Version: 3.5.7 (using KDE 3.5.7, Kubuntu (feisty) 4:3.5.7-0ubuntu1~feisty2) Compiler: Target: x86_64-linux-gnu OS: Linux (x86_64) release 2.6.20-16-generic The html is <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/1999/xhtml"> <body> <div style="position:relative"> <img src="a.gif"> <img style="position:absolute;left:0px;height:183px;width:120px;" src="b.jpg"> <div>FOO</div> </div> </body> </html> i assume it is a bug because removing the FOO div i get the same rendering than in Firefox and IE so it seems Konqi's fault I will attach the html and both images later.
Created attachment 21617 [details] The html file
Created attachment 21618 [details] One needed image
Created attachment 21619 [details] The other needed image
Created attachment 21622 [details] online test case
> removing the FOO div i get the same rendering than in Firefox which means it's a problem with how anonymous blocks are created. It should include the statically positioned object in the inline.
Created attachment 21623 [details] patch here... I can't test at the moment but just syncing getInlineRun with what's in WebCore should do. The git of it is to skip positioned/floating objects while scanning for inlines. Anyone may test?
Your patch does not compile, replacing boundary with stop makes it compile and work, you rock :-) Do you want me to commit or you can do it?
yes, feel free, thanks... I'm not sure I'll be able to fix my build before some days. I'll regression test the patch later. Please mention the WebCore origin in the commit.
SVN commit 712197 by aacid: Patch by Germain Garand to fix the bug i reported yesterday, you rock ;-) In his words "The git of it is to skip positioned/floating objects while scanning for inlines." "Ported" from WebCore BUGS: 149809 M +21 -15 render_block.cpp WebSVN link: http://websvn.kde.org/?view=rev&revision=712197
Germain i did not commit to KDE4 because i'm not able to start konqueror and make it open html with khtmlpart instaed of katepart :-S I didn't even check the patch applies That's left for you too besides the regression testing :-)
Albert: the patch was in fact made for trunk. It's most probably OK to apply it to branch, but I won't take blame for it in case it breaks something. I just don't compile/test much branch those days.
ok, i commited to trunk too, should i revert 3.5 ?
thanks! I finally managed to have a working tree so I'll be able to regression test at last ;-/ As for 3.5, all I say is I won't test it there, so I can hardly take full responsibility for it. But OTOH the patch is really simple...