.
So how are you georeferencing the google image to aog? The original VR application had the cropping done right to the border edge which gave its own reference since it knew exactly where the boundary was.
1 Like