Make X axis optional (WIP)

[ellensp]

Description

Add support for making X optional also (so having only an Extruder is a possibility.)

Benefits

More modularity, smaller code for extruder-only projects.

[ellensp]

I've done all the obvious code, now i'm stuck in "macro hell"

Can anyone lend a brain to assist?

[thinkyhead]

I believe this is essentially impossible given the current Marlin architecture. As a basic point, the type XYZval and all the types that derive from it become empty. An empty type that occupies no memory can't be allocated or referenced, so all kinds of core Marlin code becomes broken. There is a fundamental assumption that there will be at least one linear axis, and the number of conditional changes throughout the code to drop that assumption is simply overwhelming.

[thinkyhead]

…or maybe…. Apparently a zero-length array is possible, and maybe empty structs. At any rate, it's now compiling, so that's a pleasant surprise.

[ellensp]

just talking a look at it now, complies, uploads, responds to gcode ... I need to find an extruder I can play with...

[thisiskeithb]

Some hacky stuff was done to the BTT EBB42 pins in #24964 to make Marlin compile without an x/y/z axis, so that should probably get cleaned up here as well:

Marlin/Marlin/src/pins/stm32g0/pins_BTT_EBB42_V1_1.h

Lines 80 to 93 in d3527f5

	//
	// Steppers
	//
	#define X_ENABLE_PIN -1
	#define X_STEP_PIN PA10 // Unused. Assigned so Marlin will compile
	#define X_DIR_PIN -1
	#define Y_ENABLE_PIN -1
	#define Y_STEP_PIN PA10 // Unused. Assigned so Marlin will compile
	#define Y_DIR_PIN -1
	#define Z_ENABLE_PIN -1
	#define Z_STEP_PIN PA10 // Unused. Assigned so Marlin will compile
	#define Z_DIR_PIN -1

[ellensp]

@thisiskeithb that is part of the plan, but a sperate pr

[alextrical]

Fix for 0 kinematics (Flash?) EEPROM has been submitted

@alextrical — Where exactly was it submitted?

@thinkyhead Sorry I found that there was an issue and tried to roll back the PR, it caused an issue that prevented the board from running code when Virtual EEPROM was enabled, presumably caused by memory locations.
Most likely it was this branch https://github.com/alextrical/Marlin/tree/Make-x-optional-eeprom-fix

[thinkyhead]

what else needs done here?

As PRs go this one is a biggie! I was holding it until after 2.1.3, which was anticipated to happen a lot sooner. I've merged to bring it up to date, so now it should be tested again to make sure all the basic functions work when there's only an extruder and nothing else. That includes the EEPROM, to see whether the comment above still applies.

I don't anticipate any major troubles with this PR generally, but it does continue to lean into the old ways. Once this next release is out the door (and patched up in a followup minor release), we should look at refactoring more in the direction of templates, compiler dead-code stripping, and so on. It would be neat to get more formal object classes like Motor and Fan and so on to better encapsulate common components. Working on Fixed-Time Motion got me looking at the motion code more closely, and I believe we can make it a lot more concise, but it might be a little daunting at first.

I do want to get 2.1.3 released soon. I just have to re-test the probing and G34 behavior, fix up Fixed-Time Motion as much as possible, and then do a couple of test prints to make sure there are no unexpected instabilities. At the very least, I want 2.1.3 to be an improvement over 2.1.2, effectively addressing all reported issues, and at the moment I'm cautiously optimistic.

[vosk]

Hello all, I believe this or the previous PR in the series broke the backlash menu for XYZE printers, HAS_A_AXIS is now undefined in _CAN_CALI(A) used in menu_backlash.cpp. Since this is WIP apparently, feel free to ignore. Otherwise adding another 3 clauses for X Y Z ( A B ...) looks an obvious quickfix.
EDIT: This may be the actual cause:
4233e48#diff-d815bc244111b340f481976e8586e14084caeb10c74192a35a5bba39c0b7d0a1

[xxxajk]

For laser, Z is optional (which my table is manual Z adjustment).
You might be able to find some clues as to what I did to fix things to get them right in my own branch, but I need to update it in sync with the current branch...