After commit d2e7abd the for loop following my change use pageMaxRow to determine which flags to clear.
But this pageMaxRow value is updated after the loop, which caused a crash in a very specific circumstance:
1 open the CMS OSD menu and go to profile
2 close the profile by yaw right, select exit
3 open CMS OSD menu and go to profile again
This caused a crash as runtimeEntryFlags wasn't cleared properly and the Copy Profile entry's flags still indicated OPTSTRING somehow.
which caused it's CMSEntryFuncPtr function to be cast to OPTSTRING and called without arguments.
Crash doesn't occur when VCP is disabled.
It was observed that shortly after the VCP port was enabled a crash
would occur in unrelated code. e.g. in `rxInit()`
Moving the delay before the call to HAL_PWREx_EnableUSBVoltageDetector
does not fix the issue - tested by moving the delay up one line and
doing many cold boots until a crash was observed. Crash un-repeatable
once delay is added.
Note: exact cause is unknown, behaviour occurs on some boards more frequently than others.
Emptry MODE_STRING paramaters would produce CLI output that would then generate errors if reapplied using a `diff` or `dump`. Fixed the output for blank values to be compatible with setting an empty string (clearing a previous value).
Also improve the length validation to report the allowed character range.
- Initial cut without Dshot
- Enable HAL-based DShot (no burst yet)
- Burst Dshot First working version
- Non-stop PWM refactor
- Conditionalize HAL structures inside motorTimerDma_s
[TIME] Basic def
[TIMER] timer related commits
[TIMER] Remove stale comments from timer_stm32h7xx.c
[TIMER] New timerClock() based on RM0443 Table 48
[TIMER] Add DMA handling
[TIMER] White space tidy in timer_def.h
Add TIM1_UP_IRQ handling
Fix DEF_TIM_DMA_FULL emitting error for 7th and 15th entry
STM32H7 - add TIM15-CH2 (No DMA)
Fix comment spelling errors.
timer.h change for "Burst Dshot First working version"
Fix bad TIM9 handling for non-H7 case
Touch-up after rebase (timer.h)
F4 and F7 does not have dmaIrqHandler in timerHardware_s (anymore)
Changes per PR comment
These parameters can now be configured directly with the set/get commands.
The "name" command will still function but presents a warning that it will be removed and to use "set name = " instead.
