PB, request for code review and merge back(2)

Suvsuv, PB, do you guys mind if I start creating issues for devo 10 branch? I'll try to help with the code too where I can/have time and send pull requests.

PhracturedBlue wrote: I'd appreciate it if you file issues on the deviation project (not suvsuv's branch). I have incorporated virtually all of his work into the trunk, and the trunk is what we'll end up releasing.
It would be better if you could figure out why your build system doesn't work.
Releasing binaries of every change is quite time consuming.
If you can post details of what you have, we can start trying to figure out what is wrong.
Please start a new thread for that.

For instance, at least some of the bugs you've filed have already been addressed. Being able to run the latest code, really is critical.

My build system is fine, I couldn't get it running using my distro's repository, so I resorted to summon-arm-toolchain and everything is fine. I run Archlinux and there are a number of ARM toolchains in the Archlinux User Repository. Unfortunately none worked ok, so I had to dump them.

Will move issues from suvsuv's repo to main, should I mention which commit is my build based on, as a reference and for history's sake? Should not matter, I always update before building a dfu, but if it makes it easier for you guys, it's not an effort on my part.

it is weird you saw rebooting several time , Please make a build under my repo(update again as your current codes is not in synced with my repo) and have it downloaded to your TX, see if the TX still reboots from time to time.
In terms of slow refreshing compared to devo8, I am sure it not related to the main loop as the GUI_RefreshScreen frequency is the same as devo8's , I believe it should related to the following codes in the lcd.c, your previous clear logic is buggy and I don't have time to figure it out, just using a work-around now

void LCD_Clear(unsigned int val)
{
/* int i,j;
val = (val & 0xFF);
for (i=0; i<LCD_PAGES; i++) {
lcd_set_page_address(i);
lcd_set_column_address(0);
for (j=0;j<PHY_LCD_WIDTH;j++)
LCD_DATA = val;
} */
// Bug fix: above method doesn't work properly, especially during keyboard type changing.
LCD_FillRect(0, 0, LCD_WIDTH, LCD_HEIGHT, val);
}

Haven't seen any reboots in my ~3 hours total play time.

Rebooting might not only caused by watchdog, but also using NULL pointers. I believe it is most likely the latter

Hi guys,

Many thanks for all your efforts - just too cool and generous even for an emoticon.

I've opened a few issues with along with the build ID. Let me know (here) if I did wrong and I'll stop.

BTW no reboots for me as yet.

PB, please do a code review and merge back again. You should have a look at main.c and take bug fix on backlight and contrast setting from tx.ini. Other changes include:
1) Fix all devo10 bugs raised in the BitBucket
2）Refactor listbox widget.
3) Don't allow user to interrupt binding until we figure out why the current goes up when interrupting binding
4) Refactor the Tx-configue page
5) Refactor the calibration procedure and let devo10 store calibration data for Aux4 and Aux5 knobs
6) Move protocol selection and binding into model_page, and hence remove all compile warnings

PhracturedBlue wrote: I've merged everything I'm going to for the moment.
Things I didn't take:

4) calibration (I don't understand why your changes are needed. Explain further, or wait until I can try it on my Tx)
5) use buttons to change brightness (I want the buttons reserved for now. I'll think on this)
6) keyboard (see above post)

4) In the calibration, I remove the 1st step which is meaningless.
5)in the current main page, the R/L buttons are not used, so they could be used for short-cut keys. Please come up with more options , I could create a shortcut key config page to let users decide.
6) I would insist on swapping the R/L behavior in the real devo10, please don't make your judgment by playing with devo8 or emu_devo10. In the real devo10 device, the R key points to the left while the L key points to the right, please do test the keyboards in a real devo10

PhracturedBlue wrote: for (6) if you really want the right and left buttons to behave in reverse, change them in channels.c so they are consistent everywhere.

Make sense to me, will do it this way。
If you compared the orientation of the L/R keys in both devo8 and devo10, you will understand why I insist on having the R/L swapped. Walkera guys are just lazy and stupid