Hi,
the reason is that the asn2eth behaviour for SEQUENCE/SET OF type had to be changed.
Before that change e.g. all SEQUENCE OF INTEGER was mapped to the one Ethereal function.
But it caused problem when it was necessary to change one of all SEQUENCE OF INTEGER field.
Now only SEQUENCE OF <TypeReference> are mergeged together.
If all languaage fields shoudl be merged to one type you can use following directive:
#.TYPE_RENAME
Connect-UUIE/language Language
Setup-UUIE/language Language
AdmissionConfirm/language Language
#.END
"T_*" functions are generated as type dissectors for non-primitive types.
The ans2eth is diesigned more "type oriented" in comaprisn with current PER/BER API which
is more "field oriented".
It was designed for new PER/BER API which is unfortunately still only in my mind. :)
If you are interested in how ASN.1 fields and types are mapped to Ethereal fields and
types, you can use "asn2eth.py -d t ... > debug_file" compiler option and you can see all
compiler tables in the debug_file.
Lars Roland wrote:
Hello all,
I tried to apply the h323 part of the voip call analysis patch an now I
get following warnings when I try to regenerate the dissector:
:: UserWarning: The same type names for different types. Explicit type
renaming is recommended.
T_language
T_language Connect-UUIE/language
T_language1 Setup-UUIE/language
T_language2 AdmissionConfirm/language
:: UserWarning: The same field names for different types. Explicit field
renaming is recommended.
language
language1 T_language1 Setup-UUIE/language
language2 T_language2 AdmissionConfirm/language
language T_language Connect-UUIE/language
I didn't get these warnings before and I can't see any difference for
the language items. They are still generated completely from the asn
definition file. There is nothing in the cnf or the template file about it.
What are the suggested "type renaming" and "field renaming" about?
When do I need them?
Another question. Why do we have so many "T_*" functions in generated
dissectors? IMO they introduce an unnecessary function call.
Regards,
Lars
_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev
