fix(semantic): drive a macro expansion block by the repetition at its own depth

[orizi]

Stacked on #10281.

A $() expansion block whose only placeholder is bound to a deeper repetition than the block's own level expanded wrongly:

($([$($x:expr),*]),*) => { $($($x +)*)* 0 }
deeper_only!([1, 2], [3, 4])
  before: 1+2+1+2+0    // group [3,4] dropped entirely, group [1,2] emitted twice
  after:  1+2+3+4+0

Nothing was bound at the outer block's depth, so it fell back to driving on $x too and both loops walked the same group.

macro_rules! accepts this shape and expands it correctly at any depth (verified under rustc 1.96), so the fix is to iterate, not to reject. What rustc does reject — ($($a:expr),*) => { $($a + $($a +)*)* 0 }, "attempted to repeat an expression containing no syntax variables matched as repeating at this depth" — is exactly our E2202, and the existing max_path_len_in_block >= curr_rep_depth check is a faithful port of that rule.

Change

A block nested in d blocks drives the pattern repetition at depth d, reached by climbing rep_parents from a placeholder's own repetition (repetition_at_depth). depth is threaded through expand_macro_rule_ex.

This makes the (is_iterated, pattern_depth) ranking from #10270 dead — a block at depth d drives a repetition no enclosing block can be iterating — so find_repetition_driver loses it, and repetition_group's parent lookup tightens to an expect (the no-parent arm stays: every depth-1 block takes it).

register_repetition_placeholders no longer lets a shallower repetition overwrite a deeper binding. It re-runs on every iteration of the enclosing repetition, and a nested repetition matching zero times that iteration is never entered, so it would not restore what was overwritten. Without this the depth rule regresses two_deeper_only!([(1)], []) into expanding not at all.

Tests

Seven goldens in expansion_test_data/inline_macros, every expected value cross-checked against rustc. Four fail at the parent commit; the other three are guards on the new expect and on zero-match inputs, which have masked three separate bugs in this code already.

[orizi]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• fix(semantic): drive a macro expansion block by the repetition at its own depth #10282 👈 (View in Graphite)
• fix(semantic): expand a nested repetition over its own group of captures #10281
• fix(semantic): select a macro repetition's driver placeholder correctly #10270
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[reviewable-StarkWare]

This change is