I'm not 100% sure the semantics here are nailed down - but I think there's no guarantee that `a` continues to exist after assignment to `b`. The value in it has been moved out of it after all... The memory which was used for the variable `a` can probably be re-used for something else, e.g. for some inlined variable used by `println!`...
In normal rust `let a = b` where the variable is of a non-Copy type (including String) is "destructive" in the sense that you can no longer use b.
The question about semantics in normal rust turns to "so if I have a raw-pointer to a hanging around and use unsafe code to copy the value out of it what do I get" and I'm not 100% sure... but I think the answer is probably it's a use after free and you get undefined behavior. The rust-- version is basically just this except you don't have to explicitly make that raw pointer to read the old memory.
In normal rust `let a = b` where the variable is of a non-Copy type (including String) is "destructive" in the sense that you can no longer use b.
The question about semantics in normal rust turns to "so if I have a raw-pointer to a hanging around and use unsafe code to copy the value out of it what do I get" and I'm not 100% sure... but I think the answer is probably it's a use after free and you get undefined behavior. The rust-- version is basically just this except you don't have to explicitly make that raw pointer to read the old memory.