Improve raw string handling in static_join_to_fstring

Refined logic to better detect when non-raw representation is needed, especially for cases involving backslashes and quote delimiters. Updated test fixtures and snapshots to clarify behavior and expected fixes for edge cases mixing raw and non-raw strings.
This commit is contained in:
Dan 2025-10-28 21:00:19 -04:00
parent d371dde2bf
commit b6caf8c92b
3 changed files with 32 additions and 26 deletions

View File

@ -36,7 +36,7 @@ nok14 = "".join((r"", '"')) # First is raw, second is not - would break syntax
nok15 = "".join((r"", "\\")) # First is raw, second has backslash - would break syntax nok15 = "".join((r"", "\\")) # First is raw, second has backslash - would break syntax
nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
# Test that all-raw strings still work (should be OK) # Test that all-raw strings still work (should be OK)
ok7 = "".join((r"", r"something")) # Both are raw - OK ok7 = "".join((r"", r"something")) # Both are raw - OK

View File

@ -109,10 +109,14 @@ fn build_fstring(joiner: &str, joinees: &[Expr], flags: FStringFlags) -> Option<
// use non-raw representation. This handles cases where raw strings would create invalid // use non-raw representation. This handles cases where raw strings would create invalid
// syntax or behavior changes. // syntax or behavior changes.
if any_raw && !content.is_empty() { if any_raw && !content.is_empty() {
let needs_non_raw = content.contains(['\n', '\r', '\0']) || { let needs_non_raw = content.contains(['\r', '\0']) || {
// Check if content contains characters that would break raw string syntax // Check if content contains characters that would break raw string syntax
let quote_char = flags.quote_str(); let quote_char = flags.quote_str();
content.contains(quote_char) || content.contains('\\') // A raw string cannot end with a single backslash if it's immediately
// followed by the quote delimiter, as that would be invalid syntax.
let ends_with_backslash = content.ends_with('\\')
&& (content.len() == 1 || content.chars().nth_back(1) != Some('\\'));
content.contains(quote_char) || ends_with_backslash
}; };
if needs_non_raw { if needs_non_raw {

View File

@ -125,7 +125,7 @@ help: Replace with `f"{secrets.token_urlsafe()}a{secrets.token_hex()}"`
13 | nok2 = a.join(["1", "2", "3"]) # Not OK (not a static joiner) 13 | nok2 = a.join(["1", "2", "3"]) # Not OK (not a static joiner)
note: This is an unsafe fix and may change runtime behavior note: This is an unsafe fix and may change runtime behavior
FLY002 [*] Consider `'line1\nline2'` instead of string join FLY002 [*] Consider f-string instead of string join
--> FLY002.py:20:8 --> FLY002.py:20:8
| |
18 | nok7 = "a".join([f"foo{8}", "bar"]) # Not OK (contains an f-string) 18 | nok7 = "a".join([f"foo{8}", "bar"]) # Not OK (contains an f-string)
@ -134,15 +134,16 @@ FLY002 [*] Consider `'line1\nline2'` instead of string join
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
21 | nok9 = '\n'.join([r"raw string", '<""">', "<'''>"]) # Not OK (both triple-quote delimiters appear; should bail) 21 | nok9 = '\n'.join([r"raw string", '<""">', "<'''>"]) # Not OK (both triple-quote delimiters appear; should bail)
| |
help: Replace with `'line1\nline2'` help: Replace with f-string
17 | nok6 = "a".join(x for x in "feefoofum") # Not OK (generator) 17 | nok6 = "a".join(x for x in "feefoofum") # Not OK (generator)
18 | nok7 = "a".join([f"foo{8}", "bar"]) # Not OK (contains an f-string) 18 | nok7 = "a".join([f"foo{8}", "bar"]) # Not OK (contains an f-string)
19 | # https://github.com/astral-sh/ruff/issues/19887 19 | # https://github.com/astral-sh/ruff/issues/19887
- nok8 = '\n'.join([r'line1','line2']) - nok8 = '\n'.join([r'line1','line2'])
20 + nok8 = 'line1\nline2' 20 + nok8 = r'''line1
21 | nok9 = '\n'.join([r"raw string", '<""">', "<'''>"]) # Not OK (both triple-quote delimiters appear; should bail) 21 + line2'''
22 | 22 | nok9 = '\n'.join([r"raw string", '<""">', "<'''>"]) # Not OK (both triple-quote delimiters appear; should bail)
23 | # Regression test for: https://github.com/astral-sh/ruff/issues/7197 23 |
24 | # Regression test for: https://github.com/astral-sh/ruff/issues/7197
note: This is an unsafe fix and may change runtime behavior note: This is an unsafe fix and may change runtime behavior
FLY002 [*] Consider `"raw string\n<\"\"\">\n<'''>"` instead of string join FLY002 [*] Consider `"raw string\n<\"\"\">\n<'''>"` instead of string join
@ -267,7 +268,7 @@ help: Replace with `"\\"`
36 + nok15 = "\\" # First is raw, second has backslash - would break syntax 36 + nok15 = "\\" # First is raw, second has backslash - would break syntax
37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes 37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior 38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r 39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
note: This is an unsafe fix and may change runtime behavior note: This is an unsafe fix and may change runtime behavior
FLY002 [*] Consider `"\x00"` instead of string join FLY002 [*] Consider `"\x00"` instead of string join
@ -278,7 +279,7 @@ FLY002 [*] Consider `"\x00"` instead of string join
37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes 37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
| ^^^^^^^^^^^^^^^^^^^^ | ^^^^^^^^^^^^^^^^^^^^
38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior 38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r 39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
| |
help: Replace with `"\x00"` help: Replace with `"\x00"`
34 | # Mixing raw and non-raw strings can cause syntax errors or behavior changes 34 | # Mixing raw and non-raw strings can cause syntax errors or behavior changes
@ -287,7 +288,7 @@ help: Replace with `"\x00"`
- nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes - nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
37 + nok16 = "\x00" # First is raw, second has null byte - would introduce null bytes 37 + nok16 = "\x00" # First is raw, second has null byte - would introduce null bytes
38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior 38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r 39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
40 | 40 |
note: This is an unsafe fix and may change runtime behavior note: This is an unsafe fix and may change runtime behavior
@ -298,7 +299,7 @@ FLY002 [*] Consider `"\r"` instead of string join
37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes 37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior 38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
| ^^^^^^^^^^^^^^^^^^^^ | ^^^^^^^^^^^^^^^^^^^^
39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r 39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
| |
help: Replace with `"\r"` help: Replace with `"\r"`
35 | nok14 = "".join((r"", '"')) # First is raw, second is not - would break syntax 35 | nok14 = "".join((r"", '"')) # First is raw, second is not - would break syntax
@ -306,27 +307,27 @@ help: Replace with `"\r"`
37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes 37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
- nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior - nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
38 + nok17 = "\r" # First is raw, second has carriage return - would change behavior 38 + nok17 = "\r" # First is raw, second has carriage return - would change behavior
39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r 39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
40 | 40 |
41 | # Test that all-raw strings still work (should be OK) 41 | # Test that all-raw strings still work (should be OK)
note: This is an unsafe fix and may change runtime behavior note: This is an unsafe fix and may change runtime behavior
FLY002 [*] Consider `"\\r"` instead of string join FLY002 [*] Consider `r"\r"` instead of string join
--> FLY002.py:39:9 --> FLY002.py:39:9
| |
37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes 37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior 38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r 39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
| ^^^^^^^^^^^^^^^^^^^^^ | ^^^^^^^^^^^^^^^^^^^^^
40 | 40 |
41 | # Test that all-raw strings still work (should be OK) 41 | # Test that all-raw strings still work (should be OK)
| |
help: Replace with `"\\r"` help: Replace with `r"\r"`
36 | nok15 = "".join((r"", "\\")) # First is raw, second has backslash - would break syntax 36 | nok15 = "".join((r"", "\\")) # First is raw, second has backslash - would break syntax
37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes 37 | nok16 = "".join((r"", "\0")) # First is raw, second has null byte - would introduce null bytes
38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior 38 | nok17 = "".join((r"", "\r")) # First is raw, second has carriage return - would change behavior
- nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r - nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
39 + nok18 = "\\r" # First is raw, second has backslash escape followed by r 39 + nok18 = r"\r" # First is raw, second has backslash followed by literal r - OK (no special handling needed)
40 | 40 |
41 | # Test that all-raw strings still work (should be OK) 41 | # Test that all-raw strings still work (should be OK)
42 | ok7 = "".join((r"", r"something")) # Both are raw - OK 42 | ok7 = "".join((r"", r"something")) # Both are raw - OK
@ -341,7 +342,7 @@ FLY002 [*] Consider `r"something"` instead of string join
43 | ok8 = "\n".join((r"line1", r'line2')) # Both are raw - OK 43 | ok8 = "\n".join((r"line1", r'line2')) # Both are raw - OK
| |
help: Replace with `r"something"` help: Replace with `r"something"`
39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash escape followed by r 39 | nok18 = "".join((r"", "\\r")) # First is raw, second has backslash followed by literal r - OK (no special handling needed)
40 | 40 |
41 | # Test that all-raw strings still work (should be OK) 41 | # Test that all-raw strings still work (should be OK)
- ok7 = "".join((r"", r"something")) # Both are raw - OK - ok7 = "".join((r"", r"something")) # Both are raw - OK
@ -351,7 +352,7 @@ help: Replace with `r"something"`
45 | # Test that all-non-raw strings still work (should be OK) 45 | # Test that all-non-raw strings still work (should be OK)
note: This is an unsafe fix and may change runtime behavior note: This is an unsafe fix and may change runtime behavior
FLY002 [*] Consider `"line1\nline2"` instead of string join FLY002 [*] Consider f-string instead of string join
--> FLY002.py:43:7 --> FLY002.py:43:7
| |
41 | # Test that all-raw strings still work (should be OK) 41 | # Test that all-raw strings still work (should be OK)
@ -361,15 +362,16 @@ FLY002 [*] Consider `"line1\nline2"` instead of string join
44 | 44 |
45 | # Test that all-non-raw strings still work (should be OK) 45 | # Test that all-non-raw strings still work (should be OK)
| |
help: Replace with `"line1\nline2"` help: Replace with f-string
40 | 40 |
41 | # Test that all-raw strings still work (should be OK) 41 | # Test that all-raw strings still work (should be OK)
42 | ok7 = "".join((r"", r"something")) # Both are raw - OK 42 | ok7 = "".join((r"", r"something")) # Both are raw - OK
- ok8 = "\n".join((r"line1", r'line2')) # Both are raw - OK - ok8 = "\n".join((r"line1", r'line2')) # Both are raw - OK
43 + ok8 = "line1\nline2" # Both are raw - OK 43 + ok8 = r"""line1
44 | 44 + line2""" # Both are raw - OK
45 | # Test that all-non-raw strings still work (should be OK) 45 |
46 | ok9 = "".join(("", '"')) # Both are non-raw - OK 46 | # Test that all-non-raw strings still work (should be OK)
47 | ok9 = "".join(("", '"')) # Both are non-raw - OK
note: This is an unsafe fix and may change runtime behavior note: This is an unsafe fix and may change runtime behavior
FLY002 [*] Consider `'"'` instead of string join FLY002 [*] Consider `'"'` instead of string join