Adjust docs to 2.16.12

doc/realearn/modules/ROOT/pages/sources/virtual.adoc

@@ -12,7 +12,8 @@ See xref:further-concepts/compartment.adoc#virtual-control-element-id[].
 
 === Pick menu
 
-The convenient picker provides IDs from standardized xref:further-concepts/compartment.adoc#virtual-control-scheme[virtual control schemes]:
+The convenient picker provides IDs from standardized xref:further-concepts/compartment.adoc#virtual-control-scheme[virtual control schemes].
+It also contains a category for accessing the virtual control elements that are currently exposed by the xref:key-concepts.adoc#controller-compartment[].
 
 === ID field
 

doc/realearn/modules/ROOT/pages/user-interface/mapping-panel/glue-section.adoc

@@ -377,6 +377,18 @@ The following variables/functions are available in the formula:
 You can use that value in order to calculate the new value.
 With this, you can essentially craft your own relative mode!
 
+`y_type`:: This contains the type of the produced output value.
+By default, it's 0. The following types are currently supported:
+
+`0`::: Absolute continuous value (a value between 0.0 and 1.0)
+`1`::: Relative discrete value (an integer, e.g. -5 for 5 decrements)
+
+`realearn_timestamp`:: This contains the precise timestamp of the incoming control event in seconds.
+This can be used to calculate the delta time between two consecutive control events, which in turn can be used to simulate encoder acceleration, for example.
++
+The timestamp has audio block precision.
+For MIDI events, it even takes the intra-block offset into account - which is as precise as it can get, even if you use a high audio buffer size in the REAPER audio device settings.
+
 `y_last`:: This contains the last value of the target before it was affected by this particular mapping.
 +
 Allows you to come up with a performance control mode typical for synth parameter mappings, just like the built-in <<performance-control>> mode but more customizable.
@@ -413,13 +425,19 @@ In this case, one can return `none`, in which case the target will not be touche
 +
 Good for transitions that are not continuous, especially if other mappings want to control the parameter as well from time to time.
 
-ReaLearn's control processing order is like this:
+`realearn_last_feedback_value`::
+Contains the last numeric feedback value sent through this mapping.
+Before that, it's zero.
+This can come in handy for mappings with xref:further-concepts/target.adoc#virtual-target[], for which the variable `y` is always zero because virtual targets don't know the concept of a current value.
++
+WARNING: This variable is experimental, better don't rely on it!
+
+`realearn_dbg()`::
+This is a function which takes a floating-point value as parameter.
+It prints that floating-point value to `stdout` (*not* the ReaScript debug console!).
+Useful for debugging.
 
-. Apply source interval
-. Apply transformation
-. Apply reverse
-. Apply target interval
-. Apply rounding
+Learn more about the order in which ReaLearn processes the elements in the glue section by looking at the xref:glue-signal-flow.adoc[] diagrams.
 
 [#step-size-min-max]
 == Step size Min/Max controls

doc/realearn/modules/ROOT/pages/user-interface/mapping-panel/top-section.adoc

@@ -45,6 +45,8 @@ Prevent echo feedback::
 This option mainly exists for motorized faders that don't like getting feedback while being moved.
 If checked, ReaLearn won't send feedback if the target value change was caused by incoming source events of this mapping.
 However, it will still send feedback if the target value change was caused by something else, e.g. a mouse action within REAPER itself.
++
+Since ReaLearn 2.16.12, this also works in controller mappings with virtual targets.
 
 Send feedback after control::
 This checkbox mainly exists for "fixing" controllers which allow their LEDs to be controlled via incoming MIDI/OSC _but at the same time_ insist on controlling these LEDs themselves.