[Development] The age-old T* foo vs. T *foo

[André Somers]

On 18/10/2019 02:37, Kevin Kofler wrote:
> Ville Voutilainen wrote:
>> Since we are about to do a major version upgrade, should be stop being
>> a special snowflake in the C++ world and start attaching pointer-stars
>> and reference-ampersands to the type instead of to the variable?
> No, because it is syntactically not a part of the type, as evidenced by the
> int *x, y; example that others have already pointed out. Semantically, it
> is, but the semantics are only computed after doing the syntactic parsing,
> so the whitespace should preferentially match the syntax.
>
> And I consider the idea of banning comma declarations to work around this
> issue to be nothing more than a workaround for poor notation not matching
> the syntax of the language.

Apparently there are already good reasons not to do that even before 
considering T* foo vs T *foo, as it has been banned by the style already:

https://wiki.qt.io/Qt_Coding_Style#Declaring_variables

So this potential change will not cause issues in this area.


Personally, I have always preferred the T* foo (and T& bar) style, as to 
me the pointer-ness or reference-ness *is* part of the type of the 
variable. It is evident in how big the type is, in how typeid identifies 
the type of the variable, in type deduction...

Anyway, I doubt it's worth changing it.

André