Re: [Pydev-code] Autocompletion in PyDev. A first working version
Brought to you by:
fabioz
Menu
▾
▴
Re: [Pydev-code] Autocompletion in PyDev. A first working version
|
From: Dana M. <dan...@ya...> - 2004-06-18 17:36:49
|
Parhaum,
is not a problem at all. I will take a crack at it.
best,
D
--- Parhaum Toofanian <pto...@ho...> wrote:
> To be honest, I'm still just trying to figure out how things are
> working as
> they are. I'm suffering from not having experience with Python,
> interpreters, or Eclipse, so it's slow going.
>
> - Parhaum
>
> ----- Original Message -----
> From: "Dana Moore" <dan...@ya...>
> To: <pyd...@li...>
> Sent: Friday, June 18, 2004 10:24 AM
> Subject: RE: [Pydev-code] Autocompletion in PyDev. A first working
> version
>
>
> > Fabio, Good comments.
> >
> > I guess this means that we should work toward a user selectable
> > autocompletion. We can build upon what we have now of course as an
> > initial cut, and we can experiment with spawning a CPython
> interpreter
> > to run the same code introspection package. Is that how everyone
> else
> > reads it?
> >
> > Parthoum, if you would prefer look into how we might spawn a
> CPython
> > interpreter and run the same introspection code we are using now
> (which
> > is in the jythonlib.jar), I could actually move into trying the new
> > project wizard. Can you let the group know if you'd like to attempt
> > that effort as soon as you can, so that we can adjust our planning
> > accordingly.
> >
> > Parthoum, I am looking at your thesis features list, and I would
> like
> > to help you get as much of tht done as possible; I notice that
> > autocomplet is on it as well, so maybe it's a good plan for you to
> get
> > involved in that way. By the way, we still need to code
> autocompletion
> > for an opening round brace "(", I can look at that, in the
> > PyCompletionProcessor code. The introspector I based this on
> _should_
> > handle that case; I will find out why it's not :-)
> >
> > Regards,
> >
> > --- Fabio Zadrozny <fa...@in...> wrote:
> > > Hi Dana,
> > >
> > > Well, I don't have PyAlaMode or PyAlaCarte installed (but I use
> PyDev
> > > full
> > > time at work)...
> > >
> > > Anyway, I don't think I have to use them in order to demonstrate
> that
> > > python
> > > gets c++ modules definitions...
> > >
> > > I have a c++ module called _coilib (It is actually a dll that has
> an
> > > interface with boost python).
> > >
> > > When I import it and call dir(_coilib), it returns to me all the
> > > functions
> > > the dll has without any problems (we would have to spawn a
> process
> > > and call
> > > dir on the module to get the completion suggestions).
> > >
> > > And, well, thinking in the other way, if I'm developing with
> Jython
> > > and not
> > > in CPython, I would like to import java.awt and see the code
> > > completion for
> > > that, so, I think, that having an option to develop in Jython or
> in
> > > CPython
> > > would be the ideal thing...
> > >
> > > []s
> > >
> > > Fabio
> > >
> > > -----Original Message-----
> > > From: pyd...@li...
> > > [mailto:pyd...@li...] On Behalf Of Dana
> > > Moore
> > > Sent: quinta-feira, 17 de junho de 2004 13:25
> > > To: pyd...@li...
> > > Subject: Re: [Pydev-code] Autocompletion in PyDev. A first
> working
> > > version
> > >
> > > --- Aleks Totic <a...@to...> wrote:
> > > > Hi,
> > > >
> > > > I think Fabio's problem is that his python code requires Python
> > > > version > 2.3. When Jython is used by pydev to parse this file,
> > > > it fails, and autocompletion does not work.
> > > >
> > >
> > > Aleks, that what I thought might be the problem at first; then I
> > > considered that Jython is merely running a bit of CPython code
> > > (actually developed on 2.3.x) So I then wondered whether his
> argument
> > > was valid or not. I mean: I understand that the jython
> interpreter
> > > itself is not 2.3 compliant but why should that matter if the
> > > intrspection code _is_ -- That is is there anything from the
> jython
> > > runtime system that prevents us from doing as good a job as (say)
> > > PyAlaMode. I noted the one thing which is Fabio's suggested
> little
> > > kludge to exec "class object:pass", but is there anything else???
> > > That's why I am asking Fabio to try out his part python/part C++
> code
> > > in somethign like PyAlaMode or PyAlaCarte and let us know.
> > >
> > > > Conversely, if his code required Jython to run, autocompletion
> > > > based upon CPython interpreter would fail to work. So no one
> > > > solution would please both Jython & CPython developers.
> > > >
> > > > What's the right solution?
> > > >
> > > > If scripts can run in both Python & Jython environment, then
> the
> > > > user could have an option to pick one they are using. This
> route
> > > > involves additional work: You'd have to communicate with Python
> > > > interpreter through stdio or sockets. Debugger does this, and
> > > > almost works with Jython. Not quite, since its threading model
> is
> > > > different.
> > > >
> > > > Or different features can use different interpreters, and they
> > > > just won't work for the other guy. This might sort itself out
> as
> > > > Jython catches up with CPython. If this is what we do, make
> sure
> > > > that you are clearly communicating to users what is happening.
> So
> > > > far:
> > > >
> > > > - debugger will work with CPython
> > > > - auto-completion will work with Jython
> > > >
> > > > the users might get confused.
> > > >
> > > > This is where doing an IDE in a foreign language (Java) gets to
> > > > be tricky.
> > > >
> > > > Good luck,
> > > >
> > > > Aleks
> > > >
> > > > > Is Jython's failure to be fully version compliant a problem?
> If
> > > > we're
> > > > > just running a module to introspect code (which was actually
> a
> > > > direct
> > > > > port from CPython) then does Jython's version skew actually
> > > present
> > > > a
> > > > > problem? If it does, then perhaps one that can be worked
> > > > separately;
> > > > > perhaps Kevin Altis might have a comment or two on this
> topic.
> > > > >
> > > > >
> > > > >>So, basically, what I want to say is that it is good the way
> it
> > > is
> > > > >
> > > > > for Jython development, but when it comes to develop in
> python, I
> > > > think
> > > > > that jython just hasn't arrived there yet... (still version
> 2.1)
> > > > and
> > > > > for some things like dependency on c++ modules, it probably
> > > never
> > > > will
> > > > > (and this has never been its purpose).
> > > > >
> > > > > Correct statements, both. Is there are better way to invoke
> > > Python
> > > > > introspection directly from Eclipse? I suppose we might spawn
> a
> > > > > separate command line thread "python introspector.py" and
> pipe
> > > the
> > > > > result to standard out or a temp file and then read that
> stream?
> > > >
> > > > >
> > > > >
> > > > >>Of course, we could let the user choose what to use, if we
> want
>
=== message truncated ===
=====
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
_/ Dana Moore _/
_/ BBN Technologies LLC _/
_/ M: 240.350.4196 _/
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
__________________________________
Do you Yahoo!?
Yahoo! Mail - Helps protect you from nasty viruses.
http://promotions.yahoo.com/new_mail
|
