ColorsPanel: for loop check condition simplify for optimize

[GermanAizek]

@natoscott, isn't it easier to check each step in the loops with this PR condition?
Access by index in array heavy operation vs check i index with const var LAST_COLORSCHEME

Clion IDE have diff view Code and Assembly with compiler flags.
Out of interest, I can show how much less x86 instructions it should give on -O3.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 182eeb15-0a78-4448-9bf1-cd06de4b8edf

📥 Commits

Reviewing files that changed from the base of the PR and between 1ef4226 and 673e236.

📒 Files selected for processing (1)

• ColorsPanel.c

---

📝 Walkthrough

Walkthrough

Two loops in ColorsPanel.c that iterate over color schemes were changed to use i < LAST_COLORSCHEME as the termination condition instead of checking ColorSchemeNames[i] != NULL. This affects the loop clearing CheckItem states in the event handler and the loop adding CheckItem entries during panel construction.

Changes

File	Change Summary
ColorsPanel.c	Replaced NULL-sentinel loop condition with LAST_COLORSCHEME bound in two loops (event handler and panel constructor)

Sequence Diagram(s)

Not applicable — this change is a simple loop-bound substitution with no new control flow or interactions to diagram.

Estimated code review effort: Medium

Related issues: None provided

Related PRs: None provided

Suggested labels: bug, cleanup

Suggested reviewers: None provided

Poem

A rabbit hopped through scheme and name,
Chasing NULLs, it wasn't the same.
Now bounded tight by LAST_COLORSCHEME,
No more scanning past the seam.
Hop, hop — the loop runs clean. 🐇

---

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.}

[fasterit]

@GermanAizek: Please show how much runtime / code that change saves.

[GermanAizek]

@GermanAizek: Please show how much runtime / code that change saves.

I can show assembler listing on -O3 using x86-64 instructions, I can calculate how many processor cycles are spent on this cycle, visually it will look like in this PR #2055

[GermanAizek]

I checked the assembler listing on latest versions GCC 16.1.0 and Clang 21.1.8, and for now it's better to leave it the same, since on -O3 it unroll loop itself, but my version makes CPU branch predictor work with jne instruction, which is worse for ALU and SIMD blocks.

[GermanAizek]

@fasterit indeed, checking for NULL makes the loop more vectorized and the compiler can optimize better for AVX instructions, I'll check other interesting loops for a similar case.

But usually basic simplest loop is best optimized by C/C++ compilers when the int is an index and compared to another number (for indexed loop)