Does it make sense to add attribute to operator overload functions ?
Everything is in the Q. I ask this because those functions are
hidden behind symbols and keywords (+=, ~, in, etc.). It's not
that obvious for a user who would write a custom type.
e.g:
---
struct myType
{
@safe nothrow opIndexAssign(t1 paramValue,t2 paramIndex){}
}
---
are the attributes necessary ? Are opXXX functions handled just
like any other functs ? (traversal compat. of the attribs)
Re: Does it make sense to add attribute to operator overload functions ?
On Saturday, 25 October 2014 at 17:14:51 UTC, Jkpl wrote:
Everything is in the Q. I ask this because those functions are
hidden behind symbols and keywords (+=, ~, in, etc.). It's not
that obvious for a user who would write a custom type.
e.g:
---
struct myType
{
@safe nothrow opIndexAssign(t1 paramValue,t2 paramIndex){}
}
---
are the attributes necessary ? Are opXXX functions handled just
like any other functs ? (traversal compat. of the attribs)
In every aspect they are ordinary functions, except they each
have an additional unique way of being called (the relevant
operator syntax).
In short, yes.
Re: Does it make sense to add attribute to operator overload functions ?
On Saturday, 25 October 2014 at 18:38:12 UTC, John Colvin wrote:
On Saturday, 25 October 2014 at 17:14:51 UTC, Jkpl wrote:
Everything is in the Q. I ask this because those functions are
hidden behind symbols and keywords (+=, ~, in, etc.). It's not
that obvious for a user who would write a custom type.
e.g:
---
struct myType
{
@safe nothrow opIndexAssign(t1 paramValue,t2 paramIndex){}
}
---
are the attributes necessary ? Are opXXX functions handled
just like any other functs ? (traversal compat. of the attribs)
In every aspect they are ordinary functions, except they each
have an additional unique way of being called (the relevant
operator syntax).
In short, yes.
Thx, A bit less confused by them now.
