bugfix(lowering): propagate snapshot pop success values

[jcastil0]

Summary

Propagate known values through array_snapshot_pop_front and array_snapshot_pop_back on the success path during const folding, so later lowering passes can keep folding the remaining snapshot array and the popped boxed snapshot.

---

Type of change

Please check one:

• Bug fix (fixes incorrect behavior)
• New feature
• Performance improvement
• Documentation change with concrete technical impact
• Style, wording, formatting, or typo-only change

---

Why is this change needed?

Const folding already tracks empty array_snapshot_pop_front / array_snapshot_pop_back cases, but it drops the known values on the non-empty success path.

As a result, later lowering optimization cannot keep folding operations that consume the remaining snapshot array or the popped element, even when both are fully known at compile time.

---

What was the behavior or documentation before?

For known non-empty snapshot arrays, const folding did not propagate the success outputs of array_snapshot_pop_front and array_snapshot_pop_back.

The empty path kept shape information, but the success path lost it.

---

What is the behavior or documentation after?

For known non-empty snapshot arrays, const folding now propagates:

• the remaining snapshot array to the first success output
• the popped element as Box<@T> to the second success output

---

Related issue or discussion (if any)

---

Additional context

This change stays within lowering const-folding and adds focused tests for both array_snapshot_pop_front and array_snapshot_pop_back success paths.

---

Note

Medium Risk
Touches lowering const-folding logic for array_snapshot_pop_front/back, which can affect compile-time evaluation and downstream optimizations if the propagated VarInfo is wrong, but scope is limited to these two libfuncs.

Overview
Const folding now propagates the success-path outputs of array_snapshot_pop_front/array_snapshot_pop_back: it tracks the remaining snapshot array and the popped element (as Box<@T>), instead of dropping this information on non-empty arrays.

Adds focused optimizer test cases asserting that these propagated values enable further folding (e.g., folding two pops on a known [1, 2] snapshot to a constant result).

^{Written by Cursor Bugbot for commit 1db7cc2. This will update automatically on new commits. Configure here.}

[reviewable-StarkWare]

This change is 

[orizi]

@orizi made 1 comment.
Reviewable status: 0 of 2 files reviewed, 1 unresolved discussion (waiting on jcastil0).

---

crates/cairo-lang-lowering/src/optimizations/const_folding_test.rs line 90 at r1 (raw file):

}

fn alloc_var<'db>(

add a test in the const_folding test file - not a specialized test for your case.

[jcastil0]

@orizi made 1 comment.
Reviewable status: 0 of 2 files reviewed, 1 unresolved discussion (waiting on jcastil0).

crates/cairo-lang-lowering/src/optimizations/const_folding_test.rs line 90 at r1 (raw file):

}

fn alloc_var<'db>(

add a test in the const_folding test file - not a specialized test for your case.

Moved it into the shared const_folding test file and dropped the specialized test.

@orizi made 1 comment.
Reviewable status: 0 of 2 files reviewed, 1 unresolved discussion (waiting on jcastil0).

crates/cairo-lang-lowering/src/optimizations/const_folding_test.rs line 90 at r1 (raw file):

}

fn alloc_var<'db>(

add a test in the const_folding test file - not a specialized test for your case.

@orizi Hi
Moved it into the shared const_folding test file and dropped the specialized test.