CSS2 Format of XML file orginally written one year ago or so that worked in M16, Netscape PR1, Opera 4.02, and Internet Explorer 5, fails in Mozilla 0.9.6 (also fails in Netscape 6.2). The XML and CSS files check as valid. File displays OK in Internet Explorer 5.5 and Opera 6. Originally M16 did the best followed by Opera 4.02. Internet Explorer was acceptable but a few details of format failed because IE versions I used did not support CSS2 selectors. Two lines in the test case should have the same format, white text on a black background but the second line is skewed to the right of the page without the background. Fonts appear small. I do not know if this problem occurs in other versions or other OS versions of Mozilla. The testcase file is a snippet of a book in XML for the HWG project Gutenberg project. The format problem occurs near the beginning of the file. In the test case file, intermediate text is removed between the two lines using the <title> element (not the HTML title element - this is a non-HTML XML DTD). With the full book file, text following the lines in the test case are also skewed to the right of the page in a narrow column. The files in this directory (http://www.degiere.com/mozilla/) are: index.htm [which links to the following files:] hydea10-testcase.xml [short testcase file] hydea10-testcase-no-dtd.xml [well-formed short testcase file without DTD referenced] hydea10.xml [original full book - same problem occurs at the end of the document] hydea10-no-dtd.xml [well-formed original full book without DTD referenced] gutbkplaypg1.dtd [main DTD for the XML document] bkplaypg.dtd [second DTD referenced as an entity by the gutbkplaypg1.dtd] hydea10-testcase.css [short version of CSS file for the test case XML docs (hydea10-testcase.xml and hydea10-testcase-no-dtd.xml)] hydea10.css [full CSS file for hydea10.xml document]

over to Style System adding simplesect{display:block;} fixes the problem. however, I'm not sure what the correct behaviour should be. Terence: the hydea10-testcase-no-dtd.xml link on your page points to the wrong file.

reassigning

Is this a duplicate of bug 86276?

Looks similar.... block inside inline (not really a valid CSS2 construct). Symptoms are a little different, but not much...

It's HTML where block inside inline isn't valid. It's perfectly valid in CSS.

Frank Boumphrey's DTD's specify that the element -title- is inline and that the element -simplesect- is a block element [bkplaypg.dtd] The test case example CSS does not specify -simplesect- as anything, but redefines -title- for block display. The other surrounding elements in the other DTD [gutbkplaypg1.dtd] are not defined as block or inline (I do not know what the default interpretation is supposed to be in this case). Is the user agent supposed to ignore or use block/inline definitions in the DTD for XML display with CSS? This could lead to different results if the DTD were not referenced since block/inline information would be absent without the DTD and would require it to be in the CSS. Mozilla's current behavior seems logical, forcing the issue on CSS rather than the on a DTD, which would make it work whether the DTD was present or not, but its interpretation is different from IE and Opera and some earlier Mozilla builds. Is this a case of my writing bad practice CSS, wherein I have learned a lesson, or is it an interpretation of standards issue, or is it really a bug? I have put up a slightly simpler test case file testcase-2.xml and testcase-2.css available through the index file on the test case URL. I changed the text colors in the CSS in case Mozilla is set to a default white background so you can see the text. Also, in the files with the "problem" if the browser window horizontal width is changed reflow of the first line occurs if the width is reduced, but does not occur if the width is increased. The display behavior differences between Mozilla versus IE and Opera remain the same.

Things like %block; and %inline; in the HTML4 DTD are just the names of parameter entities and have nothing to do with CSS's concepts of block and inline (if that's what you meant by "block/inline definitions in the DTD").

to dbaron -- Thanks for the clarification. I'll let you guys determine whether this is something that needs to be fixed

*** This bug has been marked as a duplicate of 86276 ***