|
From: Leif J. <le...@am...> - 2004-01-16 21:49:49
|
Hi all -
I'm having organization thoughts about the ccmtools. Here they are,
numbered for easy reference :
1. We should move the library-type .jar files (MDR/*.jar) into the
directory with the other .jar files (lib/).
2. We should move all of our metamodel .java files into the Metamodel/
directory, so the subdirs of that directory will be BaseIDL,
ComponentIDL, OCL, etc. Can we do this ?
3a. Why don't we make an OCLParser directory and put the OCL*.g files in
there, along with any needed support files from OCL/utils ? Then we
could build that directory just like the IDL3Parser directory, and
not have to keep the OCL{Parser|Lexer|TokenTypes}.java files in CVS.
3b. It looks like we can do the same thing with the uml2idl directory ;
put it in a UMLParser directory, and generate lots of it
automatically during the build process.
In fact, if we build the UMLParser directory first, then we could
use that part of the library to generate the necessary files for the
OCLParser directory. Then we could just keep the OCL.xml and OCL.dtd
files in CVS.
I'm going to keep this one from getting too long. Let me know what you
think ; I'm much in favor of keeping generated files out of CVS, since
they can be generated (duh). In particular, I think the above things
would be easy to implement, and they would help keep the source tree and
namespaces consistent.
leif
--
Leif Morgan Johnson : http://ambient.2y.net/leif/
|