bug(#3999): change @star resolving implementation

[maxonfjvipon]

Closes: #3999

[github-actions]

🚀 Performance Analysis

Test	Base Score	PR Score	Change	% Change	Unit	Mode
benchmarks.XmirBench.xmirToEO	162.711	148.522	-14.189	-8.72%	ms/op	Average Time
benchmarks.XmirBench.xmirToPhi	142.444	189.734	47.289	33.20%	ms/op	Average Time
benchmarks.XmirBench.xmirToSaltyPhi	157.049	151.896	-5.153	-3.28%	ms/op	Average Time

✅ Performance gain: benchmarks.XmirBench.xmirToEO is faster by 14.189 ms/op (8.72%)
⚠️ Performance loss: benchmarks.XmirBench.xmirToPhi is slower by 47.289 ms/op (33.20%)
✅ Performance gain: benchmarks.XmirBench.xmirToSaltyPhi is faster by 5.153 ms/op (3.28%)

[maxonfjvipon]

@yegor256 please check

[maxonfjvipon]

@yegor256 fixed, please check one more time

[0crat]

@maxonfjvipon Thanks for your contribution! Here's the breakdown of your +4 points: +4 base, +16 for 428 hits-of-code, -8 for exceeding 200 HoC, -4 for only 4 comments, and -4 to cap at 4 points. While the HoC count is impressive, aim for more concise contributions and encourage more review comments next time. Keep up the good work! Your balance is now +8.

[0crat]

@yegor256 Thanks for your code review! Here's how your +4 points were calculated: +4 base points, -24 for self-review, +1 for hits-of-code (capped at 16), -10 for <6 comments, +1 for 4 comments (capped at 8), adjusted to minimum 4 points. Keep improving! Your balance is now -596. 🚀