Suggestion Backlog Slog, 10/19/2015

[RyanCavanaugh]

• Weird behaviour with union types and 'strict' object literals #5237 Weird behavior with strict object literals and union types.
  • Anders will think about it and come up with a fix. Seems tractable.
• Report use before definitely defined errors #5207 Report use before definition errors
  • Mohamed to implement for classes and imports. A PR exists but it needs updating. Defer implementing member access of static properties
• out of range index in a tuple magically works #5203 Out of range access of tuple allowed
  • Disagreement on whether and how consistent we should be relative to dotting and string literal [] notation
• Missing compiler warning on if-clause #5109 Missing compiler warning on if clause
  • Accepting a PR that makes EmptyStatement an invalid consequent block
• Support /dev/stdout as a valid --out paramter #4841 Can't --out /dev/stdout
  • Bug, Accepting PRs that fix our use of seek()
• Suggestion: Type guard generic types #4742 Type guards on generic type parameters
  • Accepting PRs. Proposed fix is to check assignability of type guard targets using the apparent type of the target
• static property inheritance complaining when it shouldn't #4628 Static inheritance
  • Experiment with removing this check and see what happens in our RWC/baselines
• Object literal strictness and unions #4452 Object strictness and unions
  • Deferred; this doesn't seem to come up much in practice and the rule would be expensive to implement