问题
I have to archive a float
with value INFINITY
, and later to dearchive it.
Here is my example code:
Object to be archived:
@interface CodeInf : NSObject <NSCoding>
@end
@implementation CodeInf
- (void)encodeWithCoder:(NSCoder *)encoder {
float inf = INFINITY;
[encoder encodeFloat: inf forKey:@"INFINITY"];
}
- (id)initWithCoder:(NSCoder *)decoder {
if (self = [super init]) {
float decodedInf = [decoder decodeFloatForKey: @"INFINITY"];
}
return self;
}
@end
And here is the archiving/dearchiving code:
CodeInf *myCodeInf = [[CodeInf alloc] init];
NSData *data = [NSKeyedArchiver archivedDataWithRootObject:myCodeInf];
myCodeInf = [NSKeyedUnarchiver unarchiveObjectWithData:data];
Archiving works, but dearchiving raises the error:
* Terminating app due to uncaught exception 'NSRangeException', reason: '* -[NSKeyedUnarchiver decodeFloatForKey:]: value (inf) for key (INFINITY) too large to fit in 32-bit float'
Is this a bug in the dearchiver, or do I something wrong?
回答1:
Looks like a bug. Submit a bug report to Apple.
As a workaround use encodeDouble
and decodeDoubleForKey
- you can keep your value as a float
and no casts are needed by C rules.
来源:https://stackoverflow.com/questions/17059941/float-infinity-can-be-archived-by-nscoder-but-not-dearchived