The max value and null value of uint type in PrimitiveValue.java is different from documentation of SBE.

[dnbllglm]

public static final long MIN_VALUE_UINT32 = 0;
public static final long MAX_VALUE_UINT32 = 4294967293L; // 0xFFFFFFFD
public static final long NULL_VALUE_UINT32 = 4294967294L; // 0xFFFFFFFE

There is some code that define some constant value of basic type of number in PrimitiveValue.java.I can't understand why NULL_VALUE_UINT32 is 0xFFFFFFFE(2 ^ 32 - 2) instead of 0xFFFFFFFF(2 ^ 32 - 1), but NULL_VALUE_UINT16 is 65535(2 ^ 16 - 1) and NULL_VALUE_UINT8 is 255(2 ^ 8 - 1).

The documentation is in Range attributes for integer fields

[billsegall]

See section 2.4.2 of the specification at http://www.fixtradingcommunity.org/pg/file/fplpo/read/1196757/simple-binary-encoding-release-candidate-2 . Essentially this results from the default NullValue being 2**n -1 for an n bit unsigned type so 2**n -2 was used for MaxValue for these types. Signed types behave as you probably expect them to because they use -2**n for NullValue. It's far from intuitive but you can set min/max/null values in your schema to be whatever you would prefer . Bill.

…

On Wed, May 17, 2017 at 1:10 PM, dnbllglm ***@***.***> wrote: public static final long MIN_VALUE_UINT32 = 0; public static final long MAX_VALUE_UINT32 = 4294967293L; // 0xFFFFFFFD public static final long NULL_VALUE_UINT32 = 4294967294L; // 0xFFFFFFFE There is some code that define some constant value of basic type of number in PrimitiveValue.java.I can't understand why NULL_VALUE_UINT32 is 0xFFFFFFFE(2 ^ 32 - 2) instead of 0xFFFFFFFF(2 ^ 32 - 1), but NULL_VALUE_UINT16 is 65535(2 ^ 16 - 1) and NULL_VALUE_UINT8 is 255(2 ^ 8 - 1). The documentation is in Range attributes for integer fields <https://github.com/FIXTradingCommunity/fix-simple-binary-encoding/blob/master/v1-0-STANDARD/doc/02FieldEncoding.md#range-attributes-for-integer-fields> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#455>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFI1pEpEdIkRc4UchCSKp7L6BPeVbD5Hks5r6mULgaJpZM4NdUsr> .

[dnbllglm]

public static final long MIN_VALUE_UINT32 = 0;
public static final long MAX_VALUE_UINT32 = 4294967293L; // 0xFFFFFFFD
public static final long NULL_VALUE_UINT32 = 4294967294L; // 0xFFFFFFFE

Thank you for answering my question,but it is using 0xFFFFFFFE as null value of unsigned int type and using 0xFFFFFFFD as max value for unsigned int type in PrimitiveType.java.In fact,we should use 0xFFFFFFFF as null value of unsigned int type and use 0xFFFFFFFE as max value of unsigned int type like unsigned short type and unsigned byte type,for example:

public static final long MIN_VALUE_UINT32 = 0;
public static final long MAX_VALUE_UINT32 = 4294967294L; // 0xFFFFFFFE
public static final long NULL_VALUE_UINT32 = 4294967295L; // 0xFFFFFFFF

[billsegall]

Sorry, I see the confusion now. I suspect this results from Java 7's lack of support for unsigned ints but I'm unsure. Bill.

…

On Wed, May 17, 2017 at 5:19 PM, dnbllglm ***@***.***> wrote: public static final long MIN_VALUE_UINT32 = 0; public static final long MAX_VALUE_UINT32 = 4294967293L; // 0xFFFFFFFD public static final long NULL_VALUE_UINT32 = 4294967294L; // 0xFFFFFFFE Thank you for answering my question,but it is using 0xFFFFFFFE as null value of unsigned int type and using 0xFFFFFFFD as max value for unsigned int type in PrimitiveType.java.In fact,we should use 0xFFFFFFFF as null value of unsigned int type and use 0xFFFFFFFE as max value of unsigned int type like unsigned short type and unsigned byte type,for example: public static final long MIN_VALUE_UINT32 = 0; public static final long MAX_VALUE_UINT32 = 4294967294L; // 0xFFFFFFFE public static final long NULL_VALUE_UINT32 = 4294967295L; // 0xFFFFFFFF — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#455 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFI1pDz1DJZytJlaEHN5NYXYnphcueShks5r6p-YgaJpZM4NdUsr> .

[mjpt777]

I believed this may have been a hangover from early spec versions. I've pushed a correction. @tmontgomery is there something we might have missed here?

[tmontgomery]

@mjpt777 I believe the change you made is good. Not sure why the values were that way off hand.

[da4089]

This probably should be highlighted in the release notes, since schemas/stubs using default NULL values (and perhaps, MAX, although that seems less likely to be commonly used) will be incompatible?

[mjpt777]

@da4089 Agreed.