Thu, 02 Nov 2000 13:51:55 -0500
"Michael P. Wilson" wrote:
> I'm just far too attached to the fact that it's in perl to use doxy
> (which I tried and didn't like much frankly. But it's probably just
well, what I care most about is the following:
* the ability to parse actual source code, not requiring any specially
* flexibility with respect to both, the input (language, documentation style, ...)
as well as output (generating docs in a variety of formats, generating diagrams, ...)
I had hoped that the 'pyceps' tool which was talked about on this list a year ago, would
actually make it into something useful. Since this didn't happen, I started the 'synopsis'
tool. It's not that far either, but I can process IDL input already, and a C++ parser is
slowly making progress.
Note that the synopsis tool stores an AST (Abstract Syntax Tree) internally. This can
be used in quite a couple of circumstances, not merely to generate textual documentation.
While doxygen seems to be able to parse a number of languages, the code (which I looked into
a while ago) isn't nicely modular, i.e. it isn't easy to extend.
Synopsis uses plugins for the parsers as well as the formatters. Have a look at
(and it is written in python, which might be of some interest, considering the actual
mailing list address :)
Best regards, Stefan
PS: since my main hobby horse is the berlin project, my main interest into synopsis was
and is to be able to generate reference manuals for the berlin code (a mixture of IDL
and C++). But I'm sure that the scope of the project is of interest to a much larger
Departement de Physique
Universite de Montreal
...ich hab' noch einen Koffer in Berlin...