Re: [cinjug-users] Constants imply DP (was: Question on last Mond ay's meeting)

["Edward Sumerfield" <esumerfd@xxxxxxxxxxx>]

Agreed, but in this situation we are looking at modeling across languages so 
we have the choice between duplication or bugs.

Whether you are able to model out your constants for objects or not you have 
the same problem.

I also agree that avoiding code generation would be preferable, just not at 
the expense of introducing bugs.

I did think of a another approach. Since I presume that you have a java 
based website where the .js file are also deployed, you can easily parse the 
javascipt name=value format and assign your java constant at startup time.

You can also add validation steps to prevent the duplication from being 
wrong. Make an ant task that compares the two. Make your java init routing 
validate that the javascript has the same value.

Ok, one more solution, make a servlet that is called for .js suffixes and 
render the javascript live with the appropriate constant replacements as it 
goes.

An implementation variation on this would be to put your java script inside 
a jsp page so that it can access your java constants directly and use the 
script tag like this

   <script language="JavaScript" src="myjavascript.jsp"></script>

----- Original Message ----- 
From: "Amol Deshmukh" <adeshmukh@xxxxxxxxxxxxxx>
To: <users@xxxxxxxxxx>
Sent: Friday, February 25, 2005 1:13 PM
Subject: RE: [cinjug-users] Constants imply DP (was: Question on last Mond 
ay's meeting)


> ...that's still a way to deal with the problem using
> a Constants approach. (Not to mention that the fact that
> since MyConstants.java is generated during build,
> you could get into IDE code completion issues when
> accessing your constants)
>
> From what I understand the idea of a non-DP approach is
> to avoid using Constants and instead convey that
> /something/ with a method.
>
> ~ amol
>
> > -----Original Message-----
> > From: Edward Sumerfield [mailto:esumerfd@xxxxxxxxxxx]
> > Sent: Friday, February 25, 2005 12:52 PM
> > To: users@xxxxxxxxxx
> > Subject: Re: [cinjug-users] Constants imply DP (was: Question on last
> > Monday's meeting)
> >
> >
> > Automate. I didn't test it but you get the idea. Code
> > generation rocks.
> >
> > MyConstants.properties
> >     something=somethingelse
> >
> > Generate Code.sh
> >     echo "class MyConstants {" > MyConstants.java
> >     cat MyContants.properties | \
> >         sed -e "s/([^)])=([^)])/String \1 = \"\2\";" >>
> > MyConstants.java
> >     echo "}" >> MyConstants.java
> >
> >     cat MyContants.properties |
> >         sed -e "s/([^)])=([^)])/    var \1 = \"\2\";" > MyConstants.js
> >
> > A more elegant way would be to use ant and use its copy file
> > with filter
> > variables to generate the language specific files from build
> > variables.
> >
> > ----- Original Message ----- 
> > From: "Mark Windholtz" <windholtz@xxxxxxxxx>
> > To: "Jim Cook" <jimcook@xxxxxxxxxxxxx>
> > Cc: <users@xxxxxxxxxx>
> > Sent: Friday, February 25, 2005 12:36 PM
> > Subject: Re: [cinjug-users] Constants imply DP (was: Question on last
> > Monday's meeting)
> >
> >
> > > Jim Cook wrote:
> > >>
> > >> Obviously the logic to produce the total is duplicated
> > (Once in Java and
> > >> the
> > >> other in Javascript)
> > > ...
> > >> Any suggestions on how I could solve this without using constants?
> > >>
> > >
> > >   Another realization is that one reason we all drop back to
> > >   Data Processing is that DP is a common language accessible
> > >   across departments and different programming languages.
> > >
> > >   One solution is find a way to generate the JavaScript from a
> > >   primary Java definition.  Tapestry does some of this, but I'm not
> > >   sure how customizable it is.
> > >
> > >   In your case, I think your stuck with duplication.
> > >   If it's all in Java however, it worth reconciling.
> > >
> > > Regards,
> > > -Mark
> >
> >
> > ---------
> > You may unsubscribe from this mailing list
> > by sending a blank email addressed to:
> > users-unsubscribe@xxxxxxxxxx
> >
> > --
> > Find additional help by sending a blank email
> > addressed to:
> > users-help@xxxxxxxxxx
>
> ---------
> You may unsubscribe from this mailing list
> by sending a blank email addressed to:
> users-unsubscribe@xxxxxxxxxx
>
> --
> Find additional help by sending a blank email
> addressed to:
> users-help@xxxxxxxxxx