From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Alan W. Irwin" To: Norman Walsh Cc: docbook-tools-discuss@sourceware.cygnus.com Subject: Re: blank line in lists Date: Wed, 31 Oct 2001 07:19:00 -0000 Message-id: References: <877ktcjazr.fsf@nwalsh.com> X-SW-Source: 2001-q4/msg00010.html Thanks, Norman, for that clear definition. The original poster had a good example of the problem. But I responded because I have often run into that problem myself. I worked around it by being very careful of leading nl's in my DocBook-XML source paragraphs in a mixed content environment, but I always thought of it as a bug that needed to be fixed. It's been 6 months since I have looked at this so the rest of my post is based on my memory of what went on then. So I may have some of the details wrong. But I proved to my satisfaction at that time that jade propagated the nl at the start of paragraphs in mixed environments. I don't think it does this for paragraphs in more ordinary environments, but I am not sure. Let's concentrate on the jade html backend although I suspect the problem occurs for all jade backends. To verify the problem I suggest you try jade with html output, and you should see that propagated nl *looking directly at the raw html file result* at the start of paragraphs in mixed environments like lists. In other words the generated html varys depending on whether that leading nl is in the docbook source or not. Then things get murky, because the browser interpretation of that leading nl in the HTML varies between browsers. As I recall mozilla ignored it while netscape 4.7 and konqueror paid attention to it leading to the problem that the original poster was discussing. Concentrating on the html here, if the html standards say leading nl's in paragraphs are valid in the context of lists or other mixed environments then it is up to the individual browser to ignore them, and obviously the problem is not jade's fault. However, if the html standards say leading nl's for paragraphs within lists are invalid, then this is jade's problem which should be fixed. In case of doubt about the html standards, I understand there are validators which should make it easy to tell whether the jade html output is valid for the leading nl situation we have been discussing. Hope this helps to clarify the problem. Alan email: irwin@beluga.phys.uvic.ca phone: 250-727-2902 FAX: 250-721-7715 snail-mail: Dr. Alan W. Irwin Department of Physics and Astronomy, University of Victoria, P.O. Box 3055, Victoria, British Columbia, Canada, V8W 3P6 __________________________ Linux-powered astrophysics __________________________ On 31 Oct 2001, Norman Walsh wrote: > / "Alan W. Irwin" was heard to say: > | Is he saying in a mixed content environment (I presume lists are examples of > | that) that no filling or justification of paragraphs is done? That is what > | would happen if white space were preserved. Of course, "preserved" is a > | stronger statement than "significant" so I guess what I really need is > | a definition of significant. > | > | Would Norm or some other expert here be willing to expand on the remark? > > "Significant" in this case means "passed through to the application". So, > if you say: > > > Some text > > What the application sees is "Begin a paragraph" "Newline" "Some text" > "End a paragraph". > > So if the application thinks newlines are significant in paragraphs, > you get an extra newline. > > What's really odd though is that I can't reproduce this problem > without the intervening indexterm. And I expect the intervening > indexterm is just a backend bug. > > (What are you generating? TeX, RTF, HTML, ... ?) > > Be seeing you, > norm > > -- > Norman Walsh | "Abstraction, abstraction and > http://nwalsh.com/ | abstraction." This is the answer to the > | question, "What are the three most > | important words in programming?"--Paul > | Hudak >