[Mirror] server: fix memory reservations in populate_token_probs

[ngxson]

Mirror from upstream PR: ggml-org#18787

Note: @coderabbitai use my 'Mirror PR' preset for reviewing this.

Summary by CodeRabbit

• Bug Fixes
  • Fixed an issue where returned top probabilities could exceed the requested amount. The system now consistently respects both the requested number of probabilities and actual available candidates, ensuring accurate results.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

📝 Walkthrough

Walkthrough

This change introduces an explicit alias for requested top-n probabilities and ensures the returned number of probabilities is capped by both the requested amount and available candidates in both post_sampling and non-post_sampling code paths, with updated loops and reserves accordingly.

Changes

Cohort / File(s)	Summary
Probability capping refinement tools/server/server-context.cpp	Introduced n_probs_request alias for requested probabilities and implemented consistent capping logic across both post_sampling and non-post_sampling branches using min(max_probs, n_probs_request) to limit returned top-n probabilities

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Poem

🐰 With probabilities now capped just right,
The top-n values shine so bright!
No overflow, no excess here,
The balance perfect, crystal clear!

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and specifically describes the main change: fixing memory reservations in the populate_token_probs function in the server code.
Description check	✅ Passed	The description is minimal but sufficient for a mirror PR, providing the upstream source link and review instructions; however, it lacks detail about what the fix accomplishes.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

---

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