From 8ecf1531d93acda9f904f11efb2a34dfec169153 Mon Sep 17 00:00:00 2001 From: Kimplul Date: Wed, 11 Dec 2024 18:02:54 +0200 Subject: add note about closure calls --- README.md | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) (limited to 'README.md') diff --git a/README.md b/README.md index 572380d..c3ecf85 100644 --- a/README.md +++ b/README.md @@ -296,3 +296,23 @@ create_some_object((*whatever_t) next) but I'm not entirely certain if this can be automatically detected. In theory I guess we could implement the `safe` and `unsafe` attributes and provide wrappers for creating `box` pointers that immediately fix the issue, but hmm. + +### Calling closure multiple times + +This is a relatively small detail, same as in Rust. Depending on how a closure +is used, it can have different properties. If the closure is called just once +(probably the most common situation), it is allowed to move captured variables. +If called multiple times, it is not, since a previous invocation might've +already moved some variables. + +Then there are also function pointers that don't have any closure attached to +them. Currently I'm thinking of using the syntax +``` + (int) - closure that can be called once +&(int) - closure that can be called multiple times +*(int) - raw function pointer, can be called multiple times +``` + +The semantics for passing these different types around seem to match variables +fairly close, where a 'call' is a kind of destructive move. The exact semantics +are still unimplemented, though. -- cgit v1.2.3