Raise tracking support for chronos procedures #298
Conversation
| raises: [Defect, CatchableError].} = | ||
| # we have to rely on inference here | ||
| {.pop.} | ||
| proc read*(future: Future | RaiseTrackingFuture ): auto = |
There was a problem hiding this comment.
just make a separate read function for RTF? the auto return type here is unnecessarily wide as well, there's too many compromises for too little benefit
There was a problem hiding this comment.
Indeed, we can avoid it here by making a specialized read for RTF
| return retFuture | ||
|
|
||
| proc cancelAndWait*[T](fut: Future[T]): Future[void] = | ||
| proc cancelAndWait*[T](fut: Future[T]): RaiseTrackingFuture[void, (CancelledError,)] = |
There was a problem hiding this comment.
| proc cancelAndWait*[T](fut: Future[T]): RaiseTrackingFuture[void, (CancelledError,)] = | |
| proc cancelAndWait*[T](fut: Future[T]): Future[void] {.asyncraises:[(CancelledError,)] = |
the rewrite to RTF should be macro-handled, no?
There was a problem hiding this comment.
We don't have access to macros here (asyncmacro2 requires asyncfutures2)
There was a problem hiding this comment.
maybe we should move this one elsewhere then - ie cancelAndWait is a composite "utility" operation much like allFutures etc - then we keep "primitives" in asyncfuture2 and maybe one fine day we can get rid of include altogether
|
|
||
| return retFuture | ||
|
|
||
| proc wait*[T, E](fut: RaiseTrackingFuture[T, E], timeout = InfiniteDuration): auto = |
There was a problem hiding this comment.
what's the return type here? in public functions, we ideally want to avoid auto both for documentation reasons and because it serves as a type inference guide for the compiler.
There was a problem hiding this comment.
It's RTF[T, E + CanceledError] (which I just realized is wrong, should be AsyncTimeoutError)
I don't know a better solution than auto for this kind of procedures (but I don't think it has an impact on type inference)
|
superseded by #454 |
2 participants
Follow up on #251