I am using CSharpSyntaxTree.ParseText(text); to parse some code to format the code and then resave it.

I encountered a very peculiar bug, where in some very rare cases, CSharpSyntaxTree.ParseText(text).GetCompilationUnitRoot() will cause one extra parentheses to appear where it should not.

This project is the minimal reproducable case I was able to create.

Running this project as is will take the contents of Code.txt parse them to a SyntaxTree, and immediately saving the contents as text into CodeBroken.cs

This will cause an extra parentesis to appear in line CodeBroken.cs:234 SecondaryStatByRarityFrom: Create((.03, .05, .07, .1, .16),

If you were to simply change some contents of Code.txt, such as removing the comment at the top of the file, removing some logical unrelated part or simply re-ordering the contents, will cause it to write the code correctly.SecondaryStatByRarityFrom: Create(.03, .05, .07, .1, .16),

This happens when I updated the Microsoft.CodeAnalysis.CSharp from 4.12.0 to 4.13.0, This does not happen and works as expected when I downgrade back to 4.12.0