Invalid syntax error in multi-line f-string with format spec

[dhruvmanila]

Describe the bug

Basically the same bug as reported in CPython (python/cpython#110259) exists in Pyright's lexer as well.

Code or Screenshots

f"""__{
	a:d
}__"""

Running pyright on the above code gives:

$ pyright ~/playground/ruff/fstring.py
/Users/dhruv/playground/ruff/fstring.py
  /Users/dhruv/playground/ruff/fstring.py:3:2 - error: String literal is unterminated
  /Users/dhruv/playground/ruff/fstring.py:2:5 - error: "x" is not defined (reportUndefinedVariable)
2 errors, 0 warnings, 0 informations 

VS Code extension or command-line

Running through the command-line using 1.1.329

[erictraut]

I'm not able to repro the error with the code you've supplied above. Could you verify that your code sample is correct? For me, pyright emits another (correct) error indicating that a is an undefined symbol. This is in accordance with the python interpreter, which emits the same error. If I modify the code to remove that error and replace a:d with a valid expression, pyright emits no errors, nor does the Python interpreter.

a = 1
d = 1

f"""__{
	a + d
}__"""

[dhruvmanila]

Oh sorry, there are two cases here. The error thrown would be when using single quotes:

f"__{
	a:d
}__"

The problem that exists when using triple-quotes is that the format spec includes the newline character which is invalid. I'm not sure if that's relevant in Pyright.

[erictraut]

Ah, I think you meant to use single quotes rather than triple quotes.

f"__{
    x:d
}__"

Pyright's behavior matches that of the Python interpreter, so I consider it correct. I used the CPython test cases for PEP 701 to validate that pyright was working correctly here.

If the CPython devs ultimately determine that this is a bug in the Python interpreter, then I'll make the corresponding change in pyright. But for now, I consider this to be correct behavior.

[erictraut]

Oh, I just noticed that @pablogsal has already acknowledged that this is a bug in CPython and has submitted a PR to address it.

[erictraut]

This issue will be addressed in the next release of pyright.

[erictraut]

This is addressed in pyright 1.1.330, which I just published. It will also be included in an upcoming release of pylance.

[pablogsal]

Friendly reminder to review the discussion in python/cpython#110259 so you end implementing the same as CPython 😉