dotnet test html logger throws execption when using special characters in DataRow attributes.

[nohwnd]

Cannot move there, see: dotnet/sdk#38983

[daveMueller]

Is it OK when I start working on that?

[daveMueller]

I looked a bit into it and think this can only be solved by preprocessing the xml test result before transforming it into html. Like already mentioned in this comment here
https://github.com/microsoft/vstest/blob/main/src/Microsoft.TestPlatform.Extensions.HtmlLogger/HtmlTransformer.cs#L33-L35,
the test result can contain characters that aren't valid xml. Thus the html writer must also be tolerant of those characters which he isn't anymore after the change in this PR #4756.
But it also isn't an option to use own XmlWriterSettings again because this will bring back the staircase issue #4754.

[nohwnd]

I was able to reproduce this issue and have submitted a fix.

Root cause: XML 1.0 forbids control characters in the range \x00–\x08, \x0B, \x0C, \x0E–\x1F. When a [DataRow] attribute contains such characters, DataContractSerializer.WriteObject() throws XmlException: '', hexadecimal value 0x01, is an invalid character — silently suppressing the HTML report.

Fix: Added a RemoveInvalidXmlChars() helper to HtmlLogger (mirroring the pattern already used in TrxLogger's XmlPersistence) that replaces invalid characters with their \uXXXX Unicode escape representation before they are stored in the object model.

A PR has been submitted with the fix and a unit test covering this scenario.

🔍 Triaged by Issue Repro Triage & Auto-Fix 🔍