Invalid hyperlink throws exception - Invalid URI: The hostname could not be parsed.

[patrickwzl]

If the content of an Excel contains hyperlink, and the hyperlink is invalid such as "mailto:a@a,com", it will throw the exception "Invalid URI: The hostname could not be parsed.".

[tonyqus]

Can you show me the call stack?

[patrickwzl]

Just the following code can trigger the problem.

string path = "Some problem Excel files";
FileStream fs = new FileStream(path, FileMode.Open, FileAccess.Read, FileShare.ReadWrite);
book = new XSSFWorkbook(fs);

Exception stack:
[UriFormatException: Invalid URI: The hostname could not be parsed.]
System.Uri.CreateHostStringHelper(String str, UInt16 idx, UInt16 end, Flags& flags, String& scopeId) +8491635
System.Uri.CreateHostString() +251
System.Uri.EnsureHostString(Boolean allowDnsOptimization) +64
System.Uri.GetComponentsHelper(UriComponents uriComponents, UriFormat uriFormat) +89
System.Uri.ToString() +96
NPOI.POIXMLFactory.GetPackageRelationship(POIXMLDocumentPart parent, PackagePart part) in D:\github\npoi\ooxml\POIXMLFactory.cs:139
NPOI.POIXMLFactory.CreateDocumentPart(POIXMLDocumentPart parent, PackagePart part) in D:\github\npoi\ooxml\POIXMLFactory.cs:51
NPOI.POIXMLDocumentPart.Read(POIXMLFactory factory, Dictionary``2 context) in D:\github\npoi\ooxml\POIXMLDocumentPart.cs:720
NPOI.POIXMLDocumentPart.Read(POIXMLFactory factory, Dictionary``2 context) in D:\github\npoi\ooxml\POIXMLDocumentPart.cs:731
NPOI.POIXMLDocument.Load(POIXMLFactory factory) in D:\github\npoi\ooxml\POIXMLDocument.cs:164
NPOI.XSSF.UserModel.XSSFWorkbook..ctor(Stream is1) in D:\github\npoi\ooxml\XSSF\UserModel\XSSFWorkbook.cs:241

[tonyqus]

Can you upload the file? I'd like to look into the file. Something wrong with the package relationship uri.

[tonyqus]

This is where the exception comes from.

npoi/ooxml/POIXMLFactory.cs

Line 141 in b5b21c5

String packName = pr.TargetUri.ToString();// toASCIIString();

[patrickwzl]

test.xlsx

[tonyqus]

I cannot reproduce the issue with the latest NPOI code. Which NPOI version are you using?

[patrickwzl]

The latest NPOI version 2.5.6 is used.
I have tried different scenarios and found the problem.

The problem comes from the behaviour changes in .NET framework.
Uri uri = new Uri("mailto:a@a,com"); <-- Throws exception when .NET framework 4.5 or above.

Uri uri = new Uri("mailto:a@a,com");
uri.ToString(); <-- Throws exception before .NET framework 4.5, no probelm on above line.

NPOI has handled the case when Uri uri = new Uri("mailto:a@a,com"); throws exception, so the program does not have problem.
If there is no exception in Uri uri = new Uri("mailto:a@a,com");, then NPOI will throw exception when calling uri.ToString();.

For ASP.NET, even though the system is running latest .NET framework version (e.g. v4.8). If the web.config does not specify the version explicitly <httpRuntime targetFramework="4.x.x" />, the "new Uri" is still using the behaviour like before .NET framework version 4.5.

[tonyqus]

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">
  <Relationship Id="rId3" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/extended-properties" Target="docProps/app.xml"/>
  <Relationship Id="rId2" Type="http://schemas.openxmlformats.org/package/2006/relationships/metadata/core-properties" Target="docProps/core.xml"/>
  <Relationship Id="rId1" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument" Target="xl/workbook.xml"/>
</Relationships>

.rels xml example from your uploaded file - test.xlsx

The target Uri of package looks like docProps/app.xml, xl/workbook.xml, docProps/core.xml. There is no comma in it. Are you saying Uri.ToString throws exception with these kind of relative url in .NET framework before 4.5? Or the exception is from another place related to Uri in NPOI code? If so, can you show me the call stack?

[patrickwzl]

I have created a sample ASP.NET project with NPOI 2.5.6.
testNPOI.zip

Please try to comment and uncomment the line <httpRuntime targetFramework="4.8" /> in web.config.

For newly created ASP.NET projects, this config is added by default in web.config now, but it was not added some years ago by default. Hence, newly created project may not face this problem.

I have simple fix so that the NPOI can work in all situations.
XSSFHyperlink.cs (Line 85)

                Uri target = _externalRel.TargetUri;
                try
                {
                    _location = target.ToString();
                } catch
                {
                    _location = target.OriginalString.Trim().ToLower();
                }

[tonyqus]

Ok. I can reproduce the exception. Nice discovery. Looks Uri.OriginalString is much safer than Uri.ToString() in versions before .NET framework 4.5