[pypy-dev] External RPython mailing list
santagada at gmail.com
Thu Sep 16 05:01:50 CEST 2010
If anyone wants to pay developers to work on rpython it should
probably follow this wishlist and not focus on trying to merge
shedskin into pypy (for the all the reasons given above).
Here is my version of the list
On Wed, Sep 15, 2010 at 9:44 PM, Hart's Antler <bhartsho at yahoo.com> wrote:
> RPython Wishlist:
> . documentation
Yes, this should be the highest priority. I could add documentation,
doctests explaining stuff, templates for new modules and other things,
better error messages.
> . iteration over tuples of any length (with any mixed types)
> . overloading __getattr__, __setattr__
> . pickle support, if its limited thats ok.
did anyone ever really needed this? Would be cool I think, but unnecessary.
> . rstruct is incomplete
> . llvm backend, what happened to llvm support?
also I don't know about the utility of this for rpython itself, would
be interesting if there was a jit backend that used llvm to better jit
> . not having to define dummy functions on the base class to prevent 'demotion'
> . not having to use the hack `assert isinstance(a,MySubClass)` to call methods with incompatible signatures.
both are nice to have in a language with type inferences. Without this
explicit in the code maybe the error messages would be even more
complex to deal with.
> . we already have the decorator: @specialize.argtype(1), why can't we have @specialize.argtype(*) so that all arguments can have flexible types?
could be @specilize.argtype(42)
> . methods stored in a list for easy dispatch can not have mismatched signatures.
I would add:
. tool to automatically generate stubs for c libraries
. c++ support (to be able to use c++ libs directly on rpython)
. better java support (because the jvm matters).
. better tooling (profiling, debugging etc)
. separate compilation and/or some cache to speed up compilation.
More information about the Pypy-dev