On Tuesday, December 10, 2013 15:52:06 Timothee Cour wrote:
> in std.utf shouldn't we throw UTFException instead of assert(0) in stride,
> etc ?

std.utf only uses assert(0) when that code should be unreachable. It's used to catch bugs in the implementation, not in a function's input. If anything in std.utf ever hits an assert(0), it's a bug. A UTFException _is_ thrown when a std.utf function encounters invalid Unicode. It's certainly possible that there's a bug involving assert(0) in std.utf, but there shouldn't be any case in all of Phobos where an assert(0) would actually be hit.

- Jonathan M Davis

well then it's a bug:

https://d.puremagic.com/issues/show_bug.cgi?id=11721

this is the 3rd regression I'm posting today!



On Tue, Dec 10, 2013 at 4:45 PM, Jonathan M Davis <jmdavisProg@gmx.com>wrote:

> On Tuesday, December 10, 2013 15:52:06 Timothee Cour wrote:
> > in std.utf shouldn't we throw UTFException instead of assert(0) in
> stride,
> > etc ?
>
> std.utf only uses assert(0) when that code should be unreachable. It's
> used to
> catch bugs in the implementation, not in a function's input. If anything in
> std.utf ever hits an assert(0), it's a bug. A UTFException _is_ thrown
> when a
> std.utf function encounters invalid Unicode. It's certainly possible that
> there's a bug involving assert(0) in std.utf, but there shouldn't be any
> case
> in all of Phobos where an assert(0) would actually be hit.
>
> - Jonathan M Davis
>