gh-146587: fix type slot assignment incase of multiple slots for same name

[kumaraditya303]

• Issue: Sequence-like subclass __radd__ fallback regression #146587

[sergey-miryanov]

@kumaraditya303 Could you please take a look at sergey-miryanov@f26545a? WDYT?

[kumaraditya303]

Could you please take a look at sergey-miryanov@f26545a? WDYT?

It looks the same as this one except more complicated, you are also checking that if a slot has duplicates then only one of them should be filled isn't it?

I am not sure that the loop is big enough that there would be performance difference.

[sergey-miryanov]

Yeah, it looks the same. Except we don’t need to iterate all slotdefs, because we precalculates dup indexes.

[sergey-miryanov]

I wrote answer before your edit.

I don’t remember exactly how items in slotdefs, about 70 maybe 94 (edited). For dups we need iterate 2 or 3 items.

[kumaraditya303]

Yeah, it looks the same. Except we don’t need to iterate all slotdefs, because we precalculates dup indexes.

Alright, I implemented your suggestion.

[sergey-miryanov]

Thanks!