You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Language generators currently fall into three categories:
pure templated generators, which use the base generator impl. to render a single template
python templated generators, which select the generator based on the python version, but ultimately generates a single template
go templated generators, which generate multiple outputs (the CGo file and the Go file)
Each generator should be promoted to a fully-fledged class that knows everything it needs to generate the bindings for its language. This should wrap up the current filters, which perform type transformations, and offer customisation points for arbitrary types, which we will match either by type name (boost::gregorian::date) or by template-name and arguments (std::optional<std::string>).
The text was updated successfully, but these errors were encountered:
The first step, required by this ticket, is to provide the existing FFIG functionality through the new class-based approach. Once this is in place, we can add the customisation point logic.
Language generators currently fall into three categories:
Each generator should be promoted to a fully-fledged class that knows everything it needs to generate the bindings for its language. This should wrap up the current filters, which perform type transformations, and offer customisation points for arbitrary types, which we will match either by type name (
boost::gregorian::date
) or by template-name and arguments (std::optional<std::string>
).The text was updated successfully, but these errors were encountered: