@isidentical @lysnikolaou

I think we should remove the comment and keep the newline.

I expected the example to produce "SyntaxError: unterminated f-string literal (detected at line 1)", just as it would without the '{'. But given that '{' escapes the newline, I also expect it to escape the octothorpe, just as it does now. So I consider the current behavior correct.

@terryjreedy not sure I follow your reasoning, what do you mean when you say that "given that '{' escapes the newline"? There is nothing here being scaped no? Inside an f-string expression now any expression can be made as if it written between parentheses. The key of this issue is the debug expression here

I was using a different analogy for my expectation. 'a = #com\n3' is a SyntaxError. '''a= #com\n3''' is not, and the comment is keep. Your comment suggests this instead: (a * #com by itself is a Syntaxerror; (a * #com\n3 is not, because the parentheses cause the comment and newline to be ignored.

Even better, I think, are these: f'{a #com\n}' ignores the comment and its terminating newline and prints 1. The a = construct is unusual because 'a =' is quoted while 'a' is executed. What happens with text between '=' and '}'? f'{a=x}' raises SyntaxError: f-string: expecting '!', or ':', or '}'. These examples convince me that a comment after '=' should be recognized as such and ignored rather that quoted as additional text.

As for newlines: whitespace before and after = is kept, whether or not #whatever is present, and newlines can be escaped with \. Leave this alone.

If the expression component of an f-string requires comments, you're probably doing something horribly wrong. That being said, keeping the comment in the resulting string is likely to be the least surprising behavior. Either that or forbidding comments inside f-strings altogether.

I also think that the comment should be ignored, while the newline should be kept.