11630 2006-11-17 06:29:58 -0800 REGRESSION (r17068): Layout goes completely wrong 2006-11-21 14:19:05 -0800 1 1 1 Unclassified WebKit Layout and Rendering 420+ Mac (Intel) OS X 10.4 RESOLVED INVALID http://abadacapoeira.de HasReduction, Regression P1 Normal --- 1 te-web webkit-unassigned bdakin hyatt mitz sam oldest_to_newest 45613 0 te-web 2006-11-17 06:29:58 -0800 The content DIV is rendered far too wide. The rendering is a regression compared to the shipping Safari variant, and works in Opera, Firefox, and Windows IE6 and 7. Sorry, I am not enough a Webdesigner to strip it down, but the CSS & structure is already as easy as possible. The only "formal" bug is that the Webserver supplies the html as "ISO-Latin-xxxx" but the xhtml states UTF-8 encoding. 45615 1 11551 mitz 2006-11-17 07:16:07 -0800 Created attachment 11551 Reduction Opera agrees with TOT. Results are the same in strict mode. 45616 2 mitz 2006-11-17 07:29:22 -0800 The regression was in r17068 (fix for bug 9347). Based on the description of that bug, I think the new result is correct (making this bug invalid). 44938 3 11590 mitz 2006-11-20 14:06:20 -0800 Created attachment 11590 Text-transform eating up non-breaking space reduction "Completely wrong" in the bug summary may refer to the popup being on the right of the menu instead of under it. In that case, this is because of a text-transform: capitalize regression demonstrated here. In shipping Safari, Firefox and Opera, there's a space between the two 100x100 squares due to the text which includes non-breaking spaces. This is also the case in TOT if text-transform:capitalize is not specified. 44939 4 mitz 2006-11-20 14:24:54 -0800 From StringImpl::capitalize: // Replace &nbsp with a real space since ICU no longer treats &nbsp as a word separator. This was added in r13702 to address <rdar://problem/4502311> text-transform:capitalize needs to treat nbsp as a regular space when ICU changes. I think the solution might be as simple as copying from m_data[i - 1] instead of stringWithPrevious[i] in the loop that copies the non-capitalized characters. 44942 5 te-web 2006-11-20 15:49:10 -0800 Actually I meant the first reduction which - after reading a bunch of the CSS 2.1 spec - seems to be rendered correct - the second problem I considered more an cosemtic side-effect. Seriously, I did not expect for "absolute" positioning the viewport and not the (html or body) parent objects are to consider for positioning. Thanks for the promt response and reductions. 44815 6 mitz 2006-11-21 14:19:05 -0800 Closing this bug as invalid. Opened bug 11671 to track the text-transform: capitalize issue. Sorry for the mess. 11551 2006-11-17 07:16:07 -0800 2006-11-17 07:16:07 -0800 Reduction html width.html text/html 256 mitz PGh0bWwgc3R5bGU9IndpZHRoOiAxMDBweDsiPgogICAgPGJvZHk+CiAgICAgICAgPGRpdiBzdHls ZT0icG9zaXRpb246IGFic29sdXRlOyBsZWZ0OiAwOyB3aWR0aDogMTAwJTsgaGVpZ2h0OiAxMDBw eDsgYmFja2dyb3VuZDogbGlnaHRibHVlOyI+CiAgICAgICAgICAgIEluIEZpcmVmb3ggYW5kIGlu IHNoaXBwaW5nIFNhZmFyaSwgdGhpcyBpcyBhIDEwMCZ0aW1lczsxMDAgc3F1YXJlLgogICAgICAg IDwvZGl2PgogICAgPC9ib2R5Pgo8L2h0bWw+Cg== 11590 2006-11-20 14:06:20 -0800 2006-11-20 14:06:20 -0800 Text-transform eating up non-breaking space reduction capoeira.html text/html 179 mitz PHN0eWxlPgogICAgZGl2IGRpdiB7IHdpZHRoOiAxMDBweDsgaGVpZ2h0OiAxMDBweDsgYmFja2dy b3VuZDogbGlnaHRibHVlOyB9Cjwvc3R5bGU+CjxkaXYgc3R5bGU9InRleHQtdHJhbnNmb3JtOiBj YXBpdGFsaXplOyI+Cgk8ZGl2PjwvZGl2PiZuYnNwOyAgJm5ic3A7ICZuYnNwOzxkaXY+PC9kaXY+ CjwvZGl2Pgo=