fix width calculation for draw.aacircle()

[Fantastair]

For more details, please refer to the discussion in #3682 .

[coderabbitai]

📝 Walkthrough

Walkthrough

Modified circle drawing calls in src_c/draw.c to use adjusted thickness parameters by passing width - 1 instead of width, with added comments clarifying that internal drawing logic applies thickness + 1 for final rendered width.

Changes

Cohort / File(s)	Summary
Circle Rendering Thickness Adjustment src_c/draw.c	Adjusted all aacircle-related function calls to use width - 1 as the thickness parameter. Added explanatory comments noting that the internal draw_circle_xiaolinwu logic handles final thickness as thickness + 1, clarifying the relationship between parameter and actual drawn width.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% 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 accurately summarizes the main change: fixing width calculation for the aacircle drawing function, which directly corresponds to the adjusted circle drawing calls in the changeset.
Description check	✅ Passed	The description references issue #3682 which provides context for the width calculation fix in draw.aacircle(), relating directly to the changeset.

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

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

Important

Action Needed: IP Allowlist Update

If your organization protects your Git platform with IP whitelisting, please add the new CodeRabbit IP address to your allowlist:

• ✨ 136.113.208.247/32 (new)
• 34.170.211.100/32
• 35.222.179.152/32

Failure to add the new IP will result in interrupted reviews.

---

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.}

[Fantastair]

As for why I ultimately chose to modify the parameter passing on the calling side rather than the underlying implementation, it's because draw_circle_xiaolinwu() also handles drawing solid circles. Modifying the width here would cause the center pixels of solid circles to be left empty. Therefore, passing width - 1 when drawing the circle border in draw.aacircle() represents the smallest possible change.

Additionally, I added an explanatory comment inside the ‘draw_circle_xiaolinwu()’ function to serve as a reminder for future developers.

[ankith26]

I wonder if it would be better to just fix the logic here instead, maybe we should do a radius - thickness + 1 initialization?

EDIT: Just saw your comment about centre of solid circles, but I'm still not convinced. This function should ideally do the "right thing" and the centre special case could be handled separately? I need to do a bit more testing though to understand things in more detail.

[ankith26]

Okay I played around a bit with this, and now I have a clearer understanding.

Basically the main issue is that draw.aacircle always draws an extra pixel at the centre, so a 1 radius circle needs a 3x3 box, a 2 radius circle needs a 5x5 box and so on (yes, just like #1487). So there's always going to be a jump somewhere and this PR is just shifting that jump from 1 -> 3 to (r - 2) -> r. I personally think this is a reasonable compromise given the status quo.