Wrong fix for W605 in f-strings

[FreeHarry]

Python: 3.12.3
Version: ruff 0.4.4
Command: ruff check --select W605 .\test.py --fix

And let's given the following code snippet in test.py:

total = 10
ok = 7
incomplete = 3
s = f"TOTAL: {total}\nOK: {ok}\INCOMPLETE: {incomplete}\n"

It is wrongly fixed as:

total = 10
ok = 7
incomplete = 3
s = rf"TOTAL: {total}\nOK: {ok}\INCOMPLETE: {incomplete}\n"

Now the string is a raw string. The newlines will then interpreted as literals and will not remain newlines.

---

If I do the same for a normal string:

s = "TOTAL: 10\nOK: 3\INCOMPLETE: 7\n"

It is correctly fixed as:

s = "TOTAL: 10\nOK: 3\\INCOMPLETE: 7\n"

The invalid escape sequence '\I' is correctly fixed as \\, and the newlines are still there.

---

Even if it is a f-string, but without placeholders, it is done correctly:

s = f"TOTAL: 10\nOK: 3\INCOMPLETE: 7\n"

Result is correct:

s = f"TOTAL: 10\nOK: 3\\INCOMPLETE: 7\n"

[dhruvmanila]

Thank you for the report!

This is happening because each f-string element is analyzed individually. So, first TOTAL: is analyzed, then \nOK: and then \INCOMPLETE: , but the checker isn't aware that one of the element contains a newline character.

[dylwil3]

Update: The issue in the original post has been resolved, but I'm leaving this open because invalid escapes in format specifiers are not handled yet (see #14748 (comment) and the surrounding discussion).