[omniORB] omniORB for Python and new IDL compiler pre-release

[Duncan Grisby]

On Wednesday 3 November, Rick Ethington wrote:

[... IDL compiler...]

> looks interesting, and yes i have entertained an idea (showing ideas a
> good time is what i like to do best) that this might support.
> Do comments in the idl get passed to the python back-end? i had been
> looking for an up to date version of idldoc and it had been suggested
> that a new back-end for this idl compiler might be the answer. what do
> you think?

At the moment, comments do not get past the C preprocessor. Most
preprocessors have a switch which tells them not to strip comments,
but I'm not sure that all do. However, what do get through to the
back-end are any #pragmas which are not recognised by the front-end.
This could be used both for annotating IDL to generate extra code, and
for embedding documentation.

One thing which we have considered is writing a simple
pre-preprocessor which converts specially formatted comments into
suitable #pragmas before the C preprocessor is run.

Something which I need to document is how the front-end chooses which
node in the syntax tree it should attach the #pragmas to. Basically,
they are attached to the most recently created node. This means that
to document an operation, you could to do something like:

  interface I {
    void op(
#pragma doc op() is an operation
            in string arg);
  };

Which is ugly, but could easily be generated from something nicer.

Cheers,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --