Regex should fail to parse "^*"

[yetanotherchris]

I have the following C# regular expression

var regex = new Regex(@"^*_.$");
string input = "THIS_LINE_HAS_MANY_UNDERSCORES_L";

string result = regex.Match(input).Groups[0].Value;
Console.WriteLine(result);

Even though to the best of my knowledge the regex is invalid (* has no element to match on), it outputs the following (which is really just the _.$ part):

_L

Is ^ treated as a character class by the .NET Regex engine?

[danmoseley]

Other engines seem to give "The preceding token is not quantifiable" which seems appropriate. The question is whether this is worth fixing (given cost/risk of doing so, and potential to break someone who would be surprised)

[Bartolomeus-649]

@danmosemsft it'a bug! fix it!

• make it a compile error
• for existing code that has a quantifier for non-quantifiable tokens, then just silently remove/ignore the quantifier (which I guess is what you already do today?).

[danmoseley]

@danmosemsft it'a bug! fix it!

We cannot fix all bugs. We allocate engineering time based on considerations like customer impact, cost of fix and so forth. This has presumably been an issue for a long time and so the impact seems low. That being said, we'd likely accept a PR for it if you're offering to make one. It would need tests and also care to make sure the compiled and interpreted paths behave the same (I'm guessing this is easy because it's at the parsing stage)

It seems to me that if other engines fail on this, and it's very unlikely anyone is relying on it, we could make it an error. @ViktorHofer as area expert, thoughts?

[CyrusNajmabadi]

If this changes, please let me know and I'll update the Roslyn side :-)

Due to lack of recent activity, this issue has been marked as a candidate for backlog cleanup. It will be closed if no further activity occurs within 14 more days. Any new comment (by anyone, not necessarily the author) will undo this process.

This process is part of our issue cleanup automation.

[danmoseley]

would we take a PR to error on this syntax? if so, seems worth keeping open for that.

even if not, perhaps we should add a test to ensure the behavior doesn't change. at least it's the same for all engines (makes sense as it's the parser I guess)

    static void Main()
    {
        string input = "^abc_d";
        Console.WriteLine(Regex.Match(input, @"^*_.$").Groups[0].Value);
        Console.WriteLine(Regex.Match(input, @"^*_.$", RegexOptions.Compiled).Groups[0].Value);
        Console.WriteLine(Regex.Match(input, @"^*_.$", RegexOptions.NonBacktracking).Groups[0].Value);
        Console.WriteLine(GeneratedRegex().Match(input).Groups[0].Value);
    }

    [GeneratedRegex(@"^*_.$")]
    private static partial Regex GeneratedRegex();

_d
_d
_d
_d