I'm trying to understand why this code returns an empty string rather than an out of bounds error:

writeln("this"[4..$]); // Empty string

But

writeln("this"[4]);

won't compile. I thought maybe it had something to do with the $ or the .. operator, but "this"[4..4] compiles and runs while "this"[5..5] does not. Therefore it's not the fact that there are no elements in the slice.

On the other hand, "this"[3..3] also returns an empty string. That still doesn't explain why there is no out of bounds error for "this"[4..4].

On 8/11/23 12:04 PM, bachmeier wrote:

writeln("this"[4..$]); // Empty string

writeln("this"[4]);

In a slicing operation, the second index is one index beyond the data you are looking to slice. In other words, the upper bound is exclusive of the data.

The data looks like this:

data: [ t h i s ]
index:  0 1 2 3 4

So "this"[1 .. 3] will do it like this:

data: [ t h i s ]
index:  0 1 2 3 4
slice:   [h i ]

If you slice 4 .. 4, then you get:

data: [ t h i s ]
index:  0 1 2 3 4
slice:         []

This is allowed, because the pointer is still within bounds of the array.

If you slice 5 .. 5, that is out of bounds of the entire array, and even though you aren't getting any data, the pointer will point beyond the scope of the array, and isn't allowed.

-Steve

On Friday, 11 August 2023 at 16:04:21 UTC, bachmeier wrote:

writeln("this"[4..$]); // Empty string

Something goes wrong is indeed a 2nd degree problem. The important thing is that slicing a string might be wrong if it contains UTF.

SDB@79