fix!: remove LengthPrefix deref fallback that could silently drop length encoding

[Copilot]

LengthPrefix could compile and run without writing a prefix when Length was not Encodable (e.g. u32), because method resolution fell through Deref to the wrapped value’s encode. This change removes that fallback so invalid LengthPrefix usage is rejected at compile time.

• API behavior change: remove Deref on LengthPrefix

  • Deleted impl Deref for LengthPrefix<...> to prevent .encode() from auto-forwarding to the inner encodable when LengthPrefix itself does not satisfy Encodable bounds.
  • This makes mis-specified length types fail loudly instead of producing truncated output.

• Regression guard: compile-fail coverage

  • Added trybuild-based UI test asserting LengthPrefix::<_, u32, _>::new("hello").encode(...) does not compile.
  • Added expected compiler output snapshot to lock the behavior.

• Test cleanup aligned with API change

  • Removed LengthPrefix unit tests that relied on deref semantics (*lp style assertions), preserving only behavior that remains part of the type’s API.

use core::num::TryFromIntError;
use encode::combinators::LengthPrefix;
use encode::Encodable;

fn main() {
    let mut buf = Vec::new();
    LengthPrefix::<_, u32, TryFromIntError>::new("hello")
        .encode(&mut buf) // now compile error: `u32` does not satisfy `Encodable`
        .unwrap();
}

• Fixes LengthPrefix silently drops the prefix when Length is not Encodable (Deref shadows the failed impl) #47

[Altair-Bueno]

@copilot is there any combinator that has this issue too? We should fix any other deref that leads to the same bug

[Copilot]

@copilot is there any combinator that has this issue too? We should fix any other deref that leads to the same bug

Yes—I audited all combinators with Deref and removed the ones that could silently bypass combinator behavior: BE, LE, Iter, and FromError. I also added compile-fail regressions for each misuse pattern. Fixed in 19b6a9d. Cond still has Deref, but it doesn’t have the same fallback hazard in practice because its constructor and Encodable bounds line up.

[Altair-Bueno]

Cond still has Deref, but it doesn’t have the same fallback hazard in practice because its constructor and Encodable bounds line up.

@copilot I would rather have it gone to avoid potential issues in the future. Are there any more Deref uses I missed?

[Copilot]

Cond still has Deref, but it doesn’t have the same fallback hazard in practice because its constructor and Encodable bounds line up.

@copilot I would rather have it gone to avoid potential issu...

Done in 9e88b34: I removed Deref from Cond and added a compile-fail regression for invalid Cond encode usage. I also re-audited the repo—there are no remaining combinator Deref impls; the only Deref left is IoEncoder in src/encoders/std_io.rs.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (0f73ad7) to head (6d0279c).

Additional details and impacted files

@@             Coverage Diff             @@
##           master       #48      +/-   ##
===========================================
+ Coverage   99.43%   100.00%   +0.56%     
===========================================
  Files          25        25              
  Lines        1067      1025      -42     
===========================================
- Hits         1061      1025      -36     
+ Misses          6         0       -6     

Files with missing lines	Coverage Δ
src/combinators/be.rs	100.00% <ø> (ø)
src/combinators/cond.rs	100.00% <ø> (ø)
src/combinators/from_error.rs	100.00% <100.00%> (+3.44%)	⬆️
src/combinators/iter.rs	100.00% <ø> (ø)
src/combinators/le.rs	100.00% <ø> (ø)
src/combinators/length_prefix.rs	100.00% <100.00%> (+2.88%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.