RE: [Openh323-devel] RE: [OpenH323]ASN Encode bug with Extendable Integers?

[Robert J. <rob...@di...>]

Yes, I agree.

If extendable is TRUE then constraint cannot be FixedConstraint, ergo the
condition is oxymoronic.

I have applied the patch as you suggest.

Robert Jongbloed.


> -----Original Message-----
> From: ope...@li... 
> [mailto:ope...@li...] On Behalf 
> Of Greg Adams
> Sent: Friday, 5 September 2003 11:20 PM
> To: ope...@op...; ope...@li...
> Subject: [Openh323-devel] RE: [OpenH323]ASN Encode bug with 
> Extendable Integers?
> 
> 
> OK, here is what we have found:
> 
> The problem is indeed with ExtendableConstaint integers.
> 
> By looking at the decode we can see that if an 
> ExtendableConstaint integer is _not_ extended, then it is 
> decoded as a FixedConstaint (i.e. there is no length 
> prefixing the value).
> 
> When we look at the Encode, we see that depending on the 
> result of a call to ConstaintEncode we will either encode 
> with the length, or not.
> 
> ConstraintEncode looks like this:
> ---------------------------
> BOOL PASN_ConstrainedObject::ConstraintEncode(PPER_Stream & 
> strm, unsigned value) const {
>   if (!extendable)
>     return constraint != FixedConstraint;
> 
>   BOOL needsExtending = value > upperLimit;
> 
>   if (!needsExtending) {
>     if (lowerLimit < 0) {
>       if ((int)value < lowerLimit)
>         needsExtending = TRUE;
>     }
>     else {
>       if (value < (unsigned)lowerLimit)
>         needsExtending = TRUE;
>     }
>   }
> 
>   strm.SingleBitEncode(needsExtending);
> 
>   return needsExtending || constraint != FixedConstraint;
> }
> ----------------------
> 
> Only ExtendableConstraint objects have the extendable flag 
> set, so in most cases the first if condition is true and the 
> return value is correct. If it is an ExtendableConstraint 
> that has not been extended, then needsExtending is correctly 
> set to false. But the return code (needsExtending || 
> constraint != FixedConstraint) is incorrect as it will return 
> TRUE and force a length to be encoded prior to the value.
> 
> We believe that the fix should be to remove the constraint != 
> FixedConstraint check, and just to return needsExtending.
> 
> Modified code looks like this:
> 
> -------------------
> BOOL PASN_ConstrainedObject::ConstraintEncode(PPER_Stream & 
> strm, unsigned value) const {
>   if (!extendable)
>     return constraint != FixedConstraint;
> 
>   BOOL needsExtending = value > upperLimit;
> 
>   if (!needsExtending) {
>     if (lowerLimit < 0) {
>       if ((int)value < lowerLimit)
>         needsExtending = TRUE;
>     }
>     else {
>       if (value < (unsigned)lowerLimit)
>         needsExtending = TRUE;
>     }
>   }
> 
>   strm.SingleBitEncode(needsExtending);
> 
>   return needsExtending;
> }
> --------------------
> 
> Robert/Craig do you agree with this? We have tested our 
> scenario here without a problem, but need some confirmation 
> that this is a valid fix in all conditions.
> 
> Regards,
> Greg Adams
> 
> -----Original Message-----
> From: Greg Adams 
> Posted At: 04 September 2003 17:54
> Posted To: ridgeway.reflector.OpenH323
> Conversation: ASN Encode bug with Extendable Integers?
> Subject: [OpenH323]ASN Encode bug with Extendable Integers?
> 
> 
> We have the following sample code that shows up an error when 
> encoding the integer which has an asn syntax of:
> rfc-number             INTEGER(1..32768, ...),
> 
> When constructed, the print out shows the correct values, 
> when decoded back into a new class, the printout of this 
> class shows the incorrect value. We are still investigating, 
> but if anyone has any ideas or has seen this problem before, 
> then please let us know.
> 
> Regards,
> Greg Adams
> 
> ------------
> Code Sample:
> 
> H245_MediaPacketizationCapability mpc; 
> mpc.m_h261aVideoPacketization = TRUE; 
> mpc.IncludeOptionalField(H245_MediaPacketizationCapability::e_
> rtpPayload
> Type);
> mpc.m_rtpPayloadType.SetSize(1);
> H245_RTPPayloadType& rtppt = mpc.m_rtpPayloadType[0];
> rtppt.m_payloadDescriptor.SetTag(H245_RTPPayloadType_payloadDe
> scriptor::
> e_rfc_number);
> PASN_Integer& rfc = rtppt.m_payloadDescriptor;
> rfc = 2249;
> 
> PStringStream strstrm2;
> mpc.PrintOn(strstrm2);
> OutputDebugString(strstrm2);
> 
> PPER_Stream strm;
> mpc.Encode(strm);
> strm.CompleteEncoding();
> strm.ResetDecoder();
> 
> H245_MediaPacketizationCapability mpc2;
> mpc2.Decode(strm);
> 
> PStringStream strstrm;
> mpc2.PrintOn(strstrm);
> OutputDebugString(strstrm);
> 
> 
> ---------
> Output Produced:
> 
> {
>         h261aVideoPacketization = TRUE
>         rtpPayloadType = 1 entries {
>           [0]={
>             payloadDescriptor = rfc_number 2249
>           }
>         }
>       }{
>         h261aVideoPacketization = TRUE
>         rtpPayloadType = 1 entries {
>           [0]={
>             payloadDescriptor = rfc_number 521
>           }
>         }
>       }
> 
> --------------------------------------------------------------
> ----------
> Check the FAQ before asking! - 
> http://www.openh323.org/~openh323/fom.cgi
> The OpenH323 
> Project mailing list, using Mailman. To unsubscribe or change 
> your subscription options, goto 
> http://www.openh323.org/mailman/listinfo/op> enh323
> Maintained 
> by Quicknet Technologies, Inc - 
> http://www.quicknet.net
> --------------------------------------------------------------
> ----------
> 
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf 
> _______________________________________________
> Openh323-devel mailing list Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/openh323-devel
>