Pep701 fstrings

[jpe]

These changes add support for PEP 701 f-strings by bypassing the logic that would terminate an f-string if the starting quote was seen. The pre-PEP 701 behavior is kept as an option, though I don't have any plans to use the pre-PEP 701 and would be +0 on removing the old behavior.

[zufuliu]

How about merge the for and while loop as:

    if (!options.pep701StringsF) {
        // Find the deepest single quote state because that string will end; no \ continuation in f-string
        for (size_t i = 0; i < fstringStateStack.size(); i++) {
            if (IsPySingleQuoteStringState(fstringStateStack[i].state)) {
                sc.SetState(fstringStateStack[i].state);
                while (fstringStateStack.size() > i) {
                    PopFromStateStack(fstringStateStack, currentFStringExp);
                }
                break;
            }
        }
    }

this would avoid some lint warnings:

• long is always 32 bit on Windows, use size_t avoided truncate warning and extra cast.
• the scope of i can be reduced, and deepestSingleStateIndex can be removed (or can be changed to size_t and npos).

[nyamatongwe]

Committed da28b2c which is similar to 5f81ddb but with stringsF corrected to pep701StringsF in DefineProperty and with declarations of deepestSingleStateIndex and i moved inside if to avoid warnings.

Did not (yet) commit type and loop change as it isn't trivial and I didn't find a good test case. It seems that the stack popping only occurs when there are erroneous outstanding quoted strings being forcibly ended.

[zufuliu]

The type and loop changes looks non-functionality change for me:

• i is only used inside the for loop.
• if (deepestSingleStateIndex != -1) is only true when the for loop breaks.