Re: [morph-developer] appendDefaultComponents WAS Re: Copying Objects to Maps
Brought to you by:
orangeherbert,
sgarlatm
Menu
▾
▴
Re: [morph-developer] appendDefaultComponents WAS Re: Copying Objects to Maps
|
From: Matt B. <gud...@ya...> - 2007-02-28 16:28:37
|
Yeah, I thought about that, but wouldn't that be kind
of expensive having to add the default converters
every time? Especially given the caveat that we're
probably best off always calling
createDefaultComponents() and thus instantiating new
component converters every time (because nothing would
stop a user from making changes to converters obtained
from getConverters() if we were to e.g. use statically
defined components). Plus we still then have to make
the unilateral decision that a cloned SDT instance has
appendDefaultConverters set to false to make it work
right with the Composite framework, despite the fact
that a user-requested clone() operation probably wants
a true clone, appendDefaultComponents and all... you
know?
-MJB
--- Matt Sgarlata
<Mat...@wh...> wrote:
> Thinking further about option #1, how about this
> logic
>
> class SDT {
> boolean appendDefaultComponents=true;
> ...
> getComponents() {
> if (appendDefaultComponents) {
> return concat(components,
> getDefaultComponents());
> }
> else {
> if (isEmpty(components))
> throw new MorphException("No components
> found");
> return components;
> }
> }
>
> Matt S
>
> Matt Benson wrote:
> > On option 1 below: first issue: it seems that it
> > would be difficult to differentiate when to add
> what,
> > especially when no user-supplied components are
> > provided. second issue: it seems like this would
> > still cause problems with composites, wouldn't it?
> :(
> >
> > -MJB
> >
> > --- Matt Sgarlata
> > <Mat...@wh...> wrote:
> >
> >
> >> That sounds like a reasonable approach.
> >>
> >> Here is some brainstorming on other approaches:
> >> 1) Modify the getComponents method instead of the
> >> setComponents method
> >> to code the behavior of appendDefaultComponents
> >> 2) Introduce a new property like
> >> nonDefaultComponents that must be used
> >> when you want to appendDefaultComponents.
> >> Configuration is then done
> >> either with appendDefaultComponents = true and
> >> modifications done to
> >> nonDefaultComponents or when
> >> appendDefaultComponents=false directly to
> >> the components property.
> >>
> >> #2 definitely seems like a kludge and not a very
> >> good solution to me.
> >> Did you try out #1?
> >>
> >> Matt S
> >>
> >> Matt Benson wrote:
> >>
> >>> I had implemented this for
> >>> SimpleDelegatingTransformer, was working on
> >>> duplicating the effort for
> >>>
> >> SimpleDelegatingReflector,
> >>
> >>> and managed to break several test cases. I was
> >>> setting a flag and--if true--appending default
> >>> components every time components were set via
> >>> setComponents(). Debugging made it obvious that
> >>>
> >> this
> >>
> >>> completely circumvents the point of having these
> >>> classes be composites; i.e. they are no longer
> >>> specializable. So I'd like to discuss what the
> >>> options are for enabling this simplified
> >>>
> >> configuration
> >>
> >>> without breaking existing specialization
> >>> (SimpleDelegatingTransformer should in theory
> >>>
> >> suffer
> >>
> >>> from the same situation, only none of the core
> >>>
> >> code
> >>
> >>> happens to specialize an SDT):
> >>>
> >>> I think in order to do this safely the easiest
> >>>
> >> thing,
> >>
> >>> at least to start with, is to simple provide a
> >>> constructor that allows the user to specify the
> >>> component list and accepts
> appendDefaultComponents
> >>>
> >> as
> >>
> >>> a boolean parameter on a one-time-only basis. I
> >>>
> >> have
> >>
> >>> made this change to SimpleDelegatingReflector
> and
> >>>
> >> will
> >>
> >>> go ahead and modify SimpleDelegatingTransformer
> >>> similarly. This at least provides for the
> >>>
> >> behavior at
> >>
> >>> the cost of forcing the user to use constructor
> >>> injection (assuming Spring config). Perhaps we
> >>>
> >> will
> >>
> >>> then be able to see better how to proceed.
> >>>
> >>> -Matt
> >>>
> >>> --- Matt Sgarlata
> >>> <Mat...@wh...> wrote:
> >>>
> >>>
> >>>
> >>>> Matt Benson wrote:
> >>>>
> >>>>
> >>>>> Hmph--hadn't noticed the Spring file. :)
> This
> >>>>>
> >> is
> >>
> >>>>> something I would like to think more about for
> >>>>>
> >> the
> >>
> >>>>> long-term. I've been looking at this stuff
> for
> >>>>> several hours today (and once more revisited
> the
> >>>>> UNfruitful path of having SDT flat-out
> implement
> >>>>> NodeCopier, though I almost got there) and now
> I
> >>>>>
> >>>>>
> >>>> think
> >>>>
> >>>>
> >>>>> a tiny piece of effort that would have quite a
> >>>>>
> >> bit
> >>
> >>>>>
> >>>>>
> >>>> of
> >>>>
> >>>>
> >>>>> yield would simply be a boolean
> >>>>> appendDefaultComponents flag for SDT. This
> >>>>>
> >> would
> >>
> >>>>> greatly simplify Spring (or otherwise)
> >>>>>
> >>>>>
> >>>> configuration
> >>>>
> >>>>
> >>>>> of an application's top-level SDT where the
> >>>>>
> >>>>>
> >>>> developer
> >>>>
> >>>>
> >>>>> provides custom transformer implementations,
> but
> >>>>>
> >>>>>
> >>>> wants
> >>>>
> >>>>
> >>>>> to defer to Morph's default behavior for
> >>>>>
> >> standard
> >>
> >>>>> cases. Hopefully this would cover >= 90% of
> >>>>>
> >> what
> >>
> >>>>> developers would want to do, while still
> leaving
> >>>>>
> >>>>>
> >>>> the
> >>>>
> >>>>
> >>>>> door open for more complex configurations.
> >>>>>
> >>>>> "Make easy things easy and hard things
> >>>>>
> >> possible."
> >>
> >>>>>
> >>>>>
> >>>>>
> >>>> Yes, it is a huge pain in the neck to customize
> >>>>
> >> the
> >>
> >>>> SDT because of the
> >>>> need to exactly duplicate the default
> >>>>
> >> transformers
> >>
> >>>> that come with Morph
> >>>> and make sure the custom transformers come
> before
> >>>> the built-in ones. I
> >>>> hadn't really considered adding a boolean flag
> >>>>
> >> like
> >>
> >>>> this, and I do agree
> >>>> it would be a very minor change with a very
> large
> >>>> benefit. I would also
> >>>> be in favor of the default value for the flag
> to
> >>>>
> >> be
> >>
> >>>> true. I believe
> >>>> SimpleDelegatingReflector would also benefit
> from
> >>>>
> >> a
> >>
> >>>> similar change.
> >>>>
> >>>> Matt S
> >>>>
> >>>>
> >>>>
> >>>>
> >
>
-------------------------------------------------------------------------
> >
> >>>
> >>>
> >>>> Take Surveys. Earn Cash. Influence the Future
> of
> >>>>
> >> IT
> >>
> >>>> Join SourceForge.net's Techsay panel and you'll
> >>>>
> >> get
> >>
> >>>> the chance to share your
> >>>> opinions on IT & business topics through brief
> >>>> surveys-and earn cash
> >>>>
> >>>>
> >>>>
> >
>
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> >
> >>>
> >>>
> >>>> _______________________________________________
> >>>> morph-developer mailing list
> >>>> mor...@li...
> >>>>
> >>>>
> >>>>
> >
>
https://lists.sourceforge.net/lists/listinfo/morph-developer
> >
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >
>
____________________________________________________________________________________
> >
> >>> Do you Yahoo!?
> >>> Everyone is raving about the all-new Yahoo! Mail
> >>>
> >> beta.
> >>
> >>> http://new.mail.yahoo.com
> >>>
> >>>
> >>>
> >
>
-------------------------------------------------------------------------
> >
> >>> Take Surveys. Earn Cash. Influence the Future of
> >>>
> >> IT
> >>
> >>> Join SourceForge.net's Techsay panel and you'll
> >>>
> >> get the chance to share your
> >>
> >>> opinions on IT & business topics through brief
> >>>
> >> surveys-and earn cash
> >>
> >
>
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> >
> >>> _______________________________________________
> >>> morph-developer mailing list
> >>> mor...@li...
> >>>
> >>>
> >
>
https://lists.sourceforge.net/lists/listinfo/morph-developer
> >
> >>>
> >>>
> >
>
-------------------------------------------------------------------------
> >
> >> Take Surveys. Earn Cash. Influence the Future of
> IT
> >> Join SourceForge.net's Techsay panel and you'll
> get
> >> the chance to share your
> >> opinions on IT & business topics through brief
> >> surveys-and earn cash
> >>
> >>
> >
>
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV>
> > _______________________________________________
> >
> >> morph-developer mailing list
> >> mor...@li...
> >>
> >>
> >
>
https://lists.sourceforge.net/lists/listinfo/morph-developer
> >
> >
> >
> >
> >
> >
>
____________________________________________________________________________________
> > Cheap talk?
> > Check out Yahoo! Messenger's low PC-to-Phone call
> rates.
> > http://voice.yahoo.com
> >
> >
>
-------------------------------------------------------------------------
> > Take Surveys. Earn Cash. Influence the Future of
> IT
> > Join SourceForge.net's Techsay panel and you'll
> get the chance to share your
> > opinions on IT & business topics through brief
> surveys-and earn cash
> >
>
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> > _______________________________________________
> > morph-developer mailing list
> > mor...@li...
> >
>
https://lists.sourceforge.net/lists/listinfo/morph-developer
> >
> >
>
> >
-------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get
> the chance to share your
> opinions on IT & business topics through brief
> surveys-and earn cash
>
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV>
_______________________________________________
> morph-developer mailing list
> mor...@li...
>
https://lists.sourceforge.net/lists/listinfo/morph-developer
>
____________________________________________________________________________________
No need to miss a message. Get email on-the-go
with Yahoo! Mail for Mobile. Get started.
http://mobile.yahoo.com/mail
|
