[XQZone General] Re: Mozilla rendering bug affects CQ output
michael.blakeley at marklogic.com
Sat Oct 15 12:43:21 PDT 2005
> Date: Fri, 14 Oct 2005 17:37:27 -0400 (EDT)
> From: David Sewell <dsewell at virginia.edu>
> On Fri, 14 Oct 2005, Howard Katz wrote:
>>> You'll get the following output in XML:
> I don't know if this really counts as a Mozilla bug, since displaying
> unstyled XML in Mozilla/Firefox is usually not the most satisfying way
> to view XML... Anyway, I just tried out saving
> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
> to "test.xml" on my local machine, then opening it with Mozilla: it
> gives the display you document, but "View Page Source" shows the
> complete underlying XML. At least Mozilla does better than Safari, which
> displays *nothing* on screen if you try to view the same XML using
> either CQ or by opening an XML file!
Thanks for this email. I'm thinking about more on-line help for a future
release of cq, and this would make a good FAQ entry.
The cq tool is a workshop project, not an officially supported product.
But as someone who hacks on its code now and then, I don't consider this
XML display issue to be a cq bug.
By design, cq relies on your browser to render any of the three
content-types it offers to set: text/xml, text/html, and text/plain.
Your browser then does... whatever it does.
You've pointed out one way to produce strange output with text/xml. To
see a similar problem in text/html, submit a query like
'<title>hello</title>' with the html button.
Since html title tags are meaningless outside the head, most browsers
will simply ignore them, and the output will appear to be blank (but
visible in "view source"). We use this as a pitfall for the unwary, in
the developer course: it helps to teach the difference between
<title>foo</title> and <title>foo</title>/text(), and reminds us that we
rely on the browser to render our html.
Personally, I use the "text" button most of the time. The browser
renders the results more quickly, and I can always re-submit the query
for xml or html output, if I want pretty results.
So is this XML display issue a mozilla bug? That's up to the mozilla
folks, but it's been filed as
https://bugzilla.mozilla.org/show_bug.cgi?id=175946 by someone, and has
been in status NEW since 2002-10-21. Duplicates seem to be filed every
6-12 months, so lots of users do consider this behavior to be a bug.
On a happier note: we quietly released a new version of cq last week.
* streamlined UI, with tool-tips
* improved database select-list, using configured application-server roots
* improved error display, with line-number highlighting
* display of caret position (line-number and character position)
* CSS-based layout and frame-size autocalculation
* query-history tab, with worksheet support
* "list all" limits results to first 1000 documents
* resizable text areas
* requires MarkLogic Server 3.0-1 or later
The legacy release remains available for 2.2 servers.
More information about the General