Fix issue with Iri rule not handling URLs without trailing slash

fastobo · Sep 28, 2020 · 1042068 · 1042068
1 parent 4bd44dc
commit 1042068
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 2 deletions.
diff --git a/src/grammar.pest b/src/grammar.pest
@@ -394,8 +394,11 @@ ResourcePropertyValue = { RelationId ~ Id }
 
 Iri = @{ IriScheme ~ ":" ~ IriHierPart ~ ("?" ~ IriQuery)? ~ ("#" ~ IriFragment)? }
 
+// NB(@althonos): IriPathAbEmpty is not mandatory if we reach the end of input
+//                here to allow URLs to hosts without trailing slashes, e.g.
+///               'http://example.com', which would be rejected otherwise.
 IriHierPart = {
-    ("//" ~ IriAuthority ~ IriPathAbempty )
+    ("//" ~ IriAuthority ~ (IriPathAbempty | EOI) )
   | IriPathAbsolute
   | IriPathRootless
   | IriPathEmpty
@@ -411,7 +414,7 @@ IriPathAbempty  = ${ ("/" ~ IriSegment)+ }
 IriPathAbsolute = ${ "/" ~ (IriSegmentNz ~ ("/" ~ IriSegment)* )? }
 IriPathNoScheme = ${ IriSegmentNzNc ~ ("/" ~ IriSegment)* }
 IriPathRootless = ${ IriSegmentNz ~ ("/" ~ IriSegment)* }
-IriPathEmpty    = ${ "0" ~ IriIpChar}
+IriPathEmpty    = ${ "" }
 
 IriSegment     = @{ IriIpChar* }
 IriSegmentNz   = @{ IriIpChar+ }

diff --git a/tests/parser.rs b/tests/parser.rs
@@ -87,6 +87,9 @@ fn term_frame() {
     test_parse!(TermFrame, "[Term]\n id: TEST:001\ndef: \"test item\" []\n");
     test_parse!(TermFrame, "[Term]\nid: TEST:001\n def: \"test item\" []\n");
     test_parse!(TermFrame, " [Term]\nid: TEST:001\ndef: \"test item\" []\n");
+    // url id
+    test_parse!(TermFrame, "[Term]\nid: http://example.com\ndef: \"test item\" []\n");
+    test_parse!(TermFrame, "[Term]\nid: http://example.com/\ndef: \"test item\" []\n");
 }
 
 #[test]
@@ -134,3 +137,23 @@ fn property_value() {
     test_parse!(ResourcePropertyValue, "foaf:depiction http://purl.obolibrary.org/obo/plana/images/Stage2.png");
     test_not_parse!(LiteralPropertyValue, "foaf:depiction http://purl.obolibrary.org/obo/plana/images/Stage2.png");
 }
+
+#[test]
+fn iri() {
+    test_parse!(Iri, "http://example.com");
+    test_parse!(Iri, "http://example.com/");
+    test_parse!(Iri, "http://example.com/with/a/path");
+    test_parse!(Iri, "http://user:pass@example.com");
+    test_parse!(Iri, "http://user:pass@example.com/");
+    test_parse!(Iri, "http://user:pass@example.com/with/a/path");
+}
+
+#[test]
+fn term_clause() {
+    test_parse!(TermClause, "xref: http://example.com");
+    test_parse!(TermClause, "xref: http://example.com/");
+    test_parse!(TermClause, "xref: http://example.com/with/a/path");
+    test_parse!(TermClause, "xref: http://user:pass@example.com");
+    test_parse!(TermClause, "xref: http://user:pass@example.com/");
+    test_parse!(TermClause, "xref: http://user:pass@example.com/with/a/path");
+}