[html5] r1777 - /

whatwg at whatwg.org whatwg at whatwg.org
Mon Jun 16 00:19:24 PDT 2008


Author: ianh
Date: 2008-06-16 00:19:23 -0700 (Mon, 16 Jun 2008)
New Revision: 1777

Modified:
   index
   source
Log:
[e] (0) notes for urls, and typo fixes

Modified: index
===================================================================
--- index	2008-06-13 19:16:08 UTC (rev 1776)
+++ index	2008-06-16 07:19:23 UTC (rev 1777)
@@ -25,7 +25,7 @@
 
    <h1 id=html-5>HTML 5</h1>
 
-   <h2 class="no-num no-toc" id=draft>Draft Recommendation — 13 June
+   <h2 class="no-num no-toc" id=draft>Draft Recommendation — 16 June
     2008</h2>
 
    <p>You can take part in this work. <a
@@ -6210,22 +6210,39 @@
      doesn't have a path hierarchy (e.g. data:, javascript:, about:blank
      URLs)
 
-    <li>define how to handle URIs that contain non-ascii characters (and it's
-     not the same as handling as IRIs, especially if the character encoding
-     of the document isn't UTF-8)
+    <li>
+     <p>define how to handle URIs that contain non-ascii characters (and it's
+      not the same as handling as IRIs, especially if the character encoding
+      of the document isn't UTF-8)</p>
 
+     <p>First parse it (we need to define that. For some schemes it's not per
+      spec -- e.g. apparently for ftp: we should split from hosts on ';').
+      Then handle each bit as follows:</p>
+
+     <p>scheme: no further processing (treat %-escaped characters literally,
+      treat unicode characters as unicode characters).</p>
+
+     <p>host: expand %-encoded bytes to Unicode as UTF-8, treat unicode
+      characters as per IDN.</p>
+
+     <p>path: don't expand %-encoded bytes. Re-encode unicode to UTF-8 and
+      percent-encode.</p>
+
+     <p>query: don't expand %-encoded bytes. Re-encode unicode to the page's
+      encoding. Do not percent-encode.</p>
+
     <li>define URL as IRI and replace all uses of URIs and IRIs in the rest
      of the spec with URL
 
     <li>define "valid URL" in terms of valid URIs and IRIs and replace that
      concept through the spec to a pointer here, getting rid of references to
      <a href="#refsRFC3986">[RFC3986]</a> <a
-     href="#refsRFC3987">[RFC3987]</a>
+     href="#refsRFC3987">[RFC3987]</a> outside this section
 
     <li>define "valid URL reference" in terms of valid URI references and IRI
      references and replace that concept through the spec to a pointer here,
      getting rid of references to <a href="#refsRFC3986">[RFC3986]</a> <a
-     href="#refsRFC3987">[RFC3987]</a>
+     href="#refsRFC3987">[RFC3987]</a> outside this section
 
     <li>define how to parse URLs (maybe by deferring to RFCs)
 
@@ -9467,8 +9484,8 @@
          outlinee</var> element.
 
        <li>
-        <p><i>Loop:</i> If <var title="">current section</var> has no child
-         sections, stop these steps.
+        <p><i>Finding the deepest child:</i> If <var title="">current
+         section</var> has no child sections, stop these steps.
 
        <li>
         <p>Let <var title="">current section</var> be the last child <a
@@ -9476,7 +9493,7 @@
          <var title="">current section</var>.
 
        <li>
-        <p>Go back to the substep labeled <i>Loop</i>.
+        <p>Go back to the substep labeled <i>finding the deepest child</i>.
       </ol>
 
      <dt>When exiting a <a href="#sectioning0">sectioning content</a> element
@@ -9525,10 +9542,10 @@
 
        <li>
         <p>If the element being entered has a <a href="#rank">rank</a> lower
-         than the rank of the heading of the <var title="">candidate
-         section</var>, then create a new <a href="#section0"
-         title=concept-section>section</a>, and append it to <var
-         title="">candidate section</var>. (This does not change which
+         than the <a href="#rank">rank</a> of the heading of the <var
+         title="">candidate section</var>, then create a new <a
+         href="#section0" title=concept-section>section</a>, and append it to
+         <var title="">candidate section</var>. (This does not change which
          section is the last section in the outline.) Let <var
          title="">current section</var> be this new section. Let the element
          being entered be the new heading for the <var title="">current

Modified: source
===================================================================
--- source	2008-06-13 19:16:08 UTC (rev 1776)
+++ source	2008-06-16 07:19:23 UTC (rev 1777)
@@ -4671,23 +4671,35 @@
     URL doesn't have a path hierarchy (e.g. data:, javascript:,
     about:blank URLs)</li>
 
-    <li>define how to handle URIs that contain non-ascii characters
-    (and it's not the same as handling as IRIs, especially if the
-    character encoding of the document isn't UTF-8)</li>
+    <li>
 
+     <p>define how to handle URIs that contain non-ascii characters
+     (and it's not the same as handling as IRIs, especially if the
+     character encoding of the document isn't UTF-8)</p>
+
+     <p>First parse it (we need to define that. For some schemes it's
+     not per spec -- e.g. apparently for ftp: we should split from
+     hosts on ';'). Then handle each bit as follows:</p>
+     <p>scheme: no further processing (treat %-escaped characters literally, treat unicode characters as unicode characters).</p>
+     <p>host: expand %-encoded bytes to Unicode as UTF-8, treat unicode characters as per IDN.</p>
+     <p>path: don't expand %-encoded bytes. Re-encode unicode to UTF-8 and percent-encode.</p>
+     <p>query: don't expand %-encoded bytes. Re-encode unicode to the page's encoding. Do not percent-encode.</p>
+
+    </li>
+
     <li>define URL as IRI and replace all uses of URIs and IRIs in the
     rest of the spec with URL</li>
 
     <li>define "valid URL" in terms of valid URIs and IRIs and replace
     that concept through the spec to a pointer here, getting rid of
     references to <a href="#refsRFC3986">[RFC3986]</a> <a
-    href="#refsRFC3987">[RFC3987]</a></li>
+    href="#refsRFC3987">[RFC3987]</a> outside this section</li>
 
     <li>define "valid URL reference" in terms of valid URI references
     and IRI references and replace that concept through the spec to a
     pointer here, getting rid of references to <a
     href="#refsRFC3986">[RFC3986]</a> <a
-    href="#refsRFC3987">[RFC3987]</a></li>
+    href="#refsRFC3987">[RFC3987]</a> outside this section</li>
 
     <li>define how to parse URLs (maybe by deferring to RFCs)</li>
 
@@ -7687,14 +7699,16 @@
        section in the <span>outline</span> of the <var title="">current
        outlinee</var> element.</p></li>
 
-       <li><p><i>Loop:</i> If <var title="">current section</var> has
-       no child sections, stop these steps.</p></li>
+       <li><p><i>Finding the deepest child:</i> If <var
+       title="">current section</var> has no child sections, stop
+       these steps.</p></li>
 
        <li><p>Let <var title="">current section</var> be the last
        child <span title="concept-section">section</span> of the
        current <var title="">current section</var>.</p></li>
 
-       <li><p>Go back to the substep labeled <i>Loop</i>.</p></li>
+       <li><p>Go back to the substep labeled <i>finding the deepest
+       child</i>.</p></li>
 
       </ol>
 
@@ -7751,7 +7765,7 @@
        title="">current section</var>.</p></li>
 
        <li><p>If the element being entered has a <span>rank</span>
-       lower than the rank of the heading of the <var
+       lower than the <span>rank</span> of the heading of the <var
        title="">candidate section</var>, then create a new <span
        title="concept-section">section</span>, and append it to <var
        title="">candidate section</var>. (This does not change which




More information about the Commit-Watchers mailing list