Normalizes all language tags to lower case

oxigraph · Apr 4, 2020 · 3c445e0 · 3c445e0
1 parent f131f0e
commit 3c445e0
Show file tree

Hide file tree

Showing 3 changed files with 7 additions and 4 deletions.
diff --git a/api/src/model.rs b/api/src/model.rs
@@ -58,6 +58,8 @@ impl<'a> fmt::Display for BlankNode<'a> {
 ///
 /// The default string formatter is returning a N-Triples, Turtle and SPARQL compatible representation.
 ///
+/// The language tags should be lowercased  [as suggested by the RDF specification](https://www.w3.org/TR/rdf11-concepts/#dfn-language-tagged-string).
+///
 /// ```
 /// use rio_api::model::NamedNode;
 /// use rio_api::model::Literal;
@@ -88,14 +90,14 @@ pub enum Literal<'a> {
     LanguageTaggedString {
         /// The [lexical form](https://www.w3.org/TR/rdf11-concepts/#dfn-lexical-form).
         value: &'a str,
-        /// The [language tag](https://www.w3.org/TR/rdf11-concepts/#dfn-datatype-iri).
+        /// The [language tag](https://www.w3.org/TR/rdf11-concepts/#dfn-language-tag).
         language: &'a str,
     },
     /// A literal with an explicit datatype
     Typed {
         /// The [lexical form](https://www.w3.org/TR/rdf11-concepts/#dfn-lexical-form).
         value: &'a str,
-        /// The [datatype IRI](https://www.w3.org/TR/rdf11-concepts/#dfn-language-tag).
+        /// The [datatype IRI](https://www.w3.org/TR/rdf11-concepts/#dfn-datatype-iri).
         datatype: NamedNode<'a>,
     },
 }

diff --git a/turtle/src/shared.rs b/turtle/src/shared.rs
@@ -130,7 +130,7 @@ pub fn parse_langtag(
     while let Some(c) = read.current() {
         match c {
             b'a'..=b'z' | b'A'..=b'Z' | b'0'..=b'9' | b'-' => {
-                buffer.push(char::from(c));
+                buffer.push(char::from(c).to_ascii_lowercase());
                 read.consume()?;
             }
             _ => {

diff --git a/xml/src/parser.rs b/xml/src/parser.rs
@@ -302,7 +302,8 @@ impl<R: BufRead> RdfXmlReader<R> {
                         LanguageTag::parse(
                             attribute
                                 .unescape_and_decode_value(&self.reader)
-                                .map_err(RdfXmlError::from)?,
+                                .map_err(RdfXmlError::from)?
+                                .to_ascii_lowercase(),
                         )
                         .map_err(RdfXmlError::from)?,
                     );