The old code is comparing pixman_fixed_48_16_t values to pixman_fixed_16_16_t values, thus it is checking for truncation of overflow values.
It would probably be better to clamp these overflowed values, like pixman_transform_point_31_16 is doing to clamp to the pixman_fixed_48_16 result. Right now the result is an odd mix of clamping and modulus. A rewrite to go directly to clamped pixman_fixed_16_16 values would be even better. On Tue, Apr 26, 2016 at 10:59 AM, Petr Kobalíček <kobalicek.p...@gmail.com> wrote: > Doesn't this check for NaNs? > > On Sun, Apr 24, 2016 at 8:22 PM, Emil Velikov <emil.l.veli...@gmail.com> > wrote: > >> With commit ed39992564b "Use pixman_transform_point_31_16() from >> pixman_transform_point()" we added some strange hunks. >> >> Namely: we copy the data from the internal storage to the user vector >> only to compare them immediately after. >> >> Cc: Siarhei Siamashka <siarhei.siamas...@gmail.com> >> --- >> >> Siarhei, what is the intent with the original commit ? Any ideas why >> things crash ? >> >> Seemingly this can be dropped/replaced with TRUE, yet it causes one of >> the tests () to segfault in the optimised SSE2 codepath - >> scaled_bilinear_scanline_sse2_8888_8888_SRC. >> >> BILINEAR_INTERPOLATE_ONE_PIXEL -> BILINEAR_INTERPOLATE_ONE_PIXEL_HELPER >> >> Regards, >> Emil >> --- >> >> pixman/pixman-matrix.c | 8 ++------ >> 1 file changed, 2 insertions(+), 6 deletions(-) >> >> diff --git a/pixman/pixman-matrix.c b/pixman/pixman-matrix.c >> index 65b3d32..117015b 100644 >> --- a/pixman/pixman-matrix.c >> +++ b/pixman/pixman-matrix.c >> @@ -393,9 +393,7 @@ pixman_transform_point_3d (const struct >> pixman_transform *transform, >> vector->vector[1] = tmp.v[1]; >> vector->vector[2] = tmp.v[2]; >> >> - return vector->vector[0] == tmp.v[0] && >> - vector->vector[1] == tmp.v[1] && >> - vector->vector[2] == tmp.v[2]; >> + return TRUE; >> } >> >> PIXMAN_EXPORT pixman_bool_t >> @@ -414,9 +412,7 @@ pixman_transform_point (const struct pixman_transform >> *transform, >> vector->vector[1] = tmp.v[1]; >> vector->vector[2] = tmp.v[2]; >> >> - return vector->vector[0] == tmp.v[0] && >> - vector->vector[1] == tmp.v[1] && >> - vector->vector[2] == tmp.v[2]; >> + return TRUE; >> } >> >> PIXMAN_EXPORT pixman_bool_t >> -- >> 2.8.0 >> >> _______________________________________________ >> Pixman mailing list >> Pixman@lists.freedesktop.org >> https://lists.freedesktop.org/mailman/listinfo/pixman >> > > > _______________________________________________ > Pixman mailing list > Pixman@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/pixman > >
_______________________________________________ Pixman mailing list Pixman@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/pixman
