Re: Strings and NSRanges

Graham J Lee <[email protected]>
Newsgroups gmane.comp.lib.gnustep.user
Message-ID <[email protected]>
On 20 Nov 2006, at 08:24, Richard Frith-Macdonald wrote:

>
> On 19 Nov 2006, at 15:45, Graham J Lee wrote:
>
>> So I'm doing this:
>>
>>>   if ([self allowsFloats] && (NSNotFound != [useFormat  
>>> rangeOfString:@"." ].location))
>>>     {
>>>       decimalPlaceRange = [useFormat rangeOfString: @"." options:  
>>> NSBackwardsSearch];
>>>       while ([placeHolders characterIsMember: [useFormat  
>>> characterAtIndex: NSMaxRange(decimalPlaceRange)]])
>>>         {
>>>           decimalPlaceRange.length++;
>>>           if (NSMaxRange(decimalPlaceRange) == [useFormat length])
>>>             break;
>>>         }
>>>       decimalPlaces=decimalPlaceRange.length;
>>>       if (0 != decimalPlaces)
>>>         displayFractionalPart = YES;
>>>     }
>>
>> which doesn't work...decimalPlaces ends up being some garbage  
>> value such as 2412439 when the length of the string is only, say,  
>> 7.  For that to happen then the length of the *range* must have  
>> always been greater than the length of the string (so that the if 
>> (NSMaxRange(...)...) line never breaks the loop) and the loops  
>> *starts* by reading nonsense memory outside the string.  Why would  
>> that happen?
>
> It's hard to tell without a complete test program ... but I can  
> make a couple of observations.
>
> 1. the loop should not be able to start by 'reading nonsense memory  
> outside the string' because the -characterAtIndex: method should  
> raise an exception if asked to refer to an index outside the string.
> 2. if useFormat is nil then the code is totally broken since  
> sending a message to a nil object is defined to return nil ... but  
> where the method is expected to return a range this means that the  
> resulting behavior is undefined.  On sparc it will cause an  
> immediate crash, but on intel I think it probably just means the  
> returned range contains garbage ... which would explain your symptoms.
>
>
Got it in two that man :-).  I'd overlooked that _[posi|nega] 
tiveFormat don't get set to defaults in the -init method so could be  
nil.

Thanks,
Graham.
-- 
Graham J Lee
http://www.thaesofereode.info/
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.