gnustep/libs-base
0
0
Fork 0
mirror of https://github.com/gnustep/libs-base.git synced 2025-04-23 00:41:02 +00:00
Code Issues Projects Releases Packages Wiki Activity

Add special small object class for floats. Previously, NSSmallRepeatableDouble

Browse source 
would be used which resulted in the float being printed as a double when calling
-stringValue or -description.


git-svn-id: svn+ssh://svn.gna.org/svn/gnustep/libs/base/trunk@35022 72102866-910b-0410-8b05-ffd578937521
This commit is contained in:
Niels Grewe 2012-04-01 07:33:38 +00:00
parent a490907e32
commit 490dfca739
2 changed files with 48 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

7
ChangeLog
View file

@ -1,3 +1,10 @@
2012-04-01 Niels Grewe <niels.grewe@halbordnung.de>
* Source/NSNumber.m: Add special small object class for floats.
Previously, NSSmallRepeatableDouble would be used which resulted in
the float being printed as a double when calling -stringValue or
-description.
2012-03-27 Fred Kiefer <FredKiefer@gmx.de>
* Headers/Foundation/NSObject.h (-error:): Declare return type.

47
Source/NSNumber.m
View file

@ -6,7 +6,7 @@
(to compile on gnu/linux and mswindows,
to meet coding/style standards,
to restore lost functionality)
Date: February 2010
This file is part of the GNUstep Base Library.
@ -292,7 +292,7 @@ return NSOrderedSame;
*/
@interface NSFloatingPointNumber : NSNumber
@end
@implementation NSFloatingPointNumber
/* For floats, the type promotion rules say that we always promote to a
* floating point type, even if the other value is really an integer.
@ -352,10 +352,12 @@ return NSOrderedSame;
static BOOL useSmallInt;
static BOOL useSmallExtendedDouble;
static BOOL useSmallRepeatingDouble;
static BOOL useSmallFloat;
#define SMALL_INT_MASK 1
#define SMALL_EXTENDED_DOUBLE_MASK 2
#define SMALL_REPEATING_DOUBLE_MASK 3
// 4 is GSTinyString
#define SMALL_FLOAT_MASK 5
@interface NSSmallInt : NSSignedIntegerNumber
@end
@ -569,6 +571,39 @@ load
return;
}
@end
/*
* Technically, all floats are small on 64bit and fit into a NSRepeatingDouble,
* but we want to get the description FORMAT right for floats (i.e. "%0.7g" and
* not "%0.16g".
*/
@interface NSSmallFloat : NSSmallRepeatingDouble
@end
@implementation NSSmallFloat
#undef VALUE
#define VALUE (unboxSmallRepeatingDouble((uintptr_t)self))
#define FORMAT @"%0.7g"
#include "NSNumberMethods.h"
+ (void) load
{
useSmallFloat = objc_registerSmallObjectClass_np
(self, SMALL_FLOAT_MASK);
}
+ (id) alloc
{
return (id)SMALL_FLOAT_MASK;
}
+ (id) allocWithZone: (NSZone*)aZone
{
return (id)SMALL_FLOAT_MASK;
}
@end
#endif
#endif
@ -691,7 +726,7 @@ static NSBoolNumber *boolN; // Boolean NO (integer 0)
/*
* Macro for checking whether this value is the same as one of the singleton
* instances.
* instances.
*/
#define CHECK_SINGLETON(aValue) \
if (aValue >= -1 && aValue <= 12)\
@ -860,9 +895,9 @@ if (aValue >= -1 && aValue <= 12)\
objCType: @encode(float)] autorelease];
}
#if OBJC_SMALL_OBJECT_SHIFT == 3
if (useSmallRepeatingDouble)
if (useSmallFloat)
{
return boxDouble(aValue, SMALL_REPEATING_DOUBLE_MASK);
return boxDouble(aValue, SMALL_FLOAT_MASK);
}
#endif
n = NSAllocateObject (NSFloatNumberClass, 0, 0);