Merge pull request #207 from DavidVanDeursen/fix-external-url-path-regex

Fix external url path regex
spreadsheetlab · Jan 7, 2025 · c5299ad · c5299ad
2 parents 21b12fb + 02733c1
commit c5299ad
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 3 deletions.
diff --git a/src/XLParser.Tests/FormulaAnalysisTest.cs b/src/XLParser.Tests/FormulaAnalysisTest.cs
@@ -636,14 +636,38 @@ public void ExternalWorkbookUrlPathHttpWithSpaceInDocument()
             Assert.AreEqual("Sheet1", references.First().Worksheet);
         }
 
+        [TestMethod]
+        public void ExternalWorkbookUrlPathHttpWithTildeInPath()
+        {
+            // See [#202](https://github.com/spreadsheetlab/XLParser/issues/202)
+            List<ParserReference> references = new FormulaAnalyzer(@"='http://example.com/~testfolder/[Book1.xlsx]Sheet1'!$A$1").ParserReferences().ToList();
+
+            Assert.AreEqual(1, references.Count);
+            Assert.AreEqual(@"http://example.com/~testfolder/", references.First().FilePath);
+            Assert.AreEqual("Book1.xlsx", references.First().FileName);
+            Assert.AreEqual("Sheet1", references.First().Worksheet);
+        }
+
+        [TestMethod]
+        public void ExternalWorkbookUrlPathHttpWithBackslashInPath()
+        {
+            // See [#201](https://github.com/spreadsheetlab/XLParser/issues/201)
+            List<ParserReference> references = new FormulaAnalyzer(@"='http:\\example.com\testfolder\[Book1.xlsx]Sheet1'!$A$1").ParserReferences().ToList();
+
+            Assert.AreEqual(1, references.Count);
+            Assert.AreEqual(@"http:\\example.com\testfolder\", references.First().FilePath);
+            Assert.AreEqual("Book1.xlsx", references.First().FileName);
+            Assert.AreEqual("Sheet1", references.First().Worksheet);
+        }
+
         [TestMethod]
         public void ExternalWorkbookUrlPathHttpWithPortNumberInPath()
         {
             // See [#194](https://github.com/spreadsheetlab/XLParser/issues/194)
-            List<ParserReference> references = new FormulaAnalyzer(@"='http://example.com:1234/test/[Book1.xlsx]Sheet1'!$A$1").ParserReferences().ToList();
+            List<ParserReference> references = new FormulaAnalyzer(@"='http://localhost:1234/test/[Book1.xlsx]Sheet1'!$A$1").ParserReferences().ToList();
 
             Assert.AreEqual(1, references.Count);
-            Assert.AreEqual(@"http://example.com:1234/test/", references.First().FilePath);
+            Assert.AreEqual(@"http://localhost:1234/test/", references.First().FilePath);
             Assert.AreEqual("Book1.xlsx", references.First().FileName);
             Assert.AreEqual("Sheet1", references.First().Worksheet);
         }

diff --git a/src/XLParser/ExcelFormulaGrammar.cs b/src/XLParser/ExcelFormulaGrammar.cs
@@ -205,7 +205,7 @@ public class ExcelFormulaGrammar : Grammar
 
         // Source: http://stackoverflow.com/a/6416209/572635
         private const string windowsFilePathRegex = @"(?:[a-zA-Z]:|\\?\\?[\w\-.$ @~]+)\\(([^<>\"" /\|?*\\']|( |''))*\\)*";
-        private const string urlPathRegex = @"http(s?)\://([\p{L}\p{N}-_]+\.[\p{L}\p{N}-_]*)+(:[0-9]+)?/([\p{L}\p{N}\-\.\?\,\'+&%\$#_ ()]*/)*";
+        private const string urlPathRegex = @"https?\:(//|\\\\)[\p{L}\p{N}\-_.]+(:[0-9]+)?(/|\\)([\p{L}\p{N}\-_.?,'+&%\$# ()~]*(/|\\))*";
         private const string filePathRegex = @"(" + windowsFilePathRegex + @"|" + urlPathRegex + @")";
         public Terminal FilePathToken { get; } = new RegexBasedTerminal(GrammarNames.TokenFilePath, filePathRegex)
         { Priority = TerminalPriority.FileNamePath };