I know, at least with Netscape(Mozilla), the javascript engines now
support setters and getters, so that you can use this.w=xx and it would
be able to call a function to set or retrieve the variable. This
doesn't help with efficiency, and I don't know how well it is supported
in other browsers, but it might be an interesting thing to look at.
Dan Willemsen
(If I find a like referencing setters and getters/their support, i'll
send another email)
On Thu, 2003-03-06 at 16:55, Raymond Irving wrote:
> --- Benoit Marchant <[EMAIL PROTECTED]> wrote:
> > Well, I would keep it. A problem I see in dynapi3 is
> > that the naming
> > conventions of the setter methods and their instance
> > variables isn't
> > consistent. It makes it harder to remember things,
> > particularly the
> > width and height. The instance variable are w and h,
> > but the methods
> > are setWidth/getWidth, so many time I found myself
> > using alayer.width.
> > I guess it's much more efficient in javascript to do
> > this.w = 0, the
> > this.setWidth(0), right ?
>
> It's best to avoid writin to the properties of the
> objects. Please use the available functions when
> setting variables. Using the function setWidth() will
> perform certain checks on the w argument to ensure
> smooth running of the object. when reading values I
> could also suggested that you do the same. For
> example:
>
> the setCursor('hand') will cause a pointer to be
> displayed in ie, ns6+, and opera 7+. If you were to
> use lyr._cursor='hand' it will only work in ie but not
> in ns6,etc. Then same goes for getCursor(). If you
> were to use cur=lyr._cursor in ns6 after calling
> setCursor('hand') the value returned will be 'pointer'
>
>
>
> >Which means that for
> > performance reasons we
> > are (unfortunately) strongly encouraged to call as
> > less function as
> > possible ?
> >
> > Apart from the fact that it makes using the api a
> > little more
> > difficult, it makes it harder to automate setting a
> > value by a
> > dictionary that will be used by a function, like the
> > html property. The
> > method is setHTML, and the property is html. If they
> > were following a
> > consistent naming convention like
> >
> > property -> html
> > setter -> setHtml
> >
> > Then we could do something like the following:
> >
> > DynObject.prototype.takeValuesFromDictionary =
> > function(values) {
> >
> > for(var key in values) {
> > /* This would be a workaround when you want to call
> > a function without
> > relying of an existing naming convention: ex
> > aLayer.takeValuesFromDictionary({setHTML:'test'});
> > if(typeof(this[key]) == "function") {
> > this[key](values[key]);
> > }
> > else {
> > */
> > var setterName = 'set'+
> > key.charAt(0).toUpperCase(); +
> > key.slice(1);
> > if(this[setterName] &&
> > typeof(this[setterName] ==
> > "function")) {
> > this[setterName](values[key]);
> > }
> > else
> >
> > this[key] = values[key];
> > //}
> > }
> > }
> >
> > So the idea is that you first test if an setter
> > exists for this key,
> > and if it does, use it, otherwise affect the
> > instance variable directly.
> > The big win is on ease of use, and compact code,
> > which means less to
> > download.
> >
> > Related to this, I wonder what would be the
> > performance of using the
> > watch method on properties that need additional code
> > when set, like
> > html ? It would greatly reduce the use of functions.
> > Does anyone have
> > tried this from a performance point of view ?
> >
> > Benoit
> >
> > On Thursday, March 6, 2003, at 12:20 PM, Raymond
> > Irving wrote:
> >
> > >
> > > So would you suggest that we keep both methods or
> > just
> > > used the original method?
> > >
> > > Any other comments?
> > >
> > > --
> > > Raymond Irving
> > >
> > > --- Benoit Marchant <[EMAIL PROTECTED]> wrote:
> > >> I ran the 2 versions in IE on the mac, creating
> > 2000
> > >> layers with
> > >> speedtest.dynlayer-basic.html. And it looks like
> > the
> > >> regular method is
> > >> faster, but not all the time. I don't get
> > consistent
> > >> results each time.
> > >>
> > >> Benoit
> > >>
> > >> On Wednesday, March 5, 2003, at 09:20 AM,
> > Raymond
> > >> Irving wrote:
> > >>
> > >>> Hello,
> > >>>
> > >>> Benoit had originally suggested that we use a
> > >>> dictionary object (hash table) to pass
> > parameters
> > >> to
> > >>> DynLayers but that proved must slower than using
> > >>> html,x,y,w,h,color. 9I think I've found a way to
> > >> let
> > >>> the two co-exist:
> > >>>
> > >>> Old way - The original and fastest creation
> > >> method:
> > >>>
> > >>> DynLayer(html,x,y,w,h,bgcolor)
> > >>>
> > >>> New ways - slower but more flexible creation
> > >> methods:
> > >>>
> > >>> DynLayer({html:'Hello',x:100,y:100});
> > >>> DynLayer({zIndex:2,w:100,h:100});
> > >>> DynLayer({textSelectable:true});
> > >>>
> > >>
> > >
> >
> DynLayer({anchor:{top:0,bottom:0},textSelectable:true});
> > >>> DynLayer({textSelectable:false,html:'Cool!'});
> > >>> DynLayer({x:100,y:100,w:10,h:10,visible:false});
> > >>>
> > >>> inline layer - when id is set x,y,w,h,bgcolor,z
> > >> and
> > >>> html will be replaced with values from inline
> > div:
> > >>>
> > >>> DynLayer({id:'myinline'});
> > >>> DynLayer({id:'boxer',anchor:{left:0:right:0}});
> > >>>
> > >>> Optional arguments:
> > >>> x,y,w,h,html,bgColor,zIndex,textSelectable,
> > >>> anchor,visible,id
> > >>>
> > >>>
> > >>> Changes are inside the attached dynlayer_ie.js
> > >> file.
> > >>> Please perform all the speed test and
> > improvements
> > >> you
> > >>> can to see if the new creation method makes any
> > >> sense
> > >>> at all.
> > >>>
> > >>>
> > >>> --
> > >>> Raymond Irving
> > >>>
> > >>>
> > >>>
> > >>>
> > __________________________________________________
> > >>> Do you Yahoo!?
> > >>> Yahoo! Tax Center - forms, calculators, tips,
> > more
> > >>> http://taxes.yahoo.com/
> > >>>
> > >>>
> > __________________________________________________
> > >>> Do you Yahoo!?
> > >>> Yahoo! Tax Center - forms, calculators, tips,
> > more
> > >>> http://taxes.yahoo.com/<dynlayer_ie.zip>
> > >>
> > >
> > >
> > > __________________________________________________
> > > Do you Yahoo!?
> > > Yahoo! Tax Center - forms, calculators, tips, more
> > > http://taxes.yahoo.com/
> > >
> > >
> > >
> >
> -------------------------------------------------------
> > > This SF.net email is sponsored by: Etnus, makers
> > of TotalView, The
> > > debugger
> > > for complex code. Debugging C/C++ programs can
> > leave you feeling lost
> > > and
> > > disoriented. TotalView can help you find your way.
> > Available on major
> > > UNIX
> > > and Linux platforms. Try it free. www.etnus.com
> > > _______________________________________________
> >
> === message truncated ===
>
>
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - forms, calculators, tips, more
> http://taxes.yahoo.com/
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger
> for complex code. Debugging C/C++ programs can leave you feeling lost and
> disoriented. TotalView can help you find your way. Available on major UNIX
> and Linux platforms. Try it free. www.etnus.com
> _______________________________________________
> Dynapi-Dev mailing list
> [EMAIL PROTECTED]
> http://www.mail-archive.com/[EMAIL PROTECTED]/
--
Dan Willemsen <[EMAIL PROTECTED]>
-------------------------------------------------------
This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger
for complex code. Debugging C/C++ programs can leave you feeling lost and
disoriented. TotalView can help you find your way. Available on major UNIX
and Linux platforms. Try it free. www.etnus.com
_______________________________________________
Dynapi-Dev mailing list
[EMAIL PROTECTED]
http://www.mail-archive.com/[EMAIL PROTECTED]/
