Thank you for your answer, code examples and useful links.
I know about module patter and saw it in many places but still never
used it by myself.
So it will solve two of my problems: private method and referencing
window object.
Question about module patter: is it only one solution? After ES5 any
other patterns appeared that can solve similar problem with private
members?
One point that I like in Object.defineProperty - is better control.
E.x. if my method isn't writable than it cannot be override like
fs.prototype.readAsText=function(..){
//new implementation
}
In case of module patter I think I don't have such control. E.x. I
wan't people extend my fs object but doesn't change any of the
methods. What technique to use for this?
In what cases you use Object.defineProperty()?
Question about testing remain: do you have any suggestions / links how
to better do this?
P.S. Also I didn't restrict myself to use ES5, since main use of this
code in Chrome Web App that can be used using Chrome and I think
Chrome supports ES5. But I'll try to make it work in another browsers
with File API support.
Best regards,
Anton Podviaznikov
>> if (global.requestFileSystem)
I understand about feature detection. Thank you!
On Feb 1, 3:42 am, RobG <[email protected]> wrote:
> On Feb 1, 3:32 am, Anton Podviaznikov <[email protected]> wrote:
>
> > Hello everybody. This is my first post in this group.
>
> > I'm learning JavaScript and as almost everybody decided to write small
> > library. My library is kind of wrapper for the HTML (5) File API.
>
> The opening lines of fs.js:
>
> | "use strict";
> | var fs = Object.create({},
>
> I don't know why you chose ECMAScript ed 5, but it's not well
> supported. In Firefox 3.6:
>
> alert(typeof Object.create); // undefined
>
> Are you using features that require ECMAScript ed 5? You can emulate
> private members using the module pattern, rather than relying on
> underscores to indicate them.
>
>
>
> > It is not complete and it's not for production. I used it just in
> > Chrome Web App.
>
> > I'm interested in some feedback (not for the library itself) for the
> > code. If you can quickly point me to some big error or make
> > suggestions it would be very helpful for me.
>
> > Also I have few questions:
> > 1. I wanted few of my method be private. I know there are some
> > patters, but what is the best in my case? Do you have any quick link?
>
> I think the module pattern is the most commonly used:
>
> <http://yuiblog.com/blog/2007/06/12/module-pattern/>
>
> It is often attributed to Douglas Crockford, however it was developed
> primarily by Richard Cornford and others. It is great if you only need
> one of an object, but it can also be used to create a constructor.
>
> > 2. I extended native FIle and FileError objects. Is it normal? Is it
> > good style? How you will do this?
>
> It is generally not a good idea to extend built-in objects. Is there a
> significant advantage to doing it? If not, don't.
>
> > 3. What is the better way to write tests for these kind of libs.
> > 4. Is it ok that I reference to global window object?
>
> It is considered better to establish an alias for the global object
> and use that, e.g.
>
> var global = this;
> var fs = (function() {
>
> // Private members
> var _getNativeFS = {
> value: function(callback) {
> if (global.requestFileSystem) {
> global.requestFileSystem(window.PERSISTENT,
> this.maxSize, function(fs) {
> ...
> });
> } else {
> // deal with not having global.requestFileSystem
> }
> }
> };
>
> ...
>
> return {
>
> // Public members
> log:
> {
> value: false,
> writable: true
> },
> ....
>
> })();
>
> There are many resources on how to write libraries, fundamental
> studies are the module pattern and feature or capability detection.
> Peter Michaux[1] has blogged about quite a bit of it, so have others.
>
> 1. <URL:http://michaux.ca/>
>
> --
> Rob
--
To view archived discussions from the original JSMentors Mailman list:
http://www.mail-archive.com/[email protected]/
To search via a non-Google archive, visit here:
http://www.mail-archive.com/[email protected]/
To unsubscribe from this group, send email to
[email protected]
