Re: wordRangeForRange implememtation

FROM : Adam R. Maxwell
DATE : Mon May 12 00:39:33 2008

On May 11, 2008, at 2:45 PM, Jérome Laurens wrote:

> Le 11 mai 08 à 04:42, Ali Ozer a écrit :
>

>> To extract a word range from a string, it may be easier to use 
>> something like NSScanner:
>>
>> // Defining the words
>> NSCharacterSet *validWordSet = [NSCharacterSet 
>> alphanumericCharacterSet];        // Or however you want to define words
>> NSCharacterSet *wordBreakSet = [validWordSet invertedSet];
>>
>> // Set up a scanner
>> NSScanner *scanner = [NSScanner scannerWithString:yourString];
>> [scanner setCharactersToBeSkipped:nil];  // We will manage 
>> skipping ourselves (since we need to find the beginning of the word)

>
> A long time ago, I played seriously with NSScanner to implement some 
> syntax coloring.
> It was unbearably slow, and I ended with the conclusion that 
> [NSScanner scannerWithString:]; was making a copy of the given string.
> So I reimplemented my own NSScanner and it was way faster.
>
> Is it true that NSScanner duplicates the string?


Interesting question...I wrote a quick test program and set a 
breakpoint on -[NSCFString copyWithZone:], and NSScanner indeed copied 
it.  For immutable strings, that's probably a retain, but it could be 
pretty expensive for mutable strings or the NSTextStorage proxy.

The main performance problem I've had with NSScanner is that -
[NSScanner scanCharactersFromSet:] creates an autoreleased inverted 
NSCharacterSet, which can blow up an autorelease pool pretty quickly 
(rdar://problem/4652388).  My conclusion is that it's not great for 
parsing large strings or in a loop.

--
Adam_______________________________________________

Cocoa-dev mailing list (<email_removed>)

Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com

Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/cocoa-dev/<email_removed>

This email sent to <email_removed>