Short URL: http://html5.org/r/5994
| SVN | Bug | Comment | Time (UTC) |
|---|---|---|---|
| 5994 | 10202 | apply wg decision | 2011-04-12 22:13 |
Index: source
===================================================================
--- source (revision 5993)
+++ source (revision 5994)
@@ -27968,14 +27968,32 @@
when it is used to label a potential <span>media
resource</span>.</p>
- <p class="note">In the absence of a <!-- pretty crazy -->
- specification to the contrary, the <span>MIME type</span>
- "<code>application/octet-stream</code>" when used <em>with</em>
- parameters, e.g.
- "<code>application/octet-stream;codecs=theora</code>", <em>is</em>
- <span>a type that the user agent knows it cannot render</span>,
- since that parameter is not defined for that type.</p>
+ <p class="note">
+ <!--END w3c-html-->
+ Only the <span>MIME type</span> <!-- the WG decision started with the next bit, which is not in the style of the spec -->
+ <!--START w3c-html-->
+ "<code>application/octet-stream</code>"
+ <!--END w3c-html-->
+ with no parameters
+ <!--START w3c-html-->
+ is special-cased here; if any parameter appears with it, it
+ <!--END w3c-html-->
+ will <!-- the WG decision that led to this text had a "should", but this is a non-normative section -->
+ <!--START w3c-html--><!--END html--><!--END complete--><!--END epub--><!--END dev-html-->
+ should
+ <!--START html--><!--START complete--><!--START epub--><!--START dev-html-->
+ be treated just like any other <span>MIME type</span>.
+ This is a deviation from the rule <!-- in RFC 2046, section 1,
+ paragraph 3 --> that unknown <span>MIME type</span> parameters
+ should be ignored.
+
+ <!-- but not really a "willful violation" since it's not that the
+ types are not being ignored, just that before the type is handled as
+ a type, there's a special case for a particular set of strings -->
+
+ </p>
+
<dl class="domintro">
<dt><var title="">media</var> . <code title="dom-navigator-canPlayType">canPlayType</code>(<var title="">type</var>)</dt>