I'm looking to follow the "top-down" (WSDL-first) service development approach, 
as commonly recommended (and for various project-specific reasons I'd be happy 
to share).

However, the .h created by wsdl2h is (as documented) not a viable input to 
other compilers (stray characters, etc.).  What is the recommended process to 
provide function declarations and data definitions to implementers when using 
gSOAP to build a web service?

For example, I initially thought I would:
1. agree on the WSDL between platform teams
2. provide the WSDL to the web service team
3. that team generates .h with wsdl2h
4. that team generates network server implementations with soapcpp2
5. that team implements functions declared in the .h to be called by the 
"skeleton" server implementations

That separates generated/read-only sources from custom implementation to reduce 
patch/maintenance effort, and the compiler ensures adherence to the spec during 
implementation and revisions.

What is the best-practice approach here?


Shannon

Reply via email to