Re: gdi32: Correct order of matrix elements in World transform

[Konstantin Kondratyuk]

Oops, this patch is not full.

I think, there is code with bad index also:

/* Extra transformation specified by caller */
    if (lpmat && !is_identity_MAT2(lpmat))
    {
        FT_Matrix extraMat;
        extraMat.xx = FT_FixedFromFIXED(lpmat->eM11);
        extraMat.xy = FT_FixedFromFIXED(lpmat->eM21);   // <<<<<<<
        extraMat.yx = FT_FixedFromFIXED(lpmat->eM12);   // <<<<<<<
        extraMat.yy = FT_FixedFromFIXED(lpmat->eM22);
        pFT_Matrix_Multiply(&extraMat, &transMat);
        pFT_Matrix_Multiply(&extraMat, &transMatUnrotated);
        needsTransform = TRUE;
    }

What can I do? Send second fix? Or send "full" patch for both transformations?

gdi32: Correct order of matrix elements in World transform

[Dmitry Timoshkov]

"Konstantin Kondratyuk"  wrote:
> Oops, this patch is not full.
>
> I think, there is code with bad index also:
>
> /* Extra transformation specified by caller */
>    if (lpmat && !is_identity_MAT2(lpmat))
>    {
>        FT_Matrix extraMat;
>        extraMat.xx = FT_FixedFromFIXED(lpmat->eM11);
>        extraMat.xy = FT_FixedFromFIXED(lpmat->eM21);   // <<<<<<<
>        extraMat.yx = FT_FixedFromFIXED(lpmat->eM12);   // <<<<<<<
>        extraMat.yy = FT_FixedFromFIXED(lpmat->eM22);
>        pFT_Matrix_Multiply(&extraMat, &transMat);
>        pFT_Matrix_Multiply(&extraMat, &transMatUnrotated);
>        needsTransform = TRUE;
>    }
>
> What can I do? Send second fix? Or send "full" patch for both transformations?

Have you tested your patch? In my tests gdi and freetype differ in matrix usage.