Internal error (is it a bug?)

[iZzz32]

Code Select Expand


typedef int (foo)(int);

int main(void)
{
    foo bar = baz; // I know -- this is an error, I missed '*' before 'bar'.
    return 0;
}

produces the following messages:

Code Select Expand


test.c(13): error #2043: Illegal initialization of 'extern bar'.           <- OK.
test.c(13): error #2031: Addressable object required.                      <- What does it mean?
test.c(13): fatal error: Internal error: 'Access violation' at 0x0041701d. <- Is it a bug?


[DMac]

Code Select Expand

typedef int (*foo)(int); // declared function pointer to point to an object's address

int baz (int i){return 0;} // declare an addressable object

int main(void)
{
    foo bar = baz; // now the symbol baz is an addressable object and foo points to it
    return 0;
}

[Pelle]

A bug, I guess; can happen when the compile has already failed, so no show-stopper, but obviously looks better if it doesn't happen. Anyway, it seems I have already fixed this in the new version...

[AlexN]

Anyway, it seems I have already fixed this in the new version...

Witch new version? 

[Stefan Pendl]

Witch new version? 

The one Pelle is currently working on, which will support x64 too.

[Pelle]

Yup. I'm on the final 5% - which seems to take the 'usual' 95% of the time... 

[Christian]

I am looking forward this new version 

Pelle, could you tell something about the planned changes within the announcements?

[Pelle]

Maybe easier to list what hasn't change... 

Here is a partial list...
* support for AMD64/X64/X86-64/whatever you like to call it... (obviously...)
* no support for Windows 9X in POIDE, and probably most tools except POLINK & POLIB...
* support for C99 complex math...
* support for "safer" C runtime functions from ISO TR24731-1...
* basic "locale" support...
* some 30 -40 new compiler intrinsic functions...
* Windows Vista support...

[ssb]

* Windows Vista support...

Great! I was looking for that. I've experienced too many issues with IDE since I've installed Vista.