29532 2016-03-15 15:23:41 +0000 The CSS "counter-reset" in nested lists does not follow expected numeration in certain situations. 2016-03-15 15:37:43 +0000 1 1 1 Unclassified CSS Lists and Counters unspecified PC Linux RESOLVED INVALID P2 normal --- 1 iamio jackalmage public-css-bugzilla oldest_to_newest 125503 0 iamio 2016-03-15 15:23:41 +0000 Steps to reproduce: I tried to use the CSS Counter property to format the numbering of a nested ordered list. However, because the order list had slightly improper formatting, the numbering is wrong. Because the nested OL is generated in a content-editable <div>, the ordered list is not properly formatted. The sublist <ol> are direct children of the parent <ol>, instead of being of sublists being children of <li> elements. Proper formatting of the list would appear as follows: <ol> <li>item <ol> <li>subitem</li> </ol> </li> </ol> The improper html generated from manipulating lists in a content-editable div is as follows: <ol> <li>item</li> <ol> <li>subitem</li> </ol> </ol> ACTUAL RESULTS In poorly formatted lists, the CSS counter does coincide with the expected list numeration. (1) expected value: (1) (1.1) expected value: (1.1) (1.2) expected value: (1.2) (1.2.1) expected value: (1.2.1) (1.2.2) WRONG VALUE. expected value: (1.3) (1.2.3) WRONG VALUE. expected value: (1.4) (1.3) WRONG VALUE. expected value: (2) EXPLANATION This result happens because the scope of CSS counter-reset affects both descendant elements and subsequent sibling elements. The sublist counter-reset scope takes precedence over higher-level counter-reset. SCOPE OF PROBLEM This numeration issue doesn't only affect invalid HTML lists. It would also affect any HTML representation of nested elements. I want to create a visual representation of set theory by using nested div elements that are labeled with counter-reset values. <div id="whole set"> 1 <div id="first subset"> 1.1 <div id="first sub-subset"> 1.1.1 </div> <div> <div id="second subset">1.1.2</div> </div> 125504 1 iamio 2016-03-15 15:37:43 +0000 I am really sorry. I was working on reporting a possible issue, but I accidentally submitted this bug. Please disregard and close this issue.