fix(test): disable test_string_diff (does String arithmetic, not allowed after typechecker tightening)

`tests/codegen/test_string_diff.affine` was checking that distinct
string literals get distinct pointers by computing `s2 - s1` (expecting
the byte offset between them). The recent typechecker tightening
correctly rejects:

    Unification error: (Unify.TypeMismatch (String, Int))

The right way to assert "different strings have different storage" is a
cast or an `__addr_of` intrinsic, neither of which exists in the v0.1.0
surface. Renamed to `.affine.disabled` (matching the pattern of
`lib/formatter.mli.disabled`) so the codegen test runner's
`*.affine` glob skips it. Re-enable when the cast story is settled.

Signed-off-by: Jonathan D.A. Jewell <6759885+hyperpolymath@users.noreply.github.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Signed-off-by: Jonathan D.A. Jewell <6759885+hyperpolymath@users.noreply.github.com>

Author: hyperpolymath