https://bugs.winehq.org/show_bug.cgi?id=57306
--- Comment #16 from Zeb Figura z.figura12@gmail.com --- Well, Rafał, I think if you read the surrounding code, it becomes clear that the "max_points" variable, returned by a function called "get_arc_max_points", is supposed to hold the maximum possible number of points returned from get_arc_points().
I think if you're familiar with code idioms, you might understand that this kind of pattern is not uncommon, and it's typically used when you want to avoid allocating more than once (an expensive action) and performing an expensive calculation more than once.
I think if you read the regression commit, you might notice that the code used this pattern already.
Is the pattern fragile? Probably. Is it to blame for the regression in this case? I would say so. Did I consider this possibility when writing it? Why, yes, I did, and I decided to leave the code as-is, because I'm not about to change more than I absolutely need to.
But yet another wildly uninformed, wildly incorrect diagnosis, that demonstrates little to no understanding of the code in question, given with a know-it-all attitude, really irks me. I must nevertheless apologize for letting it get to me.