Update cpu instruction definition and asm programs for customasm v0.11

[sparr]

List of changes:

• #bits in global scope
• #bankdef and #bank names ".data"/".instr" are now data/instr
• consistent indenting in #bankdef
• nested #tokendef is now #subruledef in global scope
• #cpudef is now #ruledef
• token/subrule def = is now =>
• cup/rule def -> is now =>
• 3'5 constant is now 5`3 (v0.11 syntax error using a'b instead of b`a syntax for a constant with a specified width hlorenzi/customasm#158)
• #str is now #d
• customasm submodule is now the upstream main branch current commit

The submodule change abandons your 4 divergent commits. They might be worth a PR?

resolves #2

[sparr]

It's difficult to compare assembled results of the example programs because it seems some of the existing ones were assembled with and without (and partially with?) the ldi optimization discussed in #6

[vascofazza]

Thank you for the PR.
It is actually difficult to compare due to the indentation. Did you make any change to the instructions due per #6 ?

I actually already merged the PRs into the original repository, it's curious the changes are not there anymore?
hlorenzi/customasm#35
hlorenzi/customasm#37

[sparr]

re issue 6, I did not change line 41, ldi {dst: reg} 0 -> 0b11 @ 0b01 @ 0b00 @ dst[1:0], other than to fix the =>. The third crumb is still 0b00 so it should still behave identically.

re the upstream repo, I'm confused because I do see your commits merged in the PRs you linked, but https://github.com/vascofazza/customasm still says it's 4 commits ahead and hlorenzi/customasm@main...vascofazza:customasm:master lists some of your merged commits. Probably not worth worrying about right now if you're willing to move the submodule dependency to the upstream commit.