[Python-Dev] Python and compilers

[Michael Foord]

On 06/04/2010 23:34, Michael Foord wrote:
> On 06/04/2010 23:31, "Martin v. Löwis" wrote:
>> willian at ufpa.br wrote:
>>> First, thank you for all opnion. Each one was considered.
>>> I think the better question would be:
>>> I have to develop a project that involves compilers, and being a fan of
>>> Python, I thought about making a compiler for it (most basic idea 
>>> involving
>>> Pythin and compilers). But I saw that I can use what I learned from
>>> compilers not only to create a compiler. What is the area of 
>>> developing the
>>> Python interpreter that I could build my project, and please give me
>>> interesting ideas for the project.
>> I don't think the question is necessarily off-topic.
>>
>> I can propose two projects, related to Python core:
>>
>> - 2to3 pattern compiler: 2to3 currently uses an interpreter for pattern
>>    matching. It does "compile" the patterns into some intermediate form,
>>    however, that is actually interpreted with an interpreter written in
>>    Python (actually, it is self-interpreted). It might be interesting to
>>    compile the pattern into actual Python code, with the hope of it
>>    executing faster than it does now (and yes, I proposed a similar, but
>>    different GSoC topic also; the two approaches are orthogonal, 
>> though).
>>
>> - IDLE code completion. Currently, IDLE has some form of code
>>    completion, which is fairly limited. It might be useful to produce a
>>    better code completion library, one that works more statically and
>>    less based on introspection. In particular, optimistic type inference
>>    might help (optimistic in the sense "if foo has a method .isalpha, it
>>    probably is a string). In code completion, exact type inference isn't
>>    necessary; giving a superset (i.e. a union type) might still be
>>    helpful.
>>
>
> This would be very useful to many Python IDE projects. Getting it 
> right is one thing, getting it fast enough to be useful is another 
> (i.e. it is a difficult problem) - but yes, could be both interesting 
> and useful.

A good basis for a project like would be PySmell (by Orestis Markou), 
which intended to provide this but was never completed:

     http://code.google.com/p/pysmell/

It would make a great gsoc project as well.

Michael

>
> Michael
>
>
>> In addition, to-python compilers may also be interesting in various HTML
>> templating languages, e.g. Django templating, Zope page templates, and
>> the like (although some of them already have compilers of some form on
>> their own).
>>
>> HTH,
>> Martin
>> _______________________________________________
>> Python-Dev mailing list
>> Python-Dev at python.org
>> http://mail.python.org/mailman/listinfo/python-dev
>> Unsubscribe: 
>> http://mail.python.org/mailman/options/python-dev/fuzzyman%40voidspace.org.uk 
>>
>
>


-- 
http://www.ironpythoninaction.com/
http://www.voidspace.org.uk/blog

READ CAREFULLY. By accepting and reading this email you agree, on behalf of your employer, to release me from all obligations and waivers arising from any and all NON-NEGOTIATED agreements, licenses, terms-of-service, shrinkwrap, clickwrap, browsewrap, confidentiality, non-disclosure, non-compete and acceptable use policies (”BOGUS AGREEMENTS”) that I have entered into with your employer, its partners, licensors, agents and assigns, in perpetuity, without prejudice to my ongoing rights and privileges. You further represent that you have the authority to release me from any BOGUS AGREEMENTS on behalf of your employer.