When I tried it, "./getddl.py -f -F /crypt/funcs -d glue", I got the error "No such file or directory: 'sql/schemas.sql'".
While the task of splitting objects into separate files could be solved by an external "wrapper tool" like yours around pg_dump,
I argue it makes more sense of putting the (minimal required) logics into pg_dump, due to a number of reasons, most importantly because it's simplier and less complex, thus less error prone.
My patch is only a few lines of code and doesn't add any logics to pg_dump, it merely reroutes the fwrite() system calls based on the toc entries.
Just the fact you and others had to create own tools to do the splitting shows the feature is important, which I think should be included in the normal pg_dump tool.
As someone whose own version of "getddl" helped inspire Dimitri to create his own version, I've both enjoyed reading this thread and seeing this wheel reinvented yet again, and wholeheartedly +1 the idea of building this directly into pg_dump. (The only thing better would be to make everything thing sql callable, but that's a problem for another day).
