If you want to setup / tear-down, use with;
so, if a is an instance of a class, Aclass:
with Aclass() as a:
b(a)
a.member_b()
# now Aclass object is freed, and any cleanup or passed on exceptions which
were raised and not handled are here.
On Mon, Jul 13, 2009 at 7:36 PM, mdipierro <mdipie...@cs.depaul.edu> wrote:
>
> Say I do
>
> with a:
> b()
>
> Perhaps I am mistaken but I do not see a way for b to refer to a. Say
> a contains a counter. I would like b() to increment that counter. How
> do I do it without explicitly passing a to b?
>
> Massimo
>
> On Jul 13, 7:24 pm, Yarko Tymciurak <yark...@gmail.com> wrote:
> > Re: Proposed syntax - I might agree; it is just a request for syntax (I
> to
> > don't see the motivation, other than - as Joe says - sytactic sugar;)
> >
> > Re: with statement - I think it has uses, and I can see immediately one
> > place I would put it in, and could probably find a few more places where
> it
> > would be of good use.
> >
> > On Mon, Jul 13, 2009 at 7:15 PM, mdipierro <mdipie...@cs.depaul.edu>
> wrote:
> >
> > > I repeat. I do not see how it would be possible to implement the
> > > proposed syntax without major changes in the definition of the helpers
> > > and that would be slow when compared with the current implementation.
> > > Perhaps I am wrong.
> >
> > > Whether or not this is a good idea, if I am wrong, I would like to see
> > > a patch that implements this for the DIV helper.
> >
> > > Massimo
> >
> > > On Jul 13, 7:07 pm, Yarko Tymciurak <yark...@gmail.com> wrote:
> > > > I think you are both missing important point:
> >
> > > > classes which support context would be able to be used in "with"
> > > contexts";
> > > > Nothing would be more complex or slower - you'd only use it to make
> > > things
> > > > cleaner (move code into class _enter_() and _exit_() methods, which
> would
> > > > only get called (I presume) on with use anyway... If you have the
> code
> > > in
> > > > one place instead of scattered, there is nothing slower or more
> complex -
> > > to
> > > > the contrary.
> >
> > > > I would agree with you that this could get overused in places (e.g.
> if it
> > > > were used as "syntactic sugar") - but that is a different story.
> >
> > > > cvs reader and writer - I had trouble (had to "hack" a fix) with this
> in
> > > the
> > > > registration system; I never provided a patch, because could not see
> a
> > > > good, clean way to decouple application setup from gluon --- THIS is
> a
> > > way
> > > > to do that.
> >
> > > > I'm pretty sure if I looked, I'd find a few (not many maybe, but a
> few)
> > > > places this would be a good idea, solve existing problems.
> >
> > > > I think both of you need to just consider this a little more
> carefully,
> > > > that's all...
> >
> > > > - Yarko
> >
> > > > On Mon, Jul 13, 2009 at 7:00 PM, mdipierro <mdipie...@cs.depaul.edu>
> > > wrote:
> >
> > > > > I agree with Joe. It require making object context dependent whoch
> > > > > would make things unnecessarily more complex and slower.
> >
> > > > > On Jul 13, 6:42 pm, Joe Barnhart <joe.barnh...@gmail.com> wrote:
> > > > > > I see no benefit to using PEP343 just to provide syntactic sugar
> for
> > > > > > this purpose. There is no REASON to use the "with" statement
> because
> > > > > > we're not doing anything with exception handling here. It's only
> > > > > > being used to create an input format that looks prettier to some
> > > eyes.
> >
> > > > > > Am I missing something??
> >
> > > > > > On Jul 13, 2:40 pm, Yarko Tymciurak <yark...@gmail.com> wrote:
> >
> > > > > > > There are 2 interesting aspects from Luis's blogger post:
> >
> > > > > > > - to use with, classes need to consider contexts;
> > > > > > > - the (to me, at least) interesting example of ASP.NET's
> xmlwriter
> > > > > class
> > > > > > > (generating SVG dynamically) is... .NET specific; I'm off busy
> > > looking
> > > > > for
> > > > > > > a python portable idiom to use in place of his last example...
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"web2py Web Framework" group.
To post to this group, send email to web2py@googlegroups.com
To unsubscribe from this group, send email to
web2py+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/web2py?hl=en
-~----------~----~----~----~------~----~------~--~---
