(I am copying this to the photran-dev mailing list for the record --
Naoki and Matthijs (Fujitsu) and Jay (NCSA) are interested in
contributing to parts of Photran.)
I promised to send a summary of where Photran is at the moment and
where we should probably be headed. Some of this is duplicated from a
previous post I sent to photran-dev.
Here is what we're doing with Photran at the moment.
- I am replacing our parser with OFP, the Open Fortran Parser
(developed at LANL), and building the rewritable AST for it. I will
also be making several modifications for scalability. The majority of
this should be done within a couple of months.
- Students in Ralph's software engineering class at UIUC are
building a number of enhancements, mostly related to the user
interface. The class ends in December, at which point I will review
their projects and decide which ones can be integrated into Photran.
I am still hoping we can have an "official" release sometime soon,
perhaps Spring/Summer 2008. Analysis/refactoring in Photran is largely
still a toy... I want our official release to be distinguished by the
fact that (1) it has a useful set of analysis-related features, i.e.,
it's not just a 200 meg text editor, and (2) these features scale for
large projects. I think we are actually quite close to making this
happen. Before an "official" release...
- After I review the student projects in December, the viable
projects need to have their code integrated into HEAD (and tested
thoroughly). There may be some conflicts (e.g., since several groups
are working on editor enhancements), so this is more than simply
applying patches, but it shouldn't be too hard.
- The parser, rewriting engine, and indexer need to be
scalability-tested. Ideally, the OFP tests and several large Fortran
codes need to be added to the test suite. This could actually be done
now, before I finish OFP integration; they would simply be testing our
existing parser.
- C preprocessor support (based on the CDT) should be integrated
into
Photran. Again, this shouldn't be too hard, but it's necessary for
scalability testing and for analysis/refactoring to be usable on real
projects.
- Photran's Managed Build System plug-ins need to be looked at; I
think some of it still needs to be updated to integrate better with CDT
4.
- I think Naoki had mentioned integrating Lahey/Fujitsu Fortran
into the Managed Build System... that would also be a great addition
before the release.
After the release, I will be focused entirely on program
analysis/refactoring and Fortran 2003/2008 support. There are lots of
directions the project can head after that. I think C/C++/Fortran
cross-language issues is an important one. We continue to have a slew
of complaints about the debugger, but I
don't have any sort of "strategic vision" for how to deal with those.
I sent an e-mail to Naoki and Matthijs earlier. Jay, what are you most
interested in? Actually, any of the five bullets above could serve as
a nice introduction to the Photran codebase... and if you're interested
in working on C/C++/Fortran cross-language support (esp. search) longer
term, I think there is a lot of interest in that. Of course, if you
have your own ideas, those are more than welcome as well!
Jeff
|
