SA1316 should not be reported for tuple deconstruction

[Arthri]

Tuples can follow two different casing conventions depending on how it's used: tuple types may follow pascal casing, while deconstruction may follow camel casing

The rule does not differentiate between the two uses and indiscriminately flags both, so I propose having a separate rule for tuple deconstruction

// no diagnostic
(string K, string V) a = ("", "");
// SA1316
(string k, string v) = a;

// SA1316
foreach ((string k, string v) in dict)
{
}

[bjornhellander]

I am not seeing this behavior. Which version are you using?

[bjornhellander]

Have you set includeInferredTupleElementNames to true in your stylecop.json configuration file? I see the behavior that you describe when I do that.

[Arthri]

Have you set includeInferredTupleElementNames to true in your stylecop.json configuration file? I see the behavior that you describe when I do that.

I have. It seems I'm misunderstanding what includeInferredTupleElementNames means. I thought by "inferred tuple names" it means x and y in var tuple = (x, y) as opposed to explicit tuple names var tuple = (X: x, Y: y)?

[bjornhellander]

var tuple = (x, y) is equivalent to var tuple = (x: x, y: y),. Setting includeInferredTupleElementNames to true means that SA1316 will also check the implicit/inferred element names if they are omitted, like they are in var tuple = (x, y). So with the default PascalCase configuration, that code would also trigger a SA1316 warning.

[bjornhellander]

It looks to me like you have found a bug. In my opinion, tuple expressions used in the "target" of a deconstruction should not be checked by this rule. @sharwell, do you agree? I am a bit surprised that this hasn't been found before, but includeInferredTupleElementNames is unfortunately false by default, so maybe it isn't used much.

[Arthri]

I see, thanks for the clarification

[bjornhellander]

It looks like setting includeInferredTupleElementNames to false doesn't work well either:
This rule does not even trigger on var tuple = (x: 1, Y: 2), which I assume it should.

[MattFromRVA]

@bjornhellander FYI I believe you fixed the rule not triggering on var tuple = (x: 1, Y: 2) with bjornhellander/NewStyleCopAnalyzers#3.

Rule now triggers with latest version of NewStyleCopAnalyzers.

I am creating a PR to fix the other issues.