Allow switch statement type guards (need help..)

[bsteephenson]

Hi. This PR resolves issue #2214 allowing the following code to work

function logNumber(v: number) { console.log("number:", v); }
function logString(v: string) { console.log("string:", v); }

function foo1(v: number|string) {
    switch (typeof v) {
        case 'number':
            logNumber(v);
            break;
        case 'string':
            logString(v);
            break;
        default:
            throw new Error("unsupported type");
    }
}

I need some help (I'm kinda new to this..)

1. TypeOfExpression doesn't have a member "text" but the Javascript equivalent does. What can I cast TypeOfExpression to to get access to "text"?
2. I'm not sure where I should be writing tests.. (Or what a .types file is)
3. This might be too permissive. In the previous example, if I were to have a case 'Watermelon' it would still work. Should I check to see that the type is a subset of v's type?

Thank you!

[msftclas]

Hi @bsteephenson, I'm your friendly neighborhood Microsoft Pull Request Bot (You can call me MSBOT). Thanks for your contribution!

This seems like a small (but important) contribution, so no Contribution License Agreement is required at this point. Real humans will now evaluate your PR.

TTYL, MSBOT;

[yuit]

Regarding test in our repository, all tests are located inside tests/cases. For Typeguard feature, there is a folder in tests/cases/conformance/expressions/typeGuards. You can certainly put your tests in there. I would try to be explicit with naming the test file. By default test file is emitted below ES6 if you would like the test to be emitted as es6 including "// @target: es6" at the first line of the file. and include "ES6" as part of the file name.

The test file is simply .ts file. When the test runner runs, it generates corresponding .js (plain javascript file), .type (this file only get generated if the ts file is compiled with no error. '.type' file specifies a type of each variable in each statement) and .errors.txt (only if there are errors).

[DanielRosenwasser]

Space after if.

[JsonFreeman]

Make sure the parent is guaranteed to be a switch statement. It is possible that the parser lets through case labels that are not parented by a switch. Please add either an assert or a bailout condition.

[DanielRosenwasser]

To add to what @yuit said, you'll need to ensure that you're writing uniquely named tests.

Decent test names for this might be

• typeGuardsSwitchStatement[SpecificTestName].ts
• typeGuardsSwitchStatement[SpecificTestName]ES6.ts

Feel free to take inspiration for existing tests; in fact, you should be testing at least the same features as are tested in if/else-style type guards, but with switches. For instance, if you notice the test tests/cases/conformance/expressions/typeGuard/typeGuardsDefeat.ts, you should do basically the same thing.

You should be testing with strings like

• "string"
• "number"
• "boolean"
• "symbol"
• "object"
• "watermelon" (or something more seasonal for March - perhaps "apple" or "rhubarb")

Fourslash tests would need to be added too. If we actually move further along, we can guide you through this process.

This might be too permissive. In the previous example, if I were to have a case 'Watermelon' it would still work. Should I check to see that the type is a subset of v's type?

Just as we don't do anything in an if type guard, nothing should be done there.

---

I would imagine that this change immediately prompts the need for getting the appropriate union type in the case of a fallthrough, or potentially eliminating accounted cases within the default clause.

function f(v: string | number | boolean) {
    switch (typeof v) {
        case 'number':
            // 'v' has type 'number'
        case 'string':
            // 'v' now has type 'number | string'
            break;
        default:
            // 'v' has type 'boolean'
    }

So we would have to do some sort of control flow analysis to do this. We could have this change temporarily, but I imagine the behavior might be somewhat surprising. I think we should discuss it further.

[JsonFreeman]

To address the question of permissiveness, please follow the same model as narrowTypeByEquality

[ghost]

@bsteephenson, this repo is using .editorextension. If you are developing using VS, you would probably want to install editorconfig extension, which automagically take editor formatting configuration from the .editorconfig file and apply to VS.

https://visualstudiogallery.msdn.microsoft.com/c8bccfe2-650c-4b42-bc5c-845e21f96328

Extensions are also available for vim, nodepad++, sublime etc.

[mhegazy]

This PR seems stale. @bsteephenson please refresh and resubmit if this is still something you still want to pursue.