[Bf-committers] Possible bug, uninitialized var in depsgraph.c

[bjornmose]

Ton Roosendaal wrote:
> Hi,
> 
>>                             subtarget= "";
>> ...
>>
>>                             target = get_constraint_target(con, 
>> &subtarget);
>> // didn't check but i think this sets subtarget to be a valid pointer  
>> OR  NULL
>> .....
>> if(subtarget[0]) <-- MSCV6 build stalls here since subtarget is NULL 
>> (  trying to READ OUT the first element subtarget points to and 
>> reading  out a NULL pointer seems illegal to MSVC6 and me )
> 
> 
> Hrms... I'm used to have strings initialized that way. For example:
> 
> char *str= "blahblah";
> 
> should make "str" a pointer to a static array.
> Using an empty string "" should give a pointer too, with the first  
> element being zero?

you're right, initializing a
char *pc;
with
pc = "blah";
is entirely OK, even in MSVC6 :)
i was confusing it with a strcpy().
OLE