Recent discussions on the zero terminated string problems and
inconsistency of string literals has me, again, wondering why D doesn't
have a 'type' to represent C's zero terminated strings. It seems to me
that having a type, and typing C functions with it would solve a lot of
problems.
The compiler could/would then error if people attempted to pass a D string
without converting it correctly.
The compiler would create literals with or without \0 as required by the
'type' being assigned, parameter passed, etc.
The conversion function from a D string to a C string would return the new
type.
A %sz format specifier could be added to writef which would be able to
type check the argument.
As the only way to get a variable of the new type would be from a literal,
conversion or C function call so we could be sure it was in fact \0
terminated(*), and so..
An implicit conversion between a C string and a D string (slice using
strlen) would be possible, and safe. (Though, not at zero runtime cost)
Existing (correct) code would continue to compile, by this I mean:
- passing literals
- calling a conversion function for each D string argument
But code which passes D string variables to C functions without conversion
would start to fail to compile, so the change will 'break' existing code.
There would be several solutions in these cases:
1) add a call to a conversion function. Introducing a conversion cost
which was not previously present.
2) re-type the variable as a C string. If it's not used for anything else
then this is more "correct". If it's passed to other code then because a
C string will implicitly converts (with a slice/strlen) to a D string this
substitution will work in most cases, however that act of conversion will
incur a cost (but it can/should be one off if the result is assigned/kept).
I am probably missing something obvious, or I have forgotten one of the
array/slice complexities which makes this a nightmare.
Thoughts?
Regan
(*) Ignoring buggy/broken C functions returning non-zero terminated
strings.. as we will crash on these no matter what in any case.
--
Using Opera's revolutionary email client: http://www.opera.com/mail/
