22771 2008-12-09 17:04:17 -0800 table with display: inline; does not work with proper DOCTYPE 2012-10-02 11:33:00 -0700 1 1 1 Unclassified WebKit Tables 525.x (Safari 3.2) PC Windows XP RESOLVED INVALID HasReduction P2 Normal --- 0 jasneet webkit-unassigned jasneet mrahaman mustaf.here robert spottabathini vswap65 oldest_to_newest 102042 0 jasneet 2008-12-09 17:04:17 -0800 I Steps: Go to attached testcase II Issue: The table should be next to the input because it has "display: inline;" III Other Browsers: IE7: ok FF3: ok IV Nightly tested: 39088 Bug in Chromium : http://code.google.com/p/chromium/issues/detail?id=2032 102043 1 25905 jasneet 2008-12-09 17:04:36 -0800 Created attachment 25905 testcase 409484 2 mustaf.here 2011-05-24 23:20:52 -0700 The issue description (Bug in Chromium : http://code.google.com/p/chromium/issues/detail?id=2032 ) mentions that the issue is present if we have the DOCTYPE, removing the DOCTYPE renders the table inline as expected. Upon investigation, it is found that RenderStyle for RenderObject (Table element ) is noninherited_flags._effectiveDisplay = 7 (INLINE_TABLE) , when DOCTYPE is NOT mentioned & table renders inline. But when DOCTYPE is mentioned, the same property has value 0 (INLINE) & the table does not render inline. The display type setting is done in CSSStyleSelector::adjustRenderStyle() { if (style->display() != NONE) { // If we have a <td> that specifies a float property, in quirks mode we just drop the float // property. // Sites also commonly use display:inline/block on <td>s and <table>s. In quirks mode we force // these tags to retain their display types. if (!m_checker.m_strictParsing && e) { if (e->hasTagName(tdTag)) { style->setDisplay(TABLE_CELL); style->setFloating(FNONE); } else if (e->hasTagName(tableTag)) style->setDisplay(style->isDisplayInlineType() ? INLINE_TABLE : TABLE); } } Further analysis found that when DOCTYPE is mentioned, the document m_compatibilityMode = LimitedQuirksMode, therefore m_checker.m_strictParsing = TRUE. When DOCTYPE is NOT mentioned m_compatibilityMode == QuirksMode, therefore m_checker.m_strictParsing = FALSE So, as we can see from above, when m_checker.m_strictParsing = FALSE, we set the display for table element as INLINE_TABLE. If m_checker.m_strictParsing == TRUE (DOCTYPE is mentioned), the if block is ommitted & the display is NOT set to INLINE. Can any one help me understand why this code is present in CSSStyleSelector::adjustRenderStyle(), I think if we remove this, the issue would be fixed. 409915 3 ap 2011-05-25 14:19:14 -0700 Did you try following svn blame to find out when this code was added? 410177 4 mustaf.here 2011-05-25 21:25:22 -0700 Thanks Alexey. I got some info as to when this code was added by using svn blame. However, I finally found that it is the problem with the test case. To make a table display inline, "display:inline" is not enough, we have to specify "display:inline-table" (enum INLINE_TABLE in the code). Please refer to http://www.webkit.org/blog/115/webcore-rendering-ii-blocks-and-inlines/ as well http://www.w3schools.com/css/pr_class_display.asp When DOCTYPE is not mentioned i.e. in Quirk mode, the code applies lesser strictness, therefore allow "display:inline" to be displayed as inline table. But when we mention DOCTYPE, applying of style becomes more strict & it displays an inline table if & only if we mention "display:inline-table". I have verified by making the following changes in the test case & now the table is rendered inline in Winlauncher & Safari. <table style="display:inline-table"> Therefore, it is a test case issue, no problem with the code. Can I change the status of the bug & How? 544881 5 mrahaman 2012-01-29 21:53:50 -0800 Based on the discussion above, can someone please mark the issue as RESOLVED WONTFIX? 25905 2008-12-09 17:04:36 -0800 2008-12-09 17:04:36 -0800 testcase testcase.zip application/zip 728 jasneet UEsDBAoAAAAAAG2IiTkAAAAAAAAAAAAAAAAJAAAAdGVzdGNhc2UvUEsDBBQAAAAIAHqJdTkRzoEc uAAAAPUAAAARAAAAdGVzdGNhc2Uvb3JnLmh0bWxNjc0KwjAQhO+C77Dm4smu0pvWHmwLCv4UiajH SoINxLSkK23e3kYqeNiZYZdvJ5qkp4Tf8wxKemnIL5v9LgE2Q7yGCWLKU7ht+WEPi2AO3BamUaQq U2jE7MiAlUT1ErFt26ANg8o+kZ+x878WHh7ijP7IQJBg8XgU+ZP3RyWcd2XqNwG5Wq4ZyY4YAPo9 FQ8toSGn5XoqVFPrwi1BGa2MXE3jiGw/IibZUIR96MX28sU8j78CHBo/UEsDBBQAAAAIABqLdTng /EzsvAAAAAUBAAAbAAAAdGVzdGNhc2UvdGVzdF9zdGFuZGFyZC5odG1sTY7NCsIwEITvgu+w5uLJ rsVbbXuwLSj4h0TUYyVBAzENzUrbt7cpCp7mY2dnduNJfsj47VjAk14ajufVdpMBmyFeFhliznO4 rvluC2EwB16XxilSlSk1YrFnwJ5ENkJsmiZoFkFVP5CfsPVdoQ9/cUZ/yUCQYOl4FHvL670SnVdl 7JuAOisTRrIlBoB+TuVdS3DUaZlMhXJWl10Eymhl5HLabwDEVA/qSaQkHcXYw2Dh4PXiawb6HcTv Bx9QSwECFAAKAAAAAABtiIk5AAAAAAAAAAAAAAAACQAAAAAAAAAAABAAAAAAAAAAdGVzdGNhc2Uv UEsBAhQAFAAAAAgAeol1ORHOgRy4AAAA9QAAABEAAAAAAAAAAAAgAAAAJwAAAHRlc3RjYXNlL29y Zy5odG1sUEsBAhQAFAAAAAgAGot1OeD8TOy8AAAABQEAABsAAAAAAAAAAAAgAAAADgEAAHRlc3Rj YXNlL3Rlc3Rfc3RhbmRhcmQuaHRtbFBLBQYAAAAAAwADAL8AAAADAgAAAAA=