dmd foreach loops throw exceptions on invalid UTF sequences, use replacementDchar instead

[kdevel]

On Sunday, 7 November 2021 at 04:18:25 UTC, Walter Bright wrote:
> On 11/4/2021 11:30 PM, FeepingCreature wrote:
>> [...] Let them crash, [...] Automagically handling errors by 
>> propagating them in an inert form robs the developers and 
>> users of a chance to avoid a mistake. It's no better than 0.0.
> It's much better than 0.0. 0.0 is indistinguishable from valid 
> data, and is a very common valid value.

Technically it makes no difference if you do not check for 0.0 or 
not for NaN. What makes a difference is using "out of band 
signalling" (exceptions) if its default behavior is process 
termination.

> NaN and ReplacementChar are not valid

The replacemment character '�' is a valid Unicode codepoint 
(U+FFFD).

> and are easily distinguished.

Someone may forget to write explicit code to handle this case 
which most likely leads to data corruption. I choose stack trace 
over potentially corrupted data.