Hi, I just installed cairo-1.0.0 and some unexpected failures seem to have occurred during make check: make[2]: Entering directory `/var/tmp/portage/cairo-1.0.0/work/cairo-1.0.0/test' a8-mask is expected to fail: image backend fails because libpixman only handles (stride % sizeof(pixman_bits) == 0) a8-mask-image: XFAIL a8-mask-xlib: UNTESTED XFAIL: a8-mask caps-sub-paths-image: PASS caps-sub-paths-xlib: UNTESTED PASS: caps-sub-paths clip-nesting-image: PASS clip-nesting-xlib: UNTESTED PASS: clip-nesting clip-operator-image: PASS clip-operator-xlib: UNTESTED PASS: clip-operator clip-twice-image: PASS clip-twice-xlib: UNTESTED PASS: clip-twice composite-integer-translate-source-image: PASS composite-integer-translate-source-xlib: UNTESTED PASS: composite-integer-translate-source composite-integer-translate-over-image: PASS composite-integer-translate-over-xlib: UNTESTED PASS: composite-integer-translate-over composite-integer-translate-over-repeat-image: PASS composite-integer-translate-over-repeat-xlib: UNTESTED PASS: composite-integer-translate-over-repeat create-from-png-image: PASS create-from-png-xlib: UNTESTED PASS: create-from-png create-from-png-image: PASS create-from-png-xlib: UNTESTED PASS: create-from-png-stream dash-offset-negative-image: PASS dash-offset-negative-xlib: UNTESTED PASS: dash-offset-negative fill-and-stroke-image: PASS fill-and-stroke-xlib: UNTESTED PASS: fill-and-stroke fill-rule-image: PASS fill-rule-xlib: UNTESTED PASS: fill-rule filter-nearest-offset is expected to fail: wrong sampling location for nearest-neighbor filter in libpixman and Render filter-nearest-offset-image: XFAIL filter-nearest-offset-xlib: UNTESTED XFAIL: filter-nearest-offset get-and-set-image: FAIL get-and-set-xlib: UNTESTED FAIL: get-and-set gradient-alpha-image: PASS gradient-alpha-xlib: UNTESTED PASS: gradient-alpha leaky-polygon-image: PASS leaky-polygon-xlib: UNTESTED PASS: leaky-polygon line-width-image: PASS line-width-xlib: UNTESTED PASS: line-width linear-gradient-image: PASS linear-gradient-xlib: UNTESTED PASS: linear-gradient mask-image: PASS mask-xlib: UNTESTED PASS: mask mask-ctm-image: FAIL mask-ctm-xlib: UNTESTED FAIL: mask-ctm mask-surface-ctm-image: FAIL mask-surface-ctm-xlib: UNTESTED FAIL: mask-surface-ctm move-to-show-surface-image: PASS move-to-show-surface-xlib: UNTESTED PASS: move-to-show-surface nil-surface-image: PASS nil-surface-xlib: UNTESTED PASS: nil-surface operator-clear-image: PASS operator-clear-xlib: UNTESTED PASS: operator-clear operator-source-image: PASS operator-source-xlib: UNTESTED PASS: operator-source paint-image: PASS paint-xlib: UNTESTED PASS: paint paint-with-alpha-image: FAIL paint-with-alpha-xlib: UNTESTED FAIL: paint-with-alpha path-data-image: PASS path-data-xlib: UNTESTED PASS: path-data pixman-rotate is expected to fail: known off-by-one bug when rotating a pixman image pixman-rotate-image: XFAIL pixman-rotate-xlib: UNTESTED XFAIL: pixman-rotate rectangle-rounding-error-image: PASS rectangle-rounding-error-xlib: UNTESTED PASS: rectangle-rounding-error scale-source-surface-paint-image: FAIL scale-source-surface-paint-xlib: UNTESTED FAIL: scale-source-surface-paint select-font-no-show-text-image: PASS select-font-no-show-text-xlib: UNTESTED PASS: select-font-no-show-text self-copy-image: PASS self-copy-xlib: UNTESTED PASS: self-copy self-intersecting is expected to fail: Self-intersecting strokes are wrong due to incremental trapezoidization. self-intersecting-image: XFAIL self-intersecting-xlib: UNTESTED XFAIL: self-intersecting set-source-image: PASS set-source-xlib: UNTESTED PASS: set-source show-text-current-point-image: PASS show-text-current-point-xlib: UNTESTED PASS: show-text-current-point source-clip-image: PASS source-clip-xlib: UNTESTED PASS: source-clip source-surface-scale-paint-image: FAIL source-surface-scale-paint-xlib: UNTESTED FAIL: source-surface-scale-paint surface-finish-twice-image: PASS surface-finish-twice-xlib: UNTESTED PASS: surface-finish-twice surface-pattern-image: PASS surface-pattern-xlib: UNTESTED PASS: surface-pattern text-antialias-gray-image: PASS text-antialias-gray-xlib: UNTESTED PASS: text-antialias-gray text-antialias-none-image: PASS text-antialias-none-xlib: UNTESTED PASS: text-antialias-none text-antialias-subpixel-image: PASS text-antialias-subpixel-xlib: UNTESTED PASS: text-antialias-subpixel text-cache-crash-image: PASS text-cache-crash-xlib: UNTESTED PASS: text-cache-crash text-pattern-image: PASS text-pattern-xlib: UNTESTED PASS: text-pattern text-rotate is expected to fail: minor bugs in positioning rotated glyphs text-rotate-image: XFAIL text-rotate-xlib: UNTESTED XFAIL: text-rotate transforms-image: PASS transforms-xlib: UNTESTED PASS: transforms translate-show-surface-image: PASS translate-show-surface-xlib: UNTESTED PASS: translate-show-surface trap-clip-image: PASS trap-clip-xlib: UNTESTED PASS: trap-clip unantialiased-shapes-image: PASS unantialiased-shapes-xlib: UNTESTED PASS: unantialiased-shapes unbounded-operator-image: PASS unbounded-operator-xlib: UNTESTED PASS: unbounded-operator PASS: user-data rel-path-image: PASS rel-path-xlib: UNTESTED PASS: rel-path ft-font-create-for-ft-face-image: PASS ft-font-create-for-ft-face-xlib: UNTESTED PASS: ft-font-create-for-ft-face PASS: xlib-surface ======================================================================== 6 of 56 tests failed Please report to http://bugs.freedesktop.org/enter_bug.cgi?product=cairo I'm on Gentoo/amd64, happy to provide you with whatever info you need.
(In reply to comment #0) > get-and-set-image: FAIL That one is quite surprising. Could you step through that in a debugger, or insert print statements into the test and see where it's failing? > mask-ctm-image: FAIL > mask-surface-ctm-image: FAIL > paint-with-alpha-image: FAIL > scale-source-surface-paint-image: FAIL > source-surface-scale-paint-image: FAIL For each of these, it might be useful to see the resulting -image-out.png images or at least a description of what the failures look like. > I'm on Gentoo/amd64, happy to provide you with whatever info you need. I'll try to get access to an amd64 machine for testing. Thanks for the report.
yep, it's now 26 out 56. 20 more failures!! :( So in between yesterday and today, I downgraded from xorg-6.9.0 rc1 back to 6.8.2 and upgraded some gnome related libs (libsoup and co.). I'll be attaching all the png files as I can and I'll try to do some debugging for you.
Created attachment 3048 [details] caps-sub-path-xlib-failed caps-sub-path-xlib test failure -- this is the .png
Created attachment 3049 [details] clip-nesting-xlib-out.png
Created attachment 3050 [details] clip-operator-xlib-out.png
Created attachment 3051 [details] clip-twice-xlib-out.png
OK, hang on. instead of all these png's I'll just update the tarball of the test/ directory so you can see for yourself. If it helps, this is the make check output: a8-mask is expected to fail: image backend fails because libpixman only handles (stride % sizeof(pixman_bits) == 0) a8-mask-image: XFAIL a8-mask-xlib: XFAIL XFAIL: a8-mask caps-sub-paths-image: PASS caps-sub-paths-xlib: FAIL FAIL: caps-sub-paths clip-nesting-image: PASS clip-nesting-xlib: FAIL FAIL: clip-nesting clip-operator-image: PASS clip-operator-xlib: FAIL FAIL: clip-operator clip-twice-image: PASS clip-twice-xlib: FAIL FAIL: clip-twice composite-integer-translate-source-image: PASS composite-integer-translate-source-xlib: PASS PASS: composite-integer-translate-source composite-integer-translate-over-image: PASS composite-integer-translate-over-xlib: PASS PASS: composite-integer-translate-over composite-integer-translate-over-repeat-image: PASS composite-integer-translate-over-repeat-xlib: PASS PASS: composite-integer-translate-over-repeat create-from-png-image: PASS create-from-png-xlib: PASS PASS: create-from-png create-from-png-image: PASS create-from-png-xlib: PASS PASS: create-from-png-stream dash-offset-negative-image: PASS dash-offset-negative-xlib: FAIL FAIL: dash-offset-negative fill-and-stroke-image: PASS fill-and-stroke-xlib: FAIL FAIL: fill-and-stroke fill-rule-image: PASS fill-rule-xlib: FAIL FAIL: fill-rule filter-nearest-offset is expected to fail: wrong sampling location for nearest-neighbor filter in libpixman and Render filter-nearest-offset-image: XFAIL filter-nearest-offset-xlib: XFAIL XFAIL: filter-nearest-offset get-and-set-image: FAIL get-and-set-xlib: FAIL FAIL: get-and-set gradient-alpha-image: PASS gradient-alpha-xlib: PASS PASS: gradient-alpha leaky-polygon-image: PASS leaky-polygon-xlib: FAIL FAIL: leaky-polygon line-width-image: PASS line-width-xlib: FAIL FAIL: line-width linear-gradient-image: PASS linear-gradient-xlib: PASS PASS: linear-gradient mask-image: PASS mask-xlib: FAIL FAIL: mask mask-ctm-image: FAIL mask-ctm-xlib: FAIL FAIL: mask-ctm mask-surface-ctm-image: FAIL mask-surface-ctm-xlib: FAIL FAIL: mask-surface-ctm move-to-show-surface-image: PASS move-to-show-surface-xlib: PASS PASS: move-to-show-surface nil-surface-image: PASS nil-surface-xlib: PASS PASS: nil-surface operator-clear-image: PASS operator-clear-xlib: PASS PASS: operator-clear operator-source-image: PASS operator-source-xlib: FAIL FAIL: operator-source paint-image: PASS paint-xlib: PASS PASS: paint paint-with-alpha-image: FAIL paint-with-alpha-xlib: FAIL FAIL: paint-with-alpha path-data-image: PASS path-data-xlib: FAIL FAIL: path-data pixman-rotate is expected to fail: known off-by-one bug when rotating a pixman image pixman-rotate-image: XFAIL pixman-rotate-xlib: XFAIL XFAIL: pixman-rotate rectangle-rounding-error-image: PASS rectangle-rounding-error-xlib: PASS PASS: rectangle-rounding-error scale-source-surface-paint-image: FAIL scale-source-surface-paint-xlib: FAIL FAIL: scale-source-surface-paint select-font-no-show-text-image: PASS select-font-no-show-text-xlib: PASS PASS: select-font-no-show-text self-copy-image: PASS self-copy-xlib: FAIL FAIL: self-copy self-intersecting is expected to fail: Self-intersecting strokes are wrong due to incremental trapezoidization. self-intersecting-image: XFAIL self-intersecting-xlib: XFAIL XFAIL: self-intersecting set-source-image: PASS set-source-xlib: PASS PASS: set-source show-text-current-point-image: PASS show-text-current-point-xlib: FAIL FAIL: show-text-current-point source-clip-image: PASS source-clip-xlib: PASS PASS: source-clip source-surface-scale-paint-image: FAIL source-surface-scale-paint-xlib: FAIL FAIL: source-surface-scale-paint surface-finish-twice-image: PASS surface-finish-twice-xlib: PASS PASS: surface-finish-twice surface-pattern-image: PASS surface-pattern-xlib: PASS PASS: surface-pattern text-antialias-gray-image: PASS text-antialias-gray-xlib: FAIL FAIL: text-antialias-gray text-antialias-none-image: PASS text-antialias-none-xlib: PASS PASS: text-antialias-none text-antialias-subpixel-image: PASS text-antialias-subpixel-xlib: FAIL FAIL: text-antialias-subpixel text-cache-crash-image: PASS text-cache-crash-xlib: PASS PASS: text-cache-crash text-pattern-image: PASS text-pattern-xlib: PASS PASS: text-pattern text-rotate is expected to fail: minor bugs in positioning rotated glyphs text-rotate-image: XFAIL text-rotate-xlib: XFAIL XFAIL: text-rotate transforms-image: PASS transforms-xlib: FAIL FAIL: transforms translate-show-surface-image: PASS translate-show-surface-xlib: PASS PASS: translate-show-surface trap-clip-image: PASS trap-clip-xlib: FAIL FAIL: trap-clip unantialiased-shapes-image: PASS unantialiased-shapes-xlib: FAIL FAIL: unantialiased-shapes unbounded-operator-image: PASS unbounded-operator-xlib: PASS PASS: unbounded-operator PASS: user-data rel-path-image: PASS rel-path-xlib: FAIL FAIL: rel-path ft-font-create-for-ft-face-image: PASS ft-font-create-for-ft-face-xlib: PASS PASS: ft-font-create-for-ft-face PASS: xlib-surface ======================================================================== 26 of 56 tests failed
Here it is: http://dev.gentoo.org/~seemant/test.tar.bz2
sorry, it is 6 -- it jumps to 26 if I run the tests in a Xvfb session instead. Wrong number there.
Created attachment 3052 [details] mask-ctm-image-out.png
Created attachment 3053 [details] mask-surface-ctm-image-out.png
Created attachment 3054 [details] paint-with-alpha-image-out.png
Created attachment 3055 [details] scale-source-surface-paint-image-out.png
Created attachment 3056 [details] source-surface-scale-paint-image-out.png
does this help: get-and-set-image: About to: settings_set (cr, &settings[0]); About to: cairo_save (cr); About to: settings_set (cr, &settings[1]); About to: settings_get (cr, &check); settings_equal (&settings[1], &check) was true About to: cairo_restore (cr); About to: settings_get (cr, &check); settings_equal (&settings[0], &check) was FALSE FAIL get-and-set-xlib: UNTESTED FAIL: get-and-set ?
I added some debugging to settings_equal() : the operators are not equal the tolerances are not equal the fill_rules are not equal the line_widths are not equal the line_joins are not equal the miter_limits are not equal the xx are not equal the xy are not equal the x0 are not equal the yx are not equal the yy are not equal the y0 are not equal none of the things are equal in a and b in that function, as it turns out.
I've got access to an amd64 machine now which helps a lot. 1 down, and 5 to go: 2005-08-31 Carl Worth <cworth@cworth.org> * test/get-and-set.c: Fix test to not try to set a non-invertible matrix. (We'd gotten away with this due to round-off error on many systems). Fixes one out of six failures in bug #4245.
4 more down and 1 to go. Quite happily, all 5 bugs I've found so far were merely bugs in the tests themselves. So, so far we haven't identified any ways in which cairo 1.0.0 itself behaves poorly on amd64. 2005-09-01 Carl Worth <cworth@cworth.org> * test/cairo-test.h: Add includes to get sized-integers types such as uint32_t. * test/filter-nearest-offset.c: (draw): * test/mask-ctm.c: (draw): * test/mask-surface-ctm.c: (draw): * test/move-to-show-surface.c: (draw): * test/paint-with-alpha.c: (draw): * test/scale-source-surface-paint.c: (draw): * test/set-source.c: (draw): * test/source-surface-scale-paint.c: (draw): * test/translate-show-surface.c: (draw): Fix declaration of image data array to be uint32_t rather than unsigned long. Fixes four out of the remaining five failures in bug #4245.
Oh, my mistake. The commit I just made did fix all 5 tests that failed in the original bug report. I miscounted since I still had one remaining test failure on my amd64 machine. But it was not in the original report, (it is text-antialias-subpixel, and it looks like it's getting hinting when it's not supposed to---could just be freetype version or font availablity). Anyway, I think this bug is now fixed.
thanks much, Carl.
You're quite welcome. It's nice to be able to "check the box" for another platform. And thanks for your help here.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.