Suggestion Backlog Slog Log

[RyanCavanaugh]

We'll be meeting weekly to get through the large number of suggestion-tagged issues that have sat idle for too long. I'll be posting notes here.

[RyanCavanaugh]

4/27/2015

Approved, Accepting PRs (11)

#6 Abstract classes: Approved. Extremely highly-requested. See design notes in my comment at the bottom
#669 External modules with AMD always requires "exports" even when it is not used: Approved. Minor codegen improvement.
#1003 Singleton types under the form of string literal types: Approved. Everyone gets excited when we talk about this one.
#1007 Support user-defined type guard functions: Approved. Meaningfully improves type guards.
#1283 type guards by constructor signature: Approved. Meaningfully improves type guards.
#1295 Type Property type: Approved. Common pattern in JS libraries.
#1544 No output with --out when using import statement: Approved. Catches a common source of user confusion
#1554 typeof this.xxx gives "identifier expected" error: Approved. Patches an oversight in the grammar
#1616 Type aliases requiring type parameters: Approved. Anders believes this will be quite difficult to implement.
#1706 Union Type inference not working in class: Approved, tentatively. Type guards on dotted expressions.
#1789 Make tsc runnable on Nashorn: Approved. Straightforward addition to our I/O wrapper to gain platform support.

Committed (1)

#247 Support looking for modules under node_modules when importing: Commited, assigned to Vlad as part of some other work he's doing

Converted to bug (1)

#282 Rename function / find all references doesn't affect inline object literals

PR desired, but not Approved yet (1)

#1213 Allow classes to be parametric in other parametric classes: Need to see an implementation here to try it out in more depth

Needs More Info / Justification (4)

#943 Use different relation when checking overload vs implementation signature compatibility: Need a concrete proposal here
#1260 Consider property access a form of type guards: Too many open questions at this point
#1790 Support super.apply from child constructor: Does the spread operator suffice here?
#1809 The object primitive type: Needs more use cases / likely errors

Declined (3)

#293 Specify (absolute) path for library projects: Out of scope. Revisit post-2.0 and see if the third-party build ecosystem has landed on anything.
#311 Traits: Too much overlap with a potential future ECMAScript feature
#983 Global type references: Not enough use cases

[fdecampredon]

Nice :) thanks for the notes !

[vvakame]

wow, great!

[RyanCavanaugh]

5/4/2015

Approved, Accepting PRs (4)

#2214 Allow switch type guards. Useful expansion of type guard syntax
#1940 Do we need to use "declare" for type aliased in d.ts files? ‘declare’ should be optional on type aliases (and interfaces), but is a no-op.
#2694 Define assignability relation for primitive-constrained type parameters. For flowing through enum types in generics.
#2369 Spread operator in 'new' expressions support for ES5? User-side workaround is really ugly.

Declined (7)

#1824 Rest binding element should contextually type initializer as a tuple. Too much complexity for not enough gain. Matches existing behavior for not inferring tuples when none exist
#1844 Add downlevel and typing support for const enum members in computed properties. Not needed until we have string enums; re-visit if we get those.
#1857 Make the Compiler and Language Service API Asynchronous. Too complex for not a lot of gain since we are not really I/O bound
#2000 Support override keyword on class methods. Doesn’t provide a consistent experience (methods not marked override would not necessarily not be overrides, which would be too confusing). Not going to add a compiler flag for this. Maybe could solve with an attribute?
#2172 Contextual typing of spread elements. Why spread literals in the first place? Low-priority.
#2328 Generic type constraint does not apply to rest parameters. Does not actually make sense – the array type does not actually flow in to ‘args’ here
#2357 Proposal: new ES6 style syntax to apply typeof on an ambient external module. Not enough gain for a fair bit of complexity

Needs More Info (2)

#2361 Allow wrapped values to be used in place of primitives. Ask for explicit proposal.
#2176 Suggestion: errors based on control flow analysis. What is this asking for, precisely? Lump in with #228 maybe

[RyanCavanaugh]

5/19/2015

Approved, Accepting PRs (2)

#2341 Allow visibility on constructors - Accepted. No other good workarounds for forcing external consumers of your class to use a factory method
#3079 tsc init to generate tsconfig.json - Very useful

Committed / In Progress (1)

#2499 Support single-module transpilation mode - Already in progress.

Needs More Info (4)

#1882 Make periodic pre-release builds available on npm - Assign to Jonathan to figure out what we should do here
#2568 Generate a single declaration file for external modules - Needs clarification on what this work item represents
#2709 Importing files other than TS modules - Needs clarification on whether a proposal in-thread covers the use cases being requested
#2749 Typescript aborts on shebang - Do people intend to run the .ts file, or the .js file? This is somewhat confusing

Postponed (1)

#2176 Suggestion: errors based on control flow analysis - Very desired feature, but the scope and complexity of this are too large to fit before 2.0. Postpone and definitely revisit

Declined (4)

#1824 Rest binding element should contextually type initializer as a tuple - Didn't really see a compelling reason to do this
#2714 Tuple suffix associated with rest element in a binding pattern - the current behavior is the intended behavior
#3022 Embedded syntax block - Far too complex
#2876 Contextual type assertion operator - Unclear that most people can actually reason about this correctly