Add assert.approx_equal for epsilon-based float assertions

[ATX24]

Pull Request Template

Issue Reference

• This PR fixes/closes #[issue number]

Changes

Adds assert.approx_equal(actual, expected, epsilon) to the BAML test stdlib for epsilon-based float comparisons.

Testing

• Unit tests added/updated
• Manual testing performed
• Tested in Cursor Cloud Linux environment

Screenshots

Not applicable.

PR Checklist

• I have read and followed the contributing guidelines
• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings

Additional Notes

The error

Before this change, the BAML test stdlib did not define assert.approx_equal, so this smallest repro failed:

test "approx_equal_missing" {
    assert.approx_equal(float.parse("0.1") + float.parse("0.2"), float.parse("0.3"), 0.000000000000001)
}

Command run against origin/canary:

cargo run -p baml_cli -- test --from /tmp/baml-approx-repro-src

Output:

warning: using the internal BAML toolchain binary directly is not recommended. Use `baml` instead.
     Loading /tmp/baml-approx-repro-src/main.baml
    Checking 1 file(s)
error: unresolved name: approx_equal
   ╭─[ main.baml:1:1 ]
   │
 1 │ test "approx_equal_missing" {
   │ │ 
   │ ╰─ unresolved name: approx_equal
   │ 
   │ Note: Error code: E0003
───╯

error: unresolved name: approx_equal
   ╭─[ main.baml:1:30 ]
   │
 1 │ ╭─▶ test "approx_equal_missing" {
 2 │ ├─▶     assert.approx_equal(float.parse("0.1") + float.parse("0.2"), float.parse("0.3"), 0.000000000000001)
   │ │                                                                                                             
   │ ╰───────────────────────────────────────────────────────────────────────────────────────────────────────────── unresolved name: approx_equal
   │     
   │     Note: Error code: E0003
───╯

Root cause

baml_language/crates/baml_builtins2/baml_std/assert/assert.baml defined the test stdlib assertions is_true, not_null, equal, and contains, but had no epsilon-based float comparison primitive. assert.equal(actual, expected) uses exact equality, so float tests for non-exact IEEE 754 results had to inline (actual - expected).abs() <= epsilon manually.

The fix

Added assert.approx_equal(actual: float, expected: float, epsilon: float) -> null in baml_language/crates/baml_builtins2/baml_std/assert/assert.baml. It computes difference = actual - expected, then delta = difference.abs(), and panics unless delta <= epsilon. The panic message includes actual, expected, epsilon, and the observed delta.

I also added BAML float tests for:

• a non-exact runtime float expression: float.parse("0.1") + float.parse("0.2") vs float.parse("0.3")
• the inclusive boundary where delta == epsilon

Snapshots were updated for the assert stdlib compiler phases, the BAML source bytecode, and the CLI package listing.

Verification

Commands run after the fix:

PATH="$HOME/.local/bin:$PATH" cargo test -p baml_tests --test baml_src baml_test -- --nocapture

Relevant passing output showing the same repro now succeeds in baml_src/ns_floats/floats.baml:

PASS ::float_approx_equal_accepts_non_exact_result
PASS ::float_approx_equal_accepts_delta_equal_to_epsilon
Finished 1616 passed, 0 failed, 1616 total in 42s
test baml_test ... ok
test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 1 filtered out; finished in 102.63s

PATH="$HOME/.local/bin:$PATH" cargo test -p baml_tests __assert_std__
PATH="$HOME/.local/bin:$PATH" cargo test -p baml_tests --test baml_src bytecode
PATH="$HOME/.local/bin:$PATH" cargo test -p baml_cli render_assert_package_listing

Passing output:

test result: ok. 8 passed; 0 failed; 0 ignored; 0 measured; 1570 filtered out; finished in 1.14s

test bytecode ... ok
test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 1 filtered out; finished in 30.45s

test describe_command_tests::render_assert_package_listing ... ok
test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 162 filtered out; finished in 0.05s

  

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
beps	Ready	Preview, Comment	Jun 3, 2026 7:41pm
promptfiddle	Ready	Preview, Comment	Jun 3, 2026 7:41pm
promptfiddle2	Ready	Preview, Comment	Jun 3, 2026 7:41pm

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 7d6ce48d-a439-44e0-80c7-17f167a6c1bf

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch cursor/baml-assert-approx-equal-bf50

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

⏭️ Performance benchmarks were skipped

Perf benchmarks (CodSpeed) are opt-in on pull requests — they no longer run on every push. They always run automatically after merge to canary/main.

To run them on this PR, do any of the following, then push a commit (or re-run CI):

• Add RUN_CODSPEED=1 to the PR description, or
• Include run-perf or /perf in the PR title or any commit message.

[github-actions]

Binary size checks passed

✅ 7 passed

	Artifact	Platform	File	Gzip	Gated on	Baseline	Delta	Status
✅	baml-cli	Linux	🔒 17.6 MB	7.5 MB	file	24.1 MB	-6.4 MB (-26.8%)	OK
✅	packed-program	Linux	🔒 12.8 MB	5.4 MB	file	15.1 MB	-2.3 MB (-15.0%)	OK
✅	baml-cli	macOS	🔒 13.5 MB	6.5 MB	file	18.2 MB	-4.7 MB (-26.0%)	OK
✅	packed-program	macOS	🔒 9.8 MB	4.7 MB	file	11.5 MB	-1.7 MB (-14.9%)	OK
✅	baml-cli	Windows	🔒 14.5 MB	6.7 MB	file	19.6 MB	-5.1 MB (-26.2%)	OK
✅	packed-program	Windows	🔒 10.3 MB	4.8 MB	file	12.2 MB	-1.9 MB (-15.5%)	OK
✅	bridge_wasm	WASM	11.8 MB	🔒 3.3 MB	gzip	3.9 MB	-524.7 KB (-13.6%)	OK

🔒 = the size this artifact is GATED on (ceiling + delta). Binaries gate on file size (installed binary); WASM gates on gzip (download size). The other size is shown for information only.

---

Generated by cargo size-gate · workflow run