Add MATLAB syntax highlighting

[MingqiZhang0]

Summary

Adds MATLAB syntax highlighting support to Focus.

This introduces a dedicated MATLAB tokenizer and wires it into Focus's language selection, comment commands, and color preview.

Highlights

The tokenizer supports:

• MATLAB reserved keywords
• true / false values
• numbers, including decimal, scientific notation, and i / j suffixes
• % line comments and %% section comments
• %{ ... %} block comments with MATLAB's standalone-line delimiter rules
• double-quoted strings
• single-quoted character vectors with doubled quote escaping
• MATLAB operators and punctuation
• ... line continuation
• function-call highlighting
• function declaration highlighting, including:
  • function foo
  • function foo(x)
  • function y = foo
  • function y = foo(x)
  • function [a,b] = foo
  • function [a,b] = foo(x)

Special handling is included for MATLAB's ' ambiguity between character vectors and conjugate transpose.

A quote directly following an expression is treated as transpose, while a quote preceded by whitespace starts a character vector.

Examples:

B = A';
C = A.';
seq = [seq 'def'];
cell_data = {A 'text'};
combo = [A' 'text'];

.m file association

Focus currently uses .m for the existing Objective-C highlighting path, so this change intentionally does not globally remap .m files to MATLAB.

MATLAB projects can opt in explicitly:

*.m : matlab

A commented example has been added to the default config.

Comment commands

• Toggle comment uses %.
• The generic block-comment command falls back to line comments for MATLAB because %{ and %} have standalone-line requirements.

Validation

A fresh Jai compilation was not performed, because a Jai compiler was not available locally.

Instead, the MATLAB tokenizer logic was ported rule-for-rule to a standalone Python test harness for behavioral validation.

The Python tokenizer and a small Tkinter editor shared the same tokenizer implementation, allowing both token-stream regression tests and visual/live-edit behavior to be checked without using a third-party MATLAB lexer.

Validation performed with the Python surrogate:

• 89 token-stream regression assertions: PASS
• quote / transpose ambiguity tests: PASS
• function declaration highlighting tests: PASS
• block comment and ellipsis tests: PASS
• identifier edge-case tests: PASS
• live editing stress tests: PASS
• visual syntax-highlighting inspection: PASS
• sample MATLAB syntax additionally verified with MATLAB R2025b

The visual/live-edit tests included cases such as:

• A'
• A.'
• A''
• [seq 'def']
• {A 'text'}
• [A' 'text']
• true'
• false'
• temporarily removing and restoring string delimiters
• changing A' to A ' and back
• removing and restoring %} in a block comment
• rapid quote insertion/removal during live highlighting

The Python test harness is only a behavioral surrogate for the tokenizer.

It does not prove that the Jai source compiles, and it is not an actual Focus runtime test.

Current verification status:

• Python tokenizer surrogate: PASS
• Visual/live-edit surrogate: PASS
• MATLAB syntax probes with MATLAB R2025b: PASS
• Fresh Jai compilation: NOT PERFORMED
• Actual patched Focus runtime: NOT PERFORMED

Scope

This PR only adds MATLAB syntax highlighting.

It does not add:

• MATLAB execution/build integration
• LSP support
• completion
• formatting
• smart indentation
• builtin function databases
• .mlx support
• automatic .m content detection
• Octave-specific syntax

Screenshot

This screenshot comes from a standalone Python editor that ports the MATLAB tokenizer rules for visual demonstration only. It is not a Focus runtime screenshot. Because a Jai compiler was not available locally, the Jai implementation itself could not be compiled or tested.