Require spaces between attributes

kornelski · May 22, 2023 · 3562bff · 3562bff
1 parent 1eefdc9
commit 3562bff
Show file tree

Hide file tree

Showing 5 changed files with 12 additions and 9 deletions.
diff --git a/src/reader/parser.rs b/src/reader/parser.rs
@@ -226,6 +226,7 @@ pub enum OpeningTagSubstate {
     AfterAttributeName,
 
     InsideAttributeValue,
+    AfterAttributeValue,
 }
 
 #[derive(Copy, Clone, PartialEq)]

diff --git a/src/reader/parser/inside_opening_tag.rs b/src/reader/parser/inside_opening_tag.rs
@@ -75,7 +75,7 @@ impl PullParser {
                                 Some(this.error(SyntaxError::CannotUndefinePrefix(ln.into())))
                             } else {
                                 this.nst.put(name.local_name.clone(), value);
-                                this.into_state_continue(State::InsideOpeningTag(OpeningTagSubstate::InsideTag))
+                                this.into_state_continue(State::InsideOpeningTag(OpeningTagSubstate::AfterAttributeValue))
                             }
                         }
 
@@ -86,7 +86,7 @@ impl PullParser {
                                     Some(this.error(SyntaxError::InvalidDefaultNamespace(value.into()))),
                                 _ => {
                                     this.nst.put(namespace::NS_NO_PREFIX, value.clone());
-                                    this.into_state_continue(State::InsideOpeningTag(OpeningTagSubstate::InsideTag))
+                                    this.into_state_continue(State::InsideOpeningTag(OpeningTagSubstate::AfterAttributeValue))
                                 }
                             },
 
@@ -96,11 +96,18 @@ impl PullParser {
                                 name: name.clone(),
                                 value
                             });
-                            this.into_state_continue(State::InsideOpeningTag(OpeningTagSubstate::InsideTag))
+                            this.into_state_continue(State::InsideOpeningTag(OpeningTagSubstate::AfterAttributeValue))
                         }
                     }
                 }
-            })
+            }),
+
+            OpeningTagSubstate::AfterAttributeValue => match t {
+                Token::Character(c) if is_whitespace_char(c) => self.into_state_continue(State::InsideOpeningTag(OpeningTagSubstate::InsideTag)),
+                Token::TagEnd => self.emit_start_element(false),
+                Token::EmptyTagEnd => self.emit_start_element(true),
+                _ => Some(self.error(SyntaxError::UnexpectedTokenInOpeningTag(t)))
+            },
         }
     }
 }
diff --git a/tests/oasis.fail.txt b/tests/oasis.fail.txt
@@ -11,8 +11,6 @@ o-p12fail6 p12fail6.xml   built-in entity refs excluded
 o-p12fail7 p12fail7.xml   The public ID has a tab character, which is disallowed  
 o-p30fail1 p30fail1.xml   An XML declaration is not the same as a TextDecl  
 o-p31fail1 p31fail1.xml   external subset excludes doctypedecl  
-o-p40fail1 p40fail1.xml   S is required between attributes  
-o-p44fail4 p44fail4.xml   Whitespace required between attributes. 
 o-p45fail2 p45fail2.xml   S before contentspec is required.  
 o-p45fail3 p45fail3.xml   only one content spec  
 o-p45fail4 p45fail4.xml   no comments in declarations (contrast with SGML)  

diff --git a/tests/sun-not-wf.fail.txt b/tests/sun-not-wf.fail.txt
@@ -7,8 +7,6 @@ attlist06 attlist06.xml   SGML's NAME attribute type is not allowed.
 attlist07 attlist07.xml   SGML's NAMES attribute type is not allowed.
 attlist08 attlist08.xml   SGML's #CURRENT is not allowed.
 attlist09 attlist09.xml   SGML's #CONREF is not allowed.
-attlist10 attlist10.xml   Whitespace required between attributes
-attlist11 attlist11.xml   Whitespace required between attributes
 cond01 cond01.xml   Only INCLUDE and IGNORE are conditional section keywords
 cond02 cond02.xml   Must have keyword in conditional sections
 content01 content01.xml   No whitespace before "?" in content model

diff --git a/tests/xmltest.fail.txt b/tests/xmltest.fail.txt
@@ -62,7 +62,6 @@ not-wf-sa-181 181.xml   Internal parsed entities must match the content  product
 not-wf-sa-182 182.xml   Internal parsed entities must match the content  production to be well formed. 
 not-wf-sa-183 183.xml   Mixed content declarations may not include content particles.
 not-wf-sa-184 184.xml   In mixed content models, element names must not be  parenthesized. 
-not-wf-sa-186 186.xml   Whitespace is required between attribute/value pairs. 
 not-wf-not-sa-001 001.xml   Conditional sections must be properly terminated ("]>" used  instead of "]]>"). 
 not-wf-not-sa-002 002.xml   Processing instruction target names may not be "XML"  in any combination of cases. 
 not-wf-not-sa-003 003.xml   Conditional sections must be properly terminated ("]]>" omitted).