[dev] Fwd: [bugs] [Bug 533] Changed - broken urls in generated code

[Max Kalika]

Quoting Anil Madhavapeddy <anil@recoil.org>:

> So for HTML, the only really valid entity form is '&blah;'.  How can
> a browser interpret a substring of a URL as an entity without the
> semicolon in there (urlencoded, of course) ?

Simple -- if the browser is buggy.  It scans the characters following the 
ampersand until it finds a string that could be considered an entity.  NS 2 did 
it way back when to "make it easier" for web page creators.  IE4 seems to have 
the same bug.  I has been fixed since.

Personally I don't think catoring to the broken-browser world is worth the 
trouble.  The only way I can think of truly fixing this problem is with 
javascript wrapper functions for every url -- bleh!  Otherwise, making sure 
that the field names match no known entity name, would be good enough.

just my 2...

--mk23