The struct core.sys.linux.sys.inotify.inotify_event contains the field "char[0] name" which corresponds to "char name[]" in C.

Why it has been translated to "char[0]" ?
For me "char*" would have been more appropriate.

On Wednesday, 9 September 2015 at 16:49:39 UTC, badlink wrote:
> The struct core.sys.linux.sys.inotify.inotify_event contains the field "char[0] name" which corresponds to "char name[]" in C.
>
> Why it has been translated to "char[0]" ?

In that structure, the name is appended directly to the end of the data instead of pointed to. The size isn't known at compile time, so a zero-length placeholder is there instead.

You'd access it by using the .ptr property and get the length out of the `len` field.

> For me "char*" would have been more appropriate.

That's typically right, but since this case does it in-place instead of pointed, the zero length array is most accurate.

On Wednesday, 9 September 2015 at 16:59:09 UTC, Adam D. Ruppe wrote:
> That's typically right, but since this case does it in-place instead of pointed, the zero length array is most accurate.

I didn't consider that the name is placed right after the struct. Thanks !

On Wednesday, 9 September 2015 at 16:49:39 UTC, badlink wrote:
> The struct core.sys.linux.sys.inotify.inotify_event contains the field "char[0] name" which corresponds to "char name[]" in C.
>
> Why it has been translated to "char[0]" ?
> For me "char*" would have been more appropriate.

It's a flexible array member [1], not a pointer. Changing it to `char*` would make it incompatible with the C functions using it.

[1]: https://en.wikipedia.org/wiki/Flexible_array_member

On Wednesday, 9 September 2015 at 19:37:54 UTC, Alex Parrill wrote:
> It's a flexible array member [1], not a pointer. Changing it to `char*` would make it incompatible with the C functions using it.
>
> [1]: https://en.wikipedia.org/wiki/Flexible_array_member

TIL a new detail about C on the D forum ;)