fix: skip emission of brillig calls which will never be executed

[TomAFrench]

Description

Problem*

Pulls across fix made in sync PR at https://github.com/AztecProtocol/aztec-packages/pull/7134/commits/e39be3dc0ee89d9a246366a1f88eb36b26f25751

Summary*

For some reason in the noir-contracts workspace, we're emitting a lot (100+ MB worth) of brillig call opcodes which have their predicate hardcoded to zero (i.e. they will never actually be executed).

This PR adds a check before we emit a brillig call opcode to see if the predicate is zero and, if so, we skip it and zero out all the outputs at compile time.

Additional Context

Documentation*

Check one:

• [x] No documentation needed.
• [ ] Documentation included in this PR.
• [ ] [For Experimental Features] Documentation to be submitted in a separate PR.

PR Checklist*

• [x] I have tested the changes locally.
• [x] I have formatted the changes with Prettier and/or cargo fmt on default settings.

[github-actions[bot]]

Changes to circuit sizes

Generated at commit: 2e2c4b383f57755a694918e9f7d1452f20d85124, compared to commit: 49e1b0c0d45565f3e87469b77f2fef0c283f6ea1

🧾 Summary (10% most significant diffs)

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
conditional_1	-5 ✅	-0.06%	0 ➖	0.00%
hashmap	-15 ✅	-0.01%	0 ➖	0.00%

---

Full diff report 👇

| Program | ACIR opcodes (+/-) | % | Circuit size (+/-) | % | |:-|-:|-:|-:|-:| | **conditional_1** | 8,545 (-5) | **-0.06%** | 38,809 (0) | **0.00%** | | **hashmap** | 209,896 (-15) | **-0.01%** | 397,324 (0) | **0.00%** |