C is Brittle D is Plastic

[Quirin Schroll]

On Monday, 13 April 2026 at 10:20:33 UTC, Dejan Lekic wrote:
> On Monday, 13 April 2026 at 09:56:19 UTC, Quirin Schroll wrote:
>> If you don’t do C interop yourself, it won’t affect you.
>>
>> If you do C interop, replacing all your `""` literals by `""z` 
>> literals is a simple regex replace away plus a linter 
>> configuration that keeps you from accidentally using `""`. 
>> However, I don’t think you pass many slice typed things to C 
>> functions. If you don’t use a linter on a big project, I’d say 
>> that’s on you as well.
>>
>> I expected this to be controversial, but not to that degree.
>
> Luckily this is just an edition so I do not need to use it. - I 
> wonder who will?
>
> Is there a comprehensive document describing this decision (DIP 
> perhaps)?

No, this was just a thought (maybe a DIP idea) that I had reading 
the posts.

> I am curious to know what do you do when you use packageX that 
> contains a function that takes a string - how do you know 
> whether you pass it a new "" string or ""z string?

If you use a package, you should generally know how to use it. If 
a function is halfway decently written, it’ll take strings as 
`const(char)[]` if it doesn’t expect zero termination. If a 
function takes a `const(char)*`, you already have to read the 
docs for details, though generally speaking, most of those expect 
a NUL-terminated string. If you call such a function with a `""` 
literal, it would become NUL-terminated because of that reason. I 
definitely mentioned this exception. Essentially, a `""` string 
passed to a `const(char)*` directly would be as if you wrote a 
`""z` string.

> You need to go and read the code to understand what that 
> function does, right? Or ""z is not a "string" but some other 
> type (zstring)?

The type of `""z` would be `string`. Essentially, `"abc"z` is 
just `"abc\0"[0..$-1]`.