| Author |
 Topic  |
|
|
beylevem
Tomato Guru
102 Posts |
 Posted - Aug 20 2009 : 3:58:32 PM
|
If I have a typedef
typedef void (*PtrToFn)();
and I then try to use it by typing
PtrT(tab) - i.e. completion,
I get
PtrToFn()
i.e. redundant/incorrect parentheses
|
|
|
feline
Whole Tomato Software
United Kingdom
18947 Posts |
Posted - Aug 20 2009 : 6:00:46 PM
|
| I am confused. From memory you need to put brackets on the end when calling a function pointer, so it is reasonable for VA to add the backets. Am I missing something obvious here? |
zen is the art of being at one with the two'ness |
 |
|
|
beylevem
Tomato Guru
102 Posts |
Posted - Aug 21 2009 : 09:11:17 AM
|
We are talking about the typedef, not the variable. To expand my example:
typedef void (*PtrToFn)();
PtrToFn aPtr = &someFn;
aPtr();
Note that this is not a pointer to member function, just a pointer to function. Also, with pointer to function, one can write either
aPtr() or (*aPtr)()
|
|
|
|
feline
Whole Tomato Software
United Kingdom
18947 Posts |
Posted - Aug 21 2009 : 12:03:17 PM
|
This makes sense, and I remember these two syntaxes for calling the function. But simply taking the "aPtr()" syntax, why would you want to write "aPtr" and not add the brackets?
When I write a function name in my code then 99.999% of the time I want the brackets on the end since I am calling the function. Am I still missing something obvious?
Off hand the only reason I can see for not wanting the brackets is if you are writing the if statement in:
if(NULL <> aPtr)
{
aPtr();
}
|
zen is the art of being at one with the two'ness |
|
|
|
beylevem
Tomato Guru
102 Posts |
Posted - Aug 21 2009 : 1:44:04 PM
|
You're still missing the point. The problem is not with
aPtr() i.e. aP(tab) ==> aPtr()
but with
PtrT(tab) i.e. PtrT(tab) ==> PtrToFn()
The paretheses should not be added when the typedef is used  |
|
|
|
feline
Whole Tomato Software
United Kingdom
18947 Posts |
Posted - Aug 24 2009 : 1:33:23 PM
|
*ah* yes, this makes sense. I thought there had to be confusion going on here
Can you try VA 1731? So far I cannot reproduce this problem with this new build, which does contain some fixes for brackets being appended when accepting listboxes.
http://www.wholetomato.com/downloads/default.asp
|
zen is the art of being at one with the two'ness |
|
|
|
beylevem
Tomato Guru
102 Posts |
Posted - Aug 25 2009 : 09:29:13 AM
|
Unfortunately, 1731 does not fix the problem for me. I opened a new instance of VS2008, added a header file, and typed
class C1 {
typedef bool (*chgFnPtr)(C1*, int);
chgFnPtr()
};
As soon as I do the second chg(tab), VA adds the parentheses.
|
|
|
|
feline
Whole Tomato Software
United Kingdom
18947 Posts |
Posted - Aug 25 2009 : 2:35:00 PM
|
I am seeing the same problem now:
case=31476
I had simplified the test a bit to much, and was testing this with the typedef inside a function, and the problem does not show up under those conditions. Oops |
zen is the art of being at one with the two'ness |
|
|
| |
Topic |
|
Pointer to function typedefs
