Christoffer Lerno
Christoffer Lerno
In terms of scope, I think it's unwise to have specialized scope rules for defer. It's just one more pitfall for people to fall into. I think defers in switches...
@bvdberg I also think my solution can be improved upon, but I ran smack into the problem with switches where you decided on opening a new scope for every case....
I feel a little bad for us breaking Duff's Device 😄
BTW I can do the change in parsing of *defer* anyway. Right now it assumes we can actually make the part AFTER the defer a subscope to defer. This is...
I want defer to work as seamless as possible. The only thing we rule out in defer is non-local jumps in defer. That means that return/continue is out (as they...
goto/label analysis in defer is already done and works in this version already.
Some additions here, but not complete #60
@bvdberg Well, the glaring problem is in the parsing code, but creating a custom hash and vector could then be reused across the codebase I think. Note that llvm's implementations...
What about f128? Should we support that where available?
I've opened an issue #87 for this. The only open question is if we use f128 when available.