Fix invalid printer IR error (#5422)

This commit is contained in:
Micha Reiser 2023-06-29 08:09:13 +02:00 committed by GitHub
parent ca5e10b5ea
commit 38189ed913
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 70 additions and 18 deletions

View file

@ -108,3 +108,13 @@ test_particular = [
'1.0000000000000000000000000000000000000000000010000' #...
'0000000000000000000000000000000000000000025',
]
# Parenthesized string continuation with messed up indentation
{
"key": (
[],
'a'
'b'
'c'
)
}

View file

@ -242,6 +242,7 @@ impl<'fmt, 'ast, 'buf> JoinCommaSeparatedBuilder<'fmt, 'ast, 'buf> {
}
pub(crate) fn finish(&mut self) -> FormatResult<()> {
self.result.and_then(|_| {
if let Some(last_end) = self.last_end.take() {
if_group_breaks(&text(",")).fmt(self.fmt)?;
@ -259,6 +260,7 @@ impl<'fmt, 'ast, 'buf> JoinCommaSeparatedBuilder<'fmt, 'ast, 'buf> {
}
Ok(())
})
}
}

View file

@ -8,7 +8,7 @@ use ruff_formatter::{format_args, write, FormatError};
use ruff_python_ast::str::is_implicit_concatenation;
use ruff_text_size::{TextLen, TextRange, TextSize};
use rustpython_parser::ast::{ExprConstant, Ranged};
use rustpython_parser::lexer::lex_starts_at;
use rustpython_parser::lexer::{lex_starts_at, LexicalError, LexicalErrorType};
use rustpython_parser::{Mode, Tok};
use std::borrow::Cow;
@ -83,7 +83,7 @@ impl Format<PyFormatContext<'_>> for FormatStringContinuation<'_> {
// Call into the lexer to extract the individual chunks and format each string on its own.
// This code does not yet implement the automatic joining of strings that fit on the same line
// because this is a black preview style.
let lexer = lex_starts_at(string_content, Mode::Module, string_range.start());
let lexer = lex_starts_at(string_content, Mode::Expression, string_range.start());
let separator = format_with(|f| match self.layout {
StringLayout::Default(_) => soft_line_break_or_space().fmt(f),
@ -93,7 +93,31 @@ impl Format<PyFormatContext<'_>> for FormatStringContinuation<'_> {
let mut joiner = f.join_with(separator);
for token in lexer {
let (token, token_range) = token.map_err(|_| FormatError::SyntaxError)?;
let (token, token_range) = match token {
Ok(spanned) => spanned,
Err(LexicalError {
error: LexicalErrorType::IndentationError,
..
}) => {
// This can happen if the string continuation appears anywhere inside of a parenthesized expression
// because the lexer doesn't know about the parentheses. For example, the following snipped triggers an Indentation error
// ```python
// {
// "key": (
// [],
// 'a'
// 'b'
// 'c'
// )
// }
// ```
// Ignoring the error here is *safe* because we know that the program once parsed to a valid AST.
continue;
}
Err(_) => {
return Err(FormatError::SyntaxError);
}
};
match token {
Tok::String { .. } => {

View file

@ -114,6 +114,16 @@ test_particular = [
'1.0000000000000000000000000000000000000000000010000' #...
'0000000000000000000000000000000000000000025',
]
# Parenthesized string continuation with messed up indentation
{
"key": (
[],
'a'
'b'
'c'
)
}
```
## Outputs
@ -259,6 +269,9 @@ test_particular = [
"1.0000000000000000000000000000000000000000000010000" # ...
"0000000000000000000000000000000000000000025",
]
# Parenthesized string continuation with messed up indentation
{"key": ([], "a" "b" "c")}
```
@ -404,6 +417,9 @@ test_particular = [
'1.0000000000000000000000000000000000000000000010000' # ...
'0000000000000000000000000000000000000000025',
]
# Parenthesized string continuation with messed up indentation
{'key': ([], 'a' 'b' 'c')}
```