Redo validation errors in the IPv4 parser

annevk · web-flow · commit f0806d616a66 · 2023-01-20T12:15:25.000+01:00
ab0e820 didn't consider the impact on the validation error infrastructure. This also makes some minor additional changes: * Modernizes the "Host parsing" section. * Clarifies that the IPv4 parser cannot be invoked directly. * Clarifies that the IPv6 parser can, but you should still not do it. Fixes #706.
diff --git a/url.bs b/url.bs
@@ -414,7 +414,8 @@ point <a for=/>URLs</a> from <var>A</var> can come from untrusted sources.
 
 <div class=example id=example-host-parsing>
  <p>A <a lt="host parser">parse</a>-<a lt="host serializer">serialize</a> roundtrip gives the
- following results, depending on the <var>isNotSpecial</var> argument to the <a>host parser</a>:
+ following results, depending on the <var ignore>isNotSpecial</var> argument to the
+ <a>host parser</a>:
 
  <table>
   <tr>
@@ -732,9 +733,10 @@ to be distinguished.
 
 <h3 id=host-parsing>Host parsing</h3>
 
+<div algorithm>
 <p>The <dfn export id=concept-host-parser lt="host parser|host parsing">host parser</dfn> takes a
 <a>scalar value string</a> <var>input</var> with an optional boolean <var>isNotSpecial</var>
-(default false), and then runs these steps:
+(default false), and then runs these steps. They return failure or a <a for=/>host</a>.
 
 <ol>
  <li>
@@ -772,11 +774,13 @@ to be distinguished.
 
  <li><p>Return <var>asciiDomain</var>.
 </ol>
+</div>
 
 <hr>
 
+<div algorithm>
 <p>The <dfn>ends in a number checker</dfn> takes an <a>ASCII string</a> <var>input</var> and then
-runs these steps:
+runs these steps. They return a boolean.
 
 <ol>
  <li><p>Let <var>parts</var> be the result of <a>strictly splitting</a> <var>input</var> on
@@ -808,26 +812,24 @@ runs these steps:
 
  <li><p>Return false.
 </ol>
+</div>
 
+<div algorithm>
 <p>The <dfn id=concept-ipv4-parser>IPv4 parser</dfn> takes an <a>ASCII string</a> <var>input</var>
-and then runs these steps:
+and then runs these steps. They return failure or an <a for=/>IPv4 address</a>.
 
-<ol>
- <li>
-  <p>Let <var>validationError</var> be false.
-
-  <p class=note>This uses <var>validationError</var> to track <a>validation errors</a> to avoid
-  reporting them before we are confident we want to parse <var>input</var> as an IPv4 address as the
-  <a>host parser</a> almost always invokes the <a>IPv4 parser</a>.
+<p class=note>The <a for=/>IPv4 parser</a> is not to be invoked directly. Instead check that the
+return value of the <a for=/>host parser</a> is an <a for=/>IPv4 address</a>.
 
+<ol>
  <li><p>Let <var>parts</var> be the result of <a>strictly splitting</a> <var>input</var> on
  U+002E (.).
 
  <li>
   <p>If the last <a for=list>item</a> in <var>parts</var> is the empty string, then:
 
   <ol>
-   <li><p>Set <var>validationError</var> to true.
+   <li><p><a>Validation error</a>.
 
    <li><p>If <var>parts</var>'s <a for=list>size</a> is greater than 1, then <a for=list>remove</a>
    the last <a for=list>item</a> from <var>parts</var>.
@@ -849,18 +851,11 @@ and then runs these steps:
 
    <li><p>If <var>result</var> is failure, <a>validation error</a>, return failure.
 
-   <li><p>If <var>result</var>[1] is true, then set <var>validationError</var> to true.
+   <li><p>If <var>result</var>[1] is true, <a>validation error</a>.
 
    <li><p><a for=list>Append</a> <var>result</var>[0] to <var>numbers</var>.
   </ol>
 
- <li>
-  <p>If <var>validationError</var> is true, <a>validation error</a>.
-
-  <p class="note">At this point each part was parsed into a number and <var>input</var> will be
-  treated as an IPv4 address (or failure). And therefore error reporting resumes.
- </li>
-
  <li><p>If any item in <var>numbers</var> is greater than 255, <a>validation error</a>.
 
  <li><p>If any but the last <a for=list>item</a> in <var>numbers</var> is greater than 255, then
@@ -888,7 +883,9 @@ and then runs these steps:
 
  <li><p>Return <var>ipv4</var>.
 </ol>
+</div>
 
+<div algorithm>
 <p>The <dfn>IPv4 number parser</dfn> takes an <a>ASCII string</a> <var>input</var> and then runs
 these steps:
 
@@ -939,11 +936,16 @@ these steps:
 
  <li><p>Return (<var>output</var>, <var>validationError</var>).
 </ol>
+</div>
 
 <hr>
 
+<div algorithm>
 <p>The <dfn id=concept-ipv6-parser>IPv6 parser</dfn> takes a <a>scalar value string</a>
-<var>input</var> and then runs these steps:
+<var>input</var> and then runs these steps. They return failure or an <a for=/>IPv6 address</a>.
+
+<p class=note>The <a for=/>IPv6 parser</a> could in theory be invoked directly, but please discuss
+actually doing that with the editors of this document first.
 
 <ol>
  <li><p>Let <var>address</var> be a new <a>IPv6 address</a> whose <a>IPv6 pieces</a> are all 0.
@@ -1089,11 +1091,14 @@ these steps:
 
  <li><p>Return <var>address</var>.
 </ol>
+</div>
 
 <hr>
 
+<div algorithm>
 <p>The <dfn export id=concept-opaque-host-parser>opaque-host parser</dfn> takes a
-<a>scalar value string</a> <var>input</var>, and then runs these steps:
+<a>scalar value string</a> <var>input</var>, and then runs these steps. They return failure or an
+<a for=/>opaque host</a>.
 
 <ol>
  <li><p>If <var>input</var> contains a <a>forbidden host code point</a>,
@@ -1108,6 +1113,7 @@ these steps:
  <li><p>Return the result of running <a for=string>UTF-8 percent-encode</a> on <var>input</var>
  using the <a>C0 control percent-encode set</a>.
 </ol>
+</div>
 
 
 <h3 id=host-serializing>Host serializing</h3>
