Re: [PATCH v2 0/2] MR4475: gdiplus: Avoid computation overflow and underflow by using hypotf

[Esme Povirk (＠madewokherd)]

Esme Povirk (@madewokherd) commented about dlls/gdiplus/graphicspath.c:
> -
>        float dy0 = distance * (p2->Y - p1->Y) / length_0;
> -
>        float dx1 = distance * (p3->X - p2->X) / length_1;
> -
>        float dy1 = distance * (p3->Y - p2->Y) / length_1;
> -
>        float det = (dy0*dx1 - dx0*dy1);
> -
>        float dx = (dx0*dx1*(dx0-dx1) + dy0*dy0*dx1 - dy1*dy1*dx0)/det;
> -
>        float dy = (dy0*dy1*(dy0-dy1) + dx0*dx0*dy1 - dx1*dx1*dy0)/det;
> -
>        REAL distance = pen_width / 2.0;
> -
>        REAL length_0 = hypotf((p2->X - p1->X), (p2->Y - p1->Y));
> -
>        REAL length_1 = hypotf((p3->X - p2->X), (p3->Y - p2->Y));
> -
>        REAL dx0 = distance * (p2->X - p1->X) / length_0;
> -
>        REAL dy0 = distance * (p2->Y - p1->Y) / length_0;
> -
>        REAL dx1 = distance * (p3->X - p2->X) / length_1;
> -
>        REAL dy1 = distance * (p3->Y - p2->Y) / length_1;
> -
>        REAL det = (dy0 * dx1 - dx0 * dy1);
> -
>        REAL dx = (dx0 * dx1 * (dx0 - dx1) + dy0 * dy0 * dx1 - dy1 * dy1 * dx0) / det;
> -
>        REAL dy = (dy0 * dy1 * (dy0 - dy1) + dx0 * dx0 * dy1 - dx1 * dx1 * dy0) / det;

There are some unrelated changes in here.