Re: content-072.htm is invalid

On 10/25/2010 02:59 PM, L. David Baron wrote:
> On Monday 2010-10-25 14:52 -0700, fantasai wrote:
>> On 10/25/2010 02:20 PM, Tab Atkins Jr. wrote:
>>>
>>> Are you looking at a different test?  The test linked above uses defer
>>> without an explicit value - "<script defer
>>> type="text/javascript"></script>".
>>>
>>> Yes, if the element had "defer='defer'", then attr(defer) should
>>> certainly return "defer".  That's not the case here.
>>
>> Yeah, I was reading through this thread and thinking "uh-oh. I know
>> exactly what the problem is here" and sure enough I get to your message...
>>
>> The build process minimizes boolean attributes when it converts from
>> XHTML to HTML. I can turn this option off, but it would be off for all
>> boolean attributes. If that's ok with everyone, I'll do that for the
>> next build. If it's not, then we'll need to figure out what the default
>> behavior should be and handle exceptions somehow.
>
> Can you generate a diff of the generated HTML before and after the
> change so people can review it?

Here we are:

Only in css2.1/: htaccess
Only in 20101027/: .htaccess
diff -ur css2.1/html4/content-072.htm 20101027/html4/content-072.htm
--- css2.1/html4/content-072.htm        2010-10-27 01:23:55.000000000 -0700
+++ 20101027/html4/content-072.htm      2010-10-27 01:16:11.000000000 -0700
@@ -22,6 +22,6 @@
      </head>
      <body>
          <p>Test passes if the word "defer" appears in the box below.</p>
-        <script defer type="text/javascript"></script>
+        <script defer="defer" type="text/javascript"></script>
      </body>
  </html>
\ No newline at end of file
Only in 20101027/: other

(I'm kinda surprised we aren't using boolean attributes anywhere else.
Granted many of them are involved in forms, and we have few tests with
forms, obviously.)

~fantasai

Received on Wednesday, 27 October 2010 09:07:15 UTC