Allow wait-for-cooling, break at threshold or if cooling stalls

[thinkyhead]

Instead of breaking immediately when M109 Rn or M190 Rn specifies a temperature below EXTRUDE_MINTEMP/2 or 30, proceed to wait, but break out when the temperature reaches the low threshold. Also check the temperature every 20 seconds. If it doesn't drop by at least 1°C in 20 seconds, stop waiting.

Reference: #4039 (comment)

[Blue-Marlin]

if ((old_temp - temp) < 1.0) break; we are cooling!
old_temp must be set to temp somewhere. Courrently it's set to 9999 and stays there.
Make clear the temperatures are float. (xxx.0 or xxxf)
5 seconds seems to be a magnitude to small. At the beginning the temperature still could rise. I suggest 60000 - 1° per minute.

[thinkyhead]

Let's see how 20 seconds works for the cooling. The whole low temperature threshold could be removed, as long as this can reliably detect that cooling has stalled.

[ghost]

After this PR was merged, I got compilation warnings.

\AppData\Local\Temp\build6674025c445f92d0b07185e4e4080804.tmp\sketch\Marlin_main.cpp: In function 'void gcode_M109()':

\AppData\Local\Temp\build6674025c445f92d0b07185e4e4080804.tmp\sketch\Marlin_main.cpp:4623:5: warning: 'wants_to_cool' may be used uninitialized in this function [-Wmaybe-uninitialized]

     if (wants_to_cool) {

     ^

\AppData\Local\Temp\build6674025c445f92d0b07185e4e4080804.tmp\sketch\Marlin_main.cpp: In function 'void gcode_M190()':

\AppData\Local\Temp\build6674025c445f92d0b07185e4e4080804.tmp\sketch\Marlin_main.cpp:4717:7: warning: 'wants_to_cool' may be used uninitialized in this function [-Wmaybe-uninitialized]

       if (wants_to_cool) {

       ^

I think that it's better that wants_to_cool is initialized, but I don't understand that which one is better true or false.

[thinkyhead]

@esenapaj That warning is overkill in this case. The compiler can't tell, but the logic is such that it will never permit wants_to_cool to be used uninitialized. Whatever value we initialize it with, it will be overwritten before it is ever used. Here's the key:

float theTarget = -1.0; // theTarget is always initialized to -1
. . .
do {
  // The actual target temperature can never be -1
  if (theTarget != thermalManager.degTargetHotend(target_extruder)) {
    // So this always happens the first time through the loop
    wants_to_cool = thermalManager.isCoolingHotend(target_extruder);
    . . .
  }
. . .
} while (!cancel_heatup && TEMP_CONDITIONS);

[Blue-Marlin]

Just a false warning. theTarget is initialized with -1. so theTarget != thermalManager.degTargetHotend(target_extruder) is always true because a temperature is always at least 0. The first time the do-while loop is executed theTarget is initialized.(Too complicated for the compiler)
If you want to avoid the warning you can initialize wants_to_cool with any value.

[thinkyhead]

I also just did a scrub-through of the temperature setters and accessors to make sure there's no place where an int is being compared as equal (or not-equal) to a float. I didn't find any problematic comparisons, but I will continue to keep an eye out for them.

[ghost]

@Blue-Marlin @thinkyhead
Thanks, I understand those.
So I leave these warnings for the present.

[Blue-Marlin]

Made some tests today.
Deactivated the lower temperature limit. Environment ~25.3°.
Testscript

M109 S50
M109 R0
M300
M106 S255
M109 R0
M300
;
M107
M190 S50
M190 R0
M300
M106 S255
M190 R0
M300

dropped out of the loop °C.

                0.25/5 1/20  1/30  1/60  1.5/60  °C/s
uncooled hotend noise  44.7  39.7  32.2  35.6
  cooled hotend noise  32.1  30.3  27.3  27.7
uncooled bed    noise  50.2  44.1  35.0  40.0
  cooled bed    noise  36.7  34.1  29.1  30.0

1/20 uncooled bed dropped out during the overshot. Later measured separately as 47.9°C

For longer times the constant must be long better unsigned long.
Numbers are only valid for the tested nozzle, bed, fan.
For the tested combination I'll stay with 1.5°C/minute - a good distance to possible noise, long enough for most over-shots (i hope), end-temperatuer around body temperature on moderate warm days.

[Blue-Marlin]

1.5°C/minute.