To maintain an accurate step count (which is required for correct
position recovery), any call to plan_set_position&co needs to be done
synchronously and from a halted state.
However, G92 E* is currently special-cased to skip the sync (likely to
avoid the associated performance cost), causing an incorrect E step
count and position to be set. This breaks absolute position recovery,
miscalculation of the LA factor and possibly other weird issues.
We rewrite the handling of G92 to always sync but still special-case the
frequent "G92 E0" for performance by using a free bit in the block flags.
To avoid a sync, we relay the request for reset first to the planner
which clears its internal state and then relays the request to the final
stepper isr.
When check_z_endstop is set, Z_MIN_PIN is checkend regardless of the
moving direction to support Z calibration. This prevents the ability
to use _just_ SG when moving upwards.
But since we know the extruder is at Z_MIN, it's always safe to raise
irregardless, so we can dodge the issue.
Do not unconditionally enqueue a Z move if no move has been requested.
Since Z is calculated using the absolute current (saved) position and
scheduled for later execution, the queue order becomes relevant.
In preparation for #2161, use MBL (G80) as a "new print" boundary
instead of just re-homing to ensure the reset is issued only once for
each print.
Similarly, do not reset the autodetection when LA is disabled via M900
K0. This can/will be used during a print if different quality settings
are used for different filling roles.
Since the advance factor is computed per-segment in LA15, there's no
need to stop the planner. Allow changing K freely at each segment.
This allows varying quality factors for different filling roles, see:
https://github.com/supermerill/Slic3r/issues/108
During pause/resume/crashdetect or powerpanic K might temporarily be out
of sync when used this way. If this becomes an issue, we might need to
store K for each block, as done for the feedrate.
Allow existing gcode using LA10 to transparently take advantage of LA15
by using a simple linear conversion function based on experimental
results with the MK3 implementation of linear advance.
Autodetect LA10 values based on the first M900 instruction contained in
the print. In order to support printing mixed files without resetting
the printer we also reset the autodetection status when starting a new
SD print and/or when explicitly disabling LA.
Since we cannot reliably detect whether a new print is started when
printing via USB, also reset the detection status when homing in G28,
which is generally performed once at each print. Note that this doesn't
clear the previous K value, it only allows a subsequent M900 to provide
LA10 values when printed after a LA15 file.
axis_relative_modes can cleanly handle G90/G91 and M82/M83 without the
need of "relative_mode" entirely.
Change G90/G91 to simply reset all axes according to the requested
state, which avoids to check two variables in get_coordinates(),
fixing the following incorrect handling:
G91 ; all axis relative
M83 ; set extruder absolute
G1 E1 ; should be an absolute move, but still relative
Correctly cleanup the printer state when stopping the current print:
- Disable interrupts while aborting the planner/queue to ensure
new serial commands are not inserted while aborting
- _Always_ call planner_abort_hard() to interrupt any pending move!
- Clear the saved_target, which might be set when calling stop
from within a paused state. Create a new function to clear the
paused state for future use.
- Do not disable/reset the MBL: doing so will destroy the ability to
restart correctly using M999.
