<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.26.0">
</HEAD>
<BODY>
First, I agree with Jim. I think removing the feature entirely until it can be properly addressed is appropriate.<BR>
<BR>
This works fine in CSS:<BR>
<BR>
<BLOCKQUOTE>
a[href $='.txt'] {<BR>
padding: 0 20px 0 0;<BR>
background: transparent url(images/icons/file-txt.png.xhtml) no-repeat<BR>
center right;<BR>
}<BR>
<BR>
Notice the relative reference-URL. This allows resource loading from any source, as long as you can "do the math" and calculate the relative URL.<BR>
</BLOCKQUOTE>
<BR>
However, on side note...<BR>
<BR>
<BLOCKQUOTE>
On Fri, 2009-10-02 at 14:39 -0700, Jim Driscoll wrote:<BR>
</BLOCKQUOTE>
<BLOCKQUOTE TYPE=CITE>
<BLOCKQUOTE>
<TT><FONT COLOR="#1a1a1a">but once we allow it, </FONT></TT><BR>
<TT><FONT COLOR="#1a1a1a">we can't "unallow" it when a better method comes along.</FONT></TT><BR>
</BLOCKQUOTE>
</BLOCKQUOTE>
<BR>
I disagree. I think that backwards compatibility has pluses and minuses, but we should not rule out removing features that don't make sense anymore, or changing features that break something in order to achieve a greater good for the future of JSF. It should be done sparingly, but is sometimes necessary.<BR>
<BR>
<TABLE CELLSPACING="0" CELLPADDING="0" WIDTH="100%">
<TR>
<TD>
--<BR>
<B>Lincoln Baxter, III</B><BR>
Co-Founder of <A HREF="http://ocpsoft.com">OcpSoft</A><BR>
<BR>
Creator of:<BR>
<A HREF="http://ocpsoft.com/prettyfaces">PrettyFaces</A>: URL rewriting for JSF<BR>
<A HREF="http://ocpsoft.com/prettytime">PrettyTime</A>: Java elapsed timestamp formatting<BR>
<BR>
<BR>
</TD>
</TR>
</TABLE>
</BODY>
</HTML>