openrcforums

ReSt wrote:So it's probably my fault when I'm having problems with the downloads. I'm used to make a "git clone" without any further selections on the clone window . Trying the selection "Branch = next" or "branch=master" did not get me really happy. Sometimes I get the old unchanged stuff and sometimes I get new stuff, that doesn't compile. So I assume, that the source code on github is not always in a consistent state.

I thought I only ever do a "push" to Github when I have files that compile. Also, I haven't been changing the master branch until I do a release, so that should have been consistent.

If you have cloned the repository, you then need to select what you want in your working directory, e.g. "switch/checkout" to the latest.

Mike.

MikeB wrote: If you have cloned the repository, you then need to select what you want in your working directory, e.g. "switch/checkout" to the latest.

Mike.

Ok. Step by Step
I open the subdirectory that contains the .git subdirectory. A right click in the empty space opens the menu where I find the option "TortoisGit"
This option opens another menu that contains, among others, "Switch/Checkout".
This opens a window with "switch to => Branch, Tag and Comit"
I select "Branch" and the input selection list of "Branch" offers me "master, remotes/origin/master, remotes/origin/next and remotes/origin/Qt5"

That's the point where I'm lost and don't know how to go further

Reinhard

You select "remotes/origin/next" for the latest stuff.

When I select "Branch = remotes/origin/next", the options field is automatically set to "Create New Branch = next"
Whether I select or deselect the "Create New Branch", when I click the ok button something works and tells me "success" but there is no change in the source code.

I had expected that the source code in my src directory would change somehow.
Do I have to export something ( the next branch) or how do I go on?

Reinhard

Unless you do something specific, you are working with you local copy of the repository. Try doing a "fetch" to update your repository.
Note that yesterday, in preparation for a release, I merged "next" into "master", so both branches are currently the same.

Mike.

I've added links on er9x.com to ersky9x and er9x test versions threads.

Mike.

Thank you

João

I posted an update to "SportSet" on the first post.

Mike.

Another update to SportSet on the first post:
Graphics for the mounting options added, I also reduced the size of the graphics data so the .exe gets smaller.
Offset angle and enable values added.
Single COM port selector with 'scan' button to re-discover available COM ports. This is useful if you start the program before you plug the STK in.

Mike.

Do you plan to make SportSet a part of eepe / eepskye package?

Do you mean put the code into eepe/eepskye or just include "SportSet" in the installation?
At present it is easier for me to work with the stand alone program.

Mike.

I meant just to include the program (as it is) in the installation. I think it will be nice to have it there!

Also, I would make the installer to add a desktop shortcut not only for the eepe.exe, but also for the eepskye and the sportset.

P.S. Though, I'm absolutely fine to download it separately and make the shortcuts myself!

I agree with andrewju

João

I had thought I'd changed the installer to add the shortcut for eepskye, but I haven't properly tested it as I aleady have a shortcut!

Mike.

Mike,
I downloaded the latest github source code and on compiling I get an errormessage about missing file "S6Rimg.lbm" (that's not in my download)
If I outcomment all "S6R"-calls, it compiles but gives me 7 warnings caused by mavlink.cpp of "dereferencing type punned pointer will break strict aliasing rule"

If I outcomment the mavlink link the rest compiles without further problems.

Could you have a look for it, please ?

Reinhard

I've added the missing file on the master branch.
You may ignore those 7 warnings about the pointers.

Mike.

I've managed to get a version of SportSet built for Linux, and it even worked in a VirtualBox. I've posted it on the first post.
Not sure about installing it on Linux, you will need some QT libs, but you may have them if you have installed eepe/eepskye.

Mike.

Mike,

I still don't have an STK, so I can't test the SportSet completely. But when I try to "Read" an S6R configuration without selecting a COM-port, the SportSet force-closes. This is a minor issue, but maybe still worth fixing...

(I'm talking about the Windows version)

I think I have the fix for that.
I've just noticed I've put the original code of "SportSet", for the UART sensors, into eepe/eepskye. Maybe I'll add the S6R config in as well!

Mike.

MikeB wrote:I've added the missing file on the master branch.
You may ignore those 7 warnings about the pointers.

Mike.

Thanks Mike.

Reinhard

I've also got a fix for those warnings. I copied the code from Mavlink itself. The "payload" of a packet is an array of 64-bit values, but these are cast to and array of 8-bit values, then, when needed, back to a 16 or 32-bit value. It's going from the 8 to the 16 or 32 that causes the warning. I've changed the code to just go from 64 down to 32 or 16. The compiled code is exactly the same, just no warnings!

Mike.

Thanks again.
Looks like I already got your fix with my latest download as the warnings disappeared and everything compiles without any errors or warnings.

Reinhard

First post updated with new SportSet files for both windows and linux(32-bit).
I've added in the accelerometer calibration and fixed the bug mentions above by "andrewju".

Mike.

The force-close issue is partially fixed. When I click 'Read', it says 'Reading' and keeps on running. But when I click 'Write', it still force-closes. That happens without an STK.


Mike,

The S6R calibration procedure in the ersky9x goes in an infinite loop. I mean, it doesn't stop after the 6th position. It goes back to the first one, then the second one again, etc. (unless I did something wrong here at 4am!). May it be a good idea to STOP calibration (perhaps, with a message like 'calibration cycle completed') after the 6th step?

And one more question. I think I need a confirmation on the Mounting options. I understand the S6R should always be mounted with antennas facing forward. So, in my understanding:
- Horizontal means the S6R label side facing up
- Horizontal Reversed should mean upside down (label facing down)
- Vertical means Rx on the side, with antennas facing forward and S6R label on the left
- Vertical Reversed is basically the same as Vertical, but Rx is mounted on the other side (so S6R label is on the right)

Am I correct? Any chance to add some graphics and visualize this part?


Thanks a lot!

I copied the FrSky program and that keeps cycling as well!

I have added graphics for the mounting, you may only see them when you have the STK running, although I see them if I just run the program and then click next and previous. You should have graphics for the wing type and for the calibration positioning as well.

Mike.

Ahh... Regarding graphics and calibration cycling, I was actually referring to ersky9x. There's no graphics there for the mounting options.

Also, I still have a feeling we should notify the user that calibration is complete, rather than give him the same screens again and again. Can you think of any reason why it should go in circle?
If a user thinks he made an error during calibration, he will likely have to go through all the 6 steps anyway. So he can just run another calibration round.

In my opinion, this should apply to both SportSet and ersky9x.

BTW, the S6R calibration LUA in Taranis stops after the 6th position and says "Calibration completed".

just imho: my vote is for any calibration process (ersky9x or Sportset) to stop after one complete cycle and notify the user - or, even a choice of 'stop' or 'run again', if you want to get fancy

But, it's all cosmetic after the first time you use it - then, you know -- as I found when I first started to use the ersky9x version.

regards
ozphoenix

andrewju wrote:Ahh... Regarding graphics and calibration cycling, I was actually referring to ersky9x. There's no graphics there for the mounting options.

Also, I still have a feeling we should notify the user that calibration is complete, rather than give him the same screens again and again. Can you think of any reason why it should go in circle?
If a user thinks he made an error during calibration, he will likely have to go through all the 6 steps anyway. So he can just run another calibration round.

In my opinion, this should apply to both SportSet and ersky9x.

SportSet now stops after the 6th position, ersky9x will take a bit more work, and may wait until after I do a release.
I've got some graphics into ersky9x for the mounting option.
SportSet now also has a RPM tab to allow you to configure the SPort RPM sensor. I've still to post this update and it may be a day or so until I have time to do so.

Mike.

Sounds great! Thanks a lot!!!