Posted By:
Stephen_Silber
Posted On:
Sunday, July 22, 2001 02:36 AM
Hi everybody, As I've been cruising through an ANTLR project (in C++, no less--thanks Peter and Ric!), it occurs to me that the anonymous {...} section of an ANTLR class definition, used to put arbitrary members into said class, really needs to be flexible enough to handle languages where the prototype and implementation of a method or function might be found in two separate files. The current code generator works fine, of course, but I am uncomfortable with the fact that any C++ methods I put in there are automatically going to be marked for inlining. I'd love to hear what anyone else thinks about this. Thanks, JSRS
More>>
Hi everybody,
As I've been cruising through an ANTLR project (in C++, no less--thanks Peter and Ric!), it occurs to me that the anonymous {...} section of an ANTLR class definition, used to put arbitrary members into said class, really needs to be flexible enough to handle languages where the prototype and implementation of a method or function might be found in two separate files.
The current code generator works fine, of course, but I am uncomfortable with the fact that any C++ methods I put in there are automatically going to be marked for inlining.
I'd love to hear what anyone else thinks about this.
Thanks,
JSRS
<<Less