Optimize UnescapeText method using StringBuilder

[y114git]

Refactor UnescapeText method to use StringBuilder for optimization.

Description

This refactors UndertaleString.UnescapeText to parse escape sequences in a single pass with StringBuilder instead of applying multiple chained string.Replace calls.

The previous implementation could re-process characters produced by an earlier replacement. For example, an escaped literal backslash followed by n ("\\n" in assembler text) could be transformed into a real newline because Replace("\\n", "\n") ran before Replace("\\\\", "\\").

The new implementation handles only the escape sequences emitted by UndertaleString.ToString(true):

• \r -> carriage return
• \n -> newline
• \" -> double quote
• \\ -> literal backslash

Unknown escape-like sequences are preserved as-is instead of dropping the backslash. This keeps strings such as Windows paths (C:\Users\Name) or unsupported escape-like text (\t, \x, \0, \U) from being silently changed.

Caveats

This intentionally does not add support for additional GML/C-style escape sequences such as \t, \xNN, octal escapes, or Unicode escapes. The method is used by the VM assembler text path and is kept as the inverse of UndertaleString.ToString(true), rather than a full GML string literal parser.

Notes

Needs tests ig