Kai, thanks for your comments! Both of the things you suggest (a .java
prettifier and a Java->NetRexx converter) are something that I'd very much like
to work on myself -- but just don't have the time.
On specific questions:
-- formatting comments from one style to another is always difficult, as
people are not consistent (for example, when is a block comment introducing
something that follows, and when is it a footnote on the preceding code?). If
you can send me a concise example of something that doesn't work well, I'll see
if I can build it in to the current algorithm.
-- the generated code uses fully-qualified classnames as that's the only way
in Java to assure the meaning. (Short names change meaning over time as
classes and packages are added, and Java doesn't warn of ambiguities as NetRexx
does.) The compiler could possibly have an option to make it use short names
rather than full names in generated code.
-- I am currently working on removing some redundancies (extra curly braces,
and unused loop labels).
-- ditto for removing dollar signs -- one possibility is only using the dollar
signs for temporaries and local variables whose name is the same as a Java
keyword. This will expose generated code to new (added) Java reserved words --
so again a new option might be best.
-- a post-processor might well be a much better way of doing this rather than
by increasing the size and complexity of the compiler .. are you perhaps going
to write one? I could use it myself...
(PS: I have two articles on NetRexx from German magazines in my files... :-)
- - - - -
Mike Cowlishaw, IBM Fellow, IBM UK Laboratories
mailto:
[hidden email] [
http://www2.hursley.ibm.com]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To unsubscribe from this mailing list ( ibm-netrexx ), please send a note to
[hidden email]
with the following message in the body of the note
unsubscribe ibm-netrexx <e-mail address>
Locked
