Code generation for WITH COMPONENTS

[shd01x]

If a sequence with optional components is fully constrained by a WITH COMPONENTS clause, the autogenerated type nevertheless contains an "exist" field, and absent components are still encoded.

For example, the following ASN.1

Test DEFINITIONS ::=

BEGIN

T-Test ::= SEQUENCE {
  field1 INTEGER,
  field2 INTEGER,
  field3 INTEGER,
  field4 INTEGER,
  field5 INTEGER
}

T-Test2 ::= SEQUENCE {
  field1 INTEGER OPTIONAL,
  field2 INTEGER OPTIONAL,
  field3 INTEGER OPTIONAL,
  field4 INTEGER OPTIONAL,
  field5 INTEGER OPTIONAL
}

T-Test3 ::= T-Test2 ( WITH COMPONENTS {
  field1 PRESENT,
  field2 PRESENT,
  field3 ABSENT,
  field4 ABSENT,
  field5 PRESENT
  } )
  
END
  

generates these types:

typedef struct {
    asn1SccSint field1;
    asn1SccSint field2;
    asn1SccSint field3;
    asn1SccSint field4;
    asn1SccSint field5;
} T_Test;

#define T_Test_REQUIRED_BYTES_FOR_ENCODING       45 

typedef struct {
    asn1SccSint field1;
    asn1SccSint field2;
    asn1SccSint field3;
    asn1SccSint field4;
    asn1SccSint field5;
    struct {
        unsigned int field1:1;
        unsigned int field2:1;
        unsigned int field3:1;
        unsigned int field4:1;
        unsigned int field5:1;
    } exist;
} T_Test2;

#define T_Test2_REQUIRED_BYTES_FOR_ENCODING       46 

typedef struct {
    asn1SccSint field1;
    asn1SccSint field2;
    asn1SccSint field3;
    asn1SccSint field4;
    asn1SccSint field5;
    struct {
        unsigned int field1:1;
        unsigned int field2:1;
        unsigned int field3:1;
        unsigned int field4:1;
        unsigned int field5:1;
    } exist;
} T_Test3_1;

#define T_Test3_1_REQUIRED_BYTES_FOR_ENCODING       46 

So space for the fields is still reserved in the encoding even when they are known to be absent, and space for the "exist" field is also unnecessary IMHO.

I would strongly prefer the ABSENT fields not to appear in the C struct. Please could this be considered?

[maxime-esa]

This was recently discussed here: http://taste.tuxfamily.org/mantis/view.php?id=543
(Let me know if you do not have an account or face issues creating one/accessing it)

Basically two conclusions were reached:

1. Removing the .exist field could break existing code so we decided to keep it, but the encoding function ignores it. Note that with uPER the bit in the encoding packet has to be there in any case, this is a rule of the uPER encodings (this subtype constraint is not visible to uPER). So even if you have a WITH COMPONENTS (PRESENT or ABSENT), it remains an optional field from the point of view of the encoder.

2. ACN was updated to solve this last "issue" by not putting this extra bit in the packets if there is a subtype constraint

[shd01x]

Thanks, Maxime. I searched mantis for this but obviously my search wasn't good enough :(

I will explore what can be done with the ACN and close this issue.