Avoid exceptions in UCI spin option validation

[luckysolanki902]

Summary

Malformed setoption input for a spin option is validated in Option::operator=, but using std::stoi for that check can throw std::invalid_argument or std::out_of_range. Because Stockfish is compiled with -fno-exceptions, malformed console input terminates the engine.

Examples on current master:

setoption name Hash value 999999999999999999  # -> std::out_of_range -> crash
setoption name Hash value notanumber          # -> std::invalid_argument -> crash

Confirmed live on stockfish-dev-20260415-b1fb50ae:

libc++abi: terminating due to uncaught exception of type std::out_of_range: stoi: out of range

Fix

Replace the throwing validation path in Option::operator= with std::strtol. This keeps validation non-throwing without depending on std::from_chars availability on older toolchains. Rejected values return early without mutating the option.

The new validation rejects:

• non-numeric values (no conversion)
• trailing garbage (partial conversion)
• overflow / underflow
• values outside [min, max]

Testing

Bench: 2984258

Malformed values are now rejected without terminating the engine.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 7513f6a5-e4f4-4e3a-9a23-1f27575fe09d

📥 Commits

Reviewing files that changed from the base of the PR and between 463afd4 and 74f97d0.

📒 Files selected for processing (1)

• src/ucioption.cpp

---

📝 Walkthrough

Walkthrough

The Option::operator=(const std::string& v) implementation in src/ucioption.cpp was changed to handle spin-type validation explicitly. Instead of calling std::stoi inside the conditional, the code now parses v with std::strtol, verifies full consumption, checks errno != ERANGE, and enforces the [min, max] range. On parse or range failure the assignment returns early without updating currentValue or calling on_change. The <cerrno> header was added.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[coderabbitai]

Actionable comments posted: 1

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: dd8e6d6f-3c47-4672-9afd-2e194f2e75bf

📥 Commits

Reviewing files that changed from the base of the PR and between ef75adc and 463afd4.

📒 Files selected for processing (1)

• src/ucioption.cpp

[github-actions]

clang-format 20 needs to be run on this PR.
If you do not have clang-format installed, the maintainer will run it when merging.
For the exact version please see https://packages.ubuntu.com/plucky/clang-format-20.

(execution 24610608033 / attempt 1)

[anematode]

This one doesn't need a fishtest run (it doesn't affect the engine's play)

[Disservin]

we avoided charconv because in the past workers on fishtest failed because it's missing

[vondele]

I recall this as well, but I do not longer recall which environment / compiler version was involved. For gcc, full support of charconv seems to be only available from 11+ and clang even a bit later. In CI our current set of supported compiler see to work fine, at least for the smoke test.

We're a bit on the same path as with the fens. In principle we rely on the gui to provide valid input.

[Disservin]

iirc it was clang 13 from sebs machine and he had a somewhat special system where it was missing..

[luckysolanki902]

Addressed the charconv concern by switching the validation path from std::from_chars to std::strtol in the latest force-push. This keeps the fix non-throwing in -fno-exceptions builds without depending on charconv support on older toolchains. Bench is still unchanged at 2984258.

[Disservin]

i'd also prefer if we simply clamp the value if it is outside the range