This bug is to track patches for exporting and converting the css3-border-image tests in http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-image/ and css3-border-radius tests in http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-radius/ to the CSSWG repository. See CSSWG Test Format http://wiki.csswg.org/test/css2.1/format and bug 691950 for background info.
assigned -> me as of now
Status: NEW → ASSIGNED
I am 99.9% sure that 'border-style: solid' declaration is missing (should be between lines 6 and 7) in http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-image/different-h-v-ref.html?force=1 Interestingly, such oversight reveals an implementation bug in Chrome 24.0.1312.56; without 'border-style: solid', we should see a blank page.
That's a very good point. I think we should copy this reference into a new test and compare it to a blank page (or a "there should be no red" page, with swatch-red.png instead of the diamonds), then fix this reference so that it works as designed.
I have fixed the reference (a patch is coming shortly..) already. I agree with you: we should create a separate, distinct test out of this discovery, from the original reftest (which had *not* 'border-style: solid'). I can do this..
Created attachment 707945 [details] [diff] [review] Patch for border-image tests **_only_** Notes about the patch and the border-image tests ------------------------------------------------ 1) '-khtml-border-image: [declarations]' has been removed everywhere since Gecko rendering engine doesn't support it and neither does Konqueror with the KHTML rendering engine! 2) Many tests would be better, more useful if border-image shorthand was broken into border-image sub-properties: for starters, the title would be easier to create, to edit ... but I guess it's because -moz- prefix was historically only applied to shorthand and not sub-properties 3) CSS code is sometimes not very compact: eg multicolor-image-3.html 4) there is a lot of inline style code: I have left such inline style code intact 5) I often could not come up with a decent title just because I couldn't figure out for sure what the test was really about 6) border-style: solid; /* added by G. Talbot */ was added in different-h-v-ref.html 7) There are about 8 tests which produce tiny+multicolor irregular polygons: not easy to examine... List of 3 files *NOT* to be switched over (full URLs) ----------------------------------------------------- http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-image/470250-2.html http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-image/470250-2-ref.html http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-image/border-image-outset-resize-1-ref.html Same list of 3 files *NOT* to be switched over (base-filenames only) -------------------------------------------------------------------- 470250-2.html 470250-2-ref.html border-image-outset-resize-1-ref.html
> we should copy this reference into a new > test and compare it to a blank page (or a "there should be no red" page, > with swatch-red.png instead of the diamonds), then fix this reference so > that it works as designed. I have added 'border-style: solid' into the patch with regards to http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-image/different-h-v-ref.html and I just created+hg-pushed http://test.csswg.org/source/contributors/gtalbot/incoming/border-image-slice-120.xht
Created attachment 709419 [details] [diff] [review] Patch for border-image tests **_and_** border-radius tests Nota bene: this patch is about both border-image and border-radius tests. **** Notes about border-radius tests ------------------------------- 1) I had to filename-rename 18 clipping-* files into corner-clipping-* 2) corner-clipping-002.html, corner-clipping-003.html, corner-joins-1.xhtml, iframe-1.html (and their respective reftests) should be using a non-red color as a neutral color. A <!-- comment --> has been inserted in those files. 3) color-layer-1a.html (and its reftest) requires over 600x600 viewport dimensions (800x1000): this test should be shrunk down proportionally 4) I do not see why color-layer-1-ref.html constitutes a valid, useful or worthy or reliable reftest for color-layer-1.html 5) do not shift over invalidate-1a.html, invalidate-1b.html, invalidate-1-ref.html, outline-circle.html, outline-ellips.html, outline-square.html, 6) intersecting-clipping-1-refc.html and intersecting-clipping-1-refi.html were respectively filename-renamed intersecting-clipping-1-ref-c.html and intersecting-clipping-1-ref-i.html 7) Tests with fractional pixel values: corner-joins-1.xhtml, scrollbar-clamping-1-ref.html, scrollbar-clamping-2-ref.html 8) No idea what is supposed to be testing zero-radius-clip-1.html; not sure what "(...) a scrollframe is 'inactive' if it can't be scrolled by blitting." in scroll-1.html is supposed to mean.. 9) Most likely these tests, all these border-radius tests, will eventually be filename-renamed once exported into http://test.csswg.org/suites/css3-background/ because of http://wiki.csswg.org/test/format#file-name-format When I do this, I will include in a <!-- comment --> what was the initial, original filename. 10) In a majority of tests, I could not figure out what the tests were actually testing. Therefore, the title text is incomplete and the link to spec may be inaccurate
> I have added 'border-style: solid' into the patch with regards to > http://mxr.mozilla.org/mozilla-central/source/layout/reftests/border-image/ > different-h-v-ref.html > > and I just created+hg-pushed > > http://test.csswg.org/source/contributors/gtalbot/incoming/border-image- > slice-120.xht This test has now been moved and filename-renamed as http://test.csswg.org/source/contributors/gtalbot/submitted/border-image-slice-003.xht and submitted: http://test.csswg.org/suites/css3-background/nightly-unstable/html4/border-image-slice-003.htm
Component: Layout: Misc Code → Layout
Product: Core Graveyard → Core
You need to log in before you can comment on or make changes to this bug.