During fstring parsing, we unleash the tokenizer on non-Python code, because we start tokenizing embedded expressions, but pass the remaining fstring content to the tokenizer. In this situation the tokenizer can throw SyntaxErrors (IndentationErrors in particular). This happens in the non-Python part, we can simply ignore these errors.
soupytwist
commented
2 years ago
During fstring parsing, we unleash the tokenizer on non-Python code, because we start tokenizing embedded expressions, but pass the remaining fstring content to the tokenizer. In this situation the tokenizer can throw SyntaxErrors (IndentationErrors in particular). This happens in the non-Python part, we can simply ignore these errors.