A collection of small fix's that only effect the new browser mode.

* Change title of html pages from "Python ..." to "PyDoc ...".
    
* Fixed unterminated div float for items returned without a header.
     example: str, None, True, False
    
* Added "topic?key=..." url command to explicitly get topics.  This is to avoid the shadowing when an object has the same name as a topic.

* Nicer parsing and error handling in the url handler.

new patch...

Adjusted a comment in the _gettopic method.

Everything else the same.

A few last minute changes.. I think this will be all.

Run topic contents through html.markup.  That makes ref:, pep:, and html: links if they exist.  (I meant to this earlier.)

Fix case where topic reference links are to objects rather than another topic.  (applies to keywords also.)

Skips making an empty reference section if there are no references with a topic.

These are all small changes, and nothing should be controversial in this as everything changed only effects the new html browser mode.

Sigh... that might have come a little earlier.

As it is, I'll review it before rc2.

Some comments on http://codereview.appspot.com/4090042

George, My apologies to you for the late corrections.  And thanks for doing this.

Eric, I replied to your comments on Rietveld.  Thanks for taking a look.

I'll wait until you have a chance to reply and test it, then upload a new patch with any needed changes.

Ron

Ok, after input from Eric,  Here is another patch.

Removed a set of unneeded parentheses.

Changed the title of the pages from PyDoc to Pydoc.

A better fix for the uncaught floats.  Wrap the main content in div with style="clear:both;". Should work on nearly everything.  Also avoided floating the "get" and "search" form inputs.  That was causing them to separate too much in IE.


To summarize ...

Minor fixes for new browser mode only.

Fix float problems with the navbar.
Change html page titles to "Pydoc" instead of "Python".
Call html.markup() on topic contents.
Fix topic/object shadowing.
Improved error handling. (Part of the shadowing fix.)

Replied on Rietveld.

Ezio, I think you have good HTML/CSS knowledge, so I’m adding you to nosy.  If you could read the short discussion and Rietveld and weigh in, it would be very helpful.

New and hopefully last patch... pydoc_misc_fix_e.diff

I removed the .html in the ?key= links as Eric suggested.

I checked the navbar float behavior on browsershots.org.  Multiple versions of MSIE, firefox, opera, chrome, and safari were tested on Ubuntu and Windows XP.  They all looked very close to each other. Better than I expected. :-)

As for clearing floats, there is the html clear="all" attribute, and a css style="clear:both;".  The browsershots.org tests confirms style="clear:both;" works as it should where I used it.

Cheers, and hopefully this is ready to go.  I believe it is.

I fixed two instances of missing HTML escaping and committed as r88261.  The code should be checked thoroughly for more such missing escaping.