publishing an intermediate release 3.1.0?

vlad_vy
Offline

29 Nov 2013 13:48 - 29 Nov 2013 13:57 #16036 by vlad_vy

Replied by vlad_vy on topic publishing an intermediate release 3.1.0?

Devo protocol telemetry monitor color indication is messed.

bitbucket.org/PhracturedBlue/deviation/i...devo8-devo-telemetry

It looks like DSM2/X telemetry monitor color indication also is messed.

Last edit: 29 Nov 2013 13:57 by vlad_vy.

The topic has been locked.

HappyHarry
Offline

29 Nov 2013 14:57 - 29 Nov 2013 14:57 #16038 by HappyHarry

Replied by HappyHarry on topic publishing an intermediate release 3.1.0?

So its ok to pull from pb's repo again now?

Last edit: 29 Nov 2013 14:57 by HappyHarry.

The topic has been locked.

PhracturedBlue
Offline

29 Nov 2013 15:15 #16040 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

I specifically chose not to merge the 'beta' tags after I manually implemented Rbe's change because there was no way to make them identical to rbe's original.
I haven't decided to continue using them (the nightlies will be updating, so I'm not sure of the value of having them in my trunk)

The topic has been locked.

PhracturedBlue
Offline

29 Nov 2013 15:38 #16041 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

vlad_vy wrote: Devo protocol telemetry monitor color indication is messed.

Is this something new? I don't see any changes to the telemetry code for a long time.
I don't see any odd behavior on the emulator. Once I break out my transmitters again, I'll take a look.

The topic has been locked.

vlad_vy
Offline

29 Nov 2013 15:45 #16042 by vlad_vy

Replied by vlad_vy on topic publishing an intermediate release 3.1.0?

I don't know how long time it has place. I didn't use telemetry monitor very long time.

The topic has been locked.

rbe2012
Topic Author
Offline
So much to do, so little time...

29 Nov 2013 16:55 #16053 by rbe2012

Replied by rbe2012 on topic publishing an intermediate release 3.1.0?

PhracturedBlue wrote: I specifically chose not to merge the 'beta' tags after I manually implemented Rbe's change because there was no way to make them identical to rbe's original.
I haven't decided to continue using them (the nightlies will be updating, so I'm not sure of the value of having them in my trunk)

Actually you have a tag "v3.0.0-beta-05" in your trunk - to avoid confusion I recommend to remove the beta tags or set a completely new one (does not matter if it's named beta-16 or else).

The topic has been locked.

PhracturedBlue
Offline

29 Nov 2013 17:06 #16055 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

sorry, I meant that the changes that were exactly the same still have them, and those that are different don't.
But yes, I removed all the beta tags. I don't think about my code as having stable point. If I push to the repo, I expect it to be stable (though it isn't always so)

The topic has been locked.

rbe2012
Topic Author
Offline
So much to do, so little time...

29 Nov 2013 17:11 #16057 by rbe2012

Replied by rbe2012 on topic publishing an intermediate release 3.1.0?

Works on my Devo8 so far, just an issue with icon selection found. I will collect them before reporting.

The topic has been locked.

vlad_vy
Offline

29 Nov 2013 18:04 - 29 Nov 2013 18:09 #16065 by vlad_vy

Replied by vlad_vy on topic publishing an intermediate release 3.1.0?

Trim setup do not work, with emulator and Devo8. I can change trim step, but nothing more.

Last edit: 29 Nov 2013 18:09 by vlad_vy.

The topic has been locked.

PhracturedBlue
Offline

29 Nov 2013 18:18 #16069 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

It works for me. Can you be more precise what doesn't work?

The topic has been locked.

PhracturedBlue
Offline

30 Nov 2013 01:59 - 30 Nov 2013 02:12 #16080 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

vlad_vy wrote: Trim setup do not work, with emulator and Devo8. I can change trim step, but nothing more.

~~This appears to only occur when using the standard mixer. I'll look into it~~
Never mind. I have no idea what I was doing earlier, but I see it now too.

Last edit: 30 Nov 2013 02:12 by PhracturedBlue.

The topic has been locked.

PhracturedBlue
Offline

30 Nov 2013 02:44 #16081 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

Rbe:
Can you explain the behavior that led you to implement f2c209f7c61e (resolve 423)? I can't reproduce that behavior on earlier builds, and this change seems to have completely broken saving trims.
That makes sense since the logic is to save the trim data in a temporary struct and copy it into 'Model' once the save happens. But obvviouslythere must have been an issue or you wouldn't have changed it.
For now I've reverted that patch.

The topic has been locked.

rbe2012
Topic Author
Offline
So much to do, so little time...

30 Nov 2013 06:56 #16083 by rbe2012

Replied by rbe2012 on topic publishing an intermediate release 3.1.0?

Yes, I'll look into. It's some time ago...

The topic has been locked.

rbe2012
Topic Author
Offline
So much to do, so little time...

30 Nov 2013 08:11 #16084 by rbe2012

Replied by rbe2012 on topic publishing an intermediate release 3.1.0?

The motivation for this change was that when editing a trim value and changing the trim step, this value was not shown on the overview page after exiting with ok, but with cancel. It seemed clear to me that the conditions are inverted so I re-inverted it.
It was the wrong solution.
The reason for this behavior is that set_trimstep_cb() works on the Model.trim, not on tp->trim while okcancelcb() copies the tp->trim where the trim step is not changed. The problem is: if we simply change this (work on tp->trim.step), in the overview page all trim step values are identical.
The solution is not to copy the complete struct trim to Model.trims[x] but to spare trim.step (pages/common/_trim_page.c):

static void okcancel_cb(guiObject_t *obj, const void *data)
{
    (void)obj;
    if (data) {
        //Reset trim here (automatically saved in the Model.trim[])
        struct Trim *trim = MIXER_GetAllTrims();
//        trim[tp->index] = tp->trim;
        trim[tp->index].neg = tp->trim.neg;
        trim[tp->index].pos = tp->trim.pos;
        trim[tp->index].src = tp->trim.src;
        trim[tp->index].sw = tp->trim.sw;
        MIXER_RegisterTrimButtons();
    }
    PAGE_RemoveAllObjects();
    PAGE_TrimInit(0);
}

I pushed a correction to my repo.

The topic has been locked.

rbe2012
Topic Author
Offline
So much to do, so little time...

30 Nov 2013 10:43 #16087 by rbe2012

Replied by rbe2012 on topic publishing an intermediate release 3.1.0?

PB, do you remember why the rf monitor was disabled? I had some requests for it and it is not too hard to re-enable, but I did not want to do that without knowing the reason for disabling.

The topic has been locked.

PhracturedBlue
Offline

30 Nov 2013 11:35 #16088 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

Thanks for debugging Rbe. I've fixed this in my repo (though I chose a different solution than you)

As for the radio monitor, in practice it is virtually useless. At worst, it gives a false sense of knowing what is going on. If you want, you can add a 'HAS_RADIO_MONITOR' but it should be left disabled by default.

The topic has been locked.

WheresWaldo
Offline

30 Nov 2013 13:44 #16091 by WheresWaldo

Replied by WheresWaldo on topic publishing an intermediate release 3.1.0?

I want to apologize in advance for this question especially if it is stupid. I am not a programmer but more of User Interface guy and I have a question about memory usage. Is there a significant difference in memory usage using separate screen pages (with no user interaction) or a modal dialog box?

The topic has been locked.

PhracturedBlue
Offline

30 Nov 2013 15:35 #16092 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

Generally, No.
We've already 'paid' for all of the graphical elements, so putting them on a page has minimal cost. Creating a new element type can be costly depending on what it is. Most of the cost of a new page is the logic that defines what it does.

The topic has been locked.

PhracturedBlue
Offline

30 Nov 2013 19:27 #16094 by PhracturedBlue

Replied by PhracturedBlue on topic publishing an intermediate release 3.1.0?

Rbe,
Another bug for you to look at:
#431
It is caused by patch a51e777df3e8
Specifically by this:

   if (type == ELEM_MODELICO) {
        if (*h > LCD_HEIGHT - 32) *h = LCD_HEIGHT - 32;
        if (*w > LCD_WIDTH)       *w = LCD_WIDTH;
        if (*x + *w > LCD_WIDTH)  *x = LCD_WIDTH - *w;
        if (*y + *h > LCD_HEIGHT) *y = LCD_HEIGHT - *h;
    }

in GetWidgetLoc()
I'm not sure what this is for. I guess the -32 is for the Devo8 header (which isn't the same size on the Devo10), but I'm not exactly sure of the intent here. If the image is too bug, it seems to drop it into the lower right corner of the screen. I'd think you would leave x and y alone, and just adjust w and h if the image doesn't fit on the screen. If you are trying to center the image than this won't work if the image is smaller, so I guess that isn't the intent.
My fix would be:

   if (type == ELEM_MODELICO) {
        if (*y + *h > LCD_HEIGHT) *h -= LCD_HEIGHT - (*y + *h);
        if (*x + *w> LCD_WIDTH)   *w -= LCD_WIDTH - (*x + *w);
    }

But I'm not sure that is what you intended.

The topic has been locked.

HappyHarry
Offline

01 Dec 2013 00:59 - 01 Dec 2013 01:08 #16099 by HappyHarry

Replied by HappyHarry on topic publishing an intermediate release 3.1.0?

I've came across a bug that happened before that was corrupting model ini's stopping access to dual rates screen, plus more importantly it totally disables the right stick output, as in on my mode 2 setup it disables aileron and elevator output completely. This is on a 7e, using a build done from pb's repo tonight, with the hisky patch applied. Version number finishes 2db21e1. I've finally worked out how to make it happen, just enable ppm as the protocol and enable standard GUI.

Last edit: 01 Dec 2013 01:08 by HappyHarry.

The topic has been locked.

Time to create page: 0.083 seconds

Force desktop display