cairo-5c ftbfs in quantal

[Matthias Klose]

cairo ftbfs in quantal, failing some test cases. quantal and sid use the same cairo version, but the tests fail in quantal and succeed in sid.

set-font failed: pixel mismatch 0xe0000000 not similar to 0x0 at 22 10
set-font-size failed: pixel mismatch 0xf8000000 not similar to 0x0 at 56 12
set-font-matrix failed: pixel mismatch 0xf8000000 not similar to 0x0 at 6 22
show-text failed: pixel mismatch 0xe0000000 not similar to 0x0 at 27 10
text-path failed: pixel mismatch 0x5b000000 not similar to 0x44000000 at 53 16
text-extents failed: pixel mismatch 0xff1f1f1f not similar to 0xffffffff at 27 30
font-extents failed: font_extents returned wrong value: {ascent = 11, descent = 3, height = 14, max_x_advance = 16, max_y_advance = 0} != {ascent = 12, descent = 3, height = 14, max_x_advance = 16, max_y_advance = 0}

all the test regressions are font related; sid has freetype 2.4.9, quantal 2.4.10, so either the test cases need adjustment, or this could be a regression in freetype.

[Steve Langasek (vorlon)]

This same build failure was reported in Debian unstable, before freetype was revved. So this isn't a freetype regression. It still could be a freetype bug at its root, but it looks more likely to be a toolchain regression.