in _load_storage the line: _storage_service.jStorage = "{}";

throws Error: QUOTA_EXCEEDED_ERR: DOM Exception 22

This line could be wrapped in a try catch, but I think it would be better to do this sooner when setting the storage engine so that an alternate mechanism could be used.

Very odd problem, and everything works fine in every other situation I've tested in. Here's what's happening...

1. Visit page in Safari in iPad simulator.
2. Save some data via jStorage.
3. Refresh page in Safari.
4. Load data via jStorage.
5. Clear data via jStorage.flush().
6. Refresh page in Safari.
7. Load data via jStorage. The saved data returns.

The very, very odd thing is that in between steps 5 and 6, if you attempt to load data from storage you can see that the data was successfully cleared. It seems like the iPad version of Safari requires a little extra magic to fully ditch the data on a flush.

Hi!

I've just packaged jstorage into debian package (and uploaded it into debian repo), so there are some files that can be used to build .deb for the following versions of jstorage.

If You want we can host them in Your repo, If You don't I will make 'git pull' from time to time :)

Thanks for Your work!

Hey @andris9.

set() method could really use TTL support.

I've seen there is a separate setTTL() method but that's just cumbersome to set a TTL after you've set the value. Many server side caching solutions allow you to define ttl when setting a value. It'd be great to add this.

Here is a proposed syntax:

$.jStorage.set(key, value, TTL);
// @var TTL - (optional) expiration time in seconds, after which the value 
//            is discarded. If omitted or set to 0 the value will never expire.

Hi,

First let me thank you for the wonderful little tool you've created, very useful!

I have run into a minor issue, which can be easily fixed (I'm no git user, so I didn't create a pull req for this).

in set:

    set: function(key, value){
        _checkKey(key);
        if(_XMLService.isXML(value)){
            value = {_is_xml:true,xml:_XMLService.encode(value)};
        }
        _storage[key] = value;
        _save();
        return value;
    }

If the value is an object, and it is later changed, the _storage cache will reflect those changes (while the localStorage won't). This is an issue, since the stored object may (and in my application does) acquire non-json-serializable properties (such as jQuery wrapped elements). when ever some other key is changed and the entire _storage object needs to be serialized, things break.

my fix was quite simple, perhaps there's a better way to go about this-

set: function(key, value){
        var clone;
        _checkKey(key);
        if(typeof(value) === 'object'){
            clone = $.extend(true, {}, value);
        }else{
            clone=value;
        }
        if(_XMLService.isXML(clone)){
            clone= {_is_xml:true,xml:_XMLService.encode(clone)};
        }
        _storage[key] = clone;
        _save();
        return clone;
}

Please note that this uses jQuery's extend function explicitly, I have no idea whether this could roll with other libraries. This works fine for me, so I hope it might help others as well.

Thanks again, o

Internal variable _storage is used for caching purposes. Suppose this example:

var b = {
    abc: function(abc) {
        return abc + '!';
    }
};

$.jStorage.set('b', b);

var b2 = $.jStorage.get('b');
log(b2.abc(123)); //shows '123! ', but should be 'error - no method abc'.

It gives the user the false understanding that he can store functions in storage. But we understand, that in browser storage his object has no 'abc' method.

I think it is better to make cache real cache :)

I just started using this for a site and it works great in IE8 and all other browsers I've tested on on all platforms. However, when I try it on my Windows XP boxes, one with IE6 and one with 7, I get the "null or not an object" error on page load. You can view my simplified test here: http://www.richpaul.com/dev/jstorage_test/

Any insight would be appreciated.

What is the max allowed size of JStorgae in different browsers(including mobile browsers)? I want to know this since I want to calculate the minimum of all the max allowed sizes and based on that I want to store some number of objects so that it can be used without any issue across most(almost any) browser(be it desktop or laptop or tablet or mobile based).

Regards, Sandip

Not even sure if the problem is in race conditions, since it involves just a single tab open. I used following code:

<!doctype html>
<script src="//cdnjs.cloudflare.com/ajax/libs/json2/20110223/json2.js"></script>
<script src="jstorage.js"></script>
<script>
  var me = "Client"+Math.random();
  var received = 0;
  $.jStorage.subscribe("aaa",function(chanel,data){
    received++;
    console.log("received message no. " + data[0] + " from client " + data[1] + " so far received "+ received + " messages");
  });
  for(var i=0;i<1000;++i){
    $.jStorage.publish("aaa",[i,me]);
  }
</script>

I'd expect it to report 1000 received messages, but actually only around ~800 pass trough in "Mozilla/5.0 (Windows NT 6.1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/41.0.2272.101 Safari/537.36" It looks like the missing messages are the initial ones.

Hey Everyone,

Everything was working great till recently and suddenly now I get the following errors throughout my website where jstorage is used Uncaught TypeError : Cannot read property 'get' of undefined

This is my code for one part of my website

if($.jStorage.get('my_guest')) {
      location = href;
    } else {
      LoginBox.checkActivate({img: img, title: title, cbRegisterURL: href,
                          onGuestClick: function(){location = href}});
    }
  } else {
    location = href;
  }

Any help will be greatly appreciated. Thanks for your time

I was wondering why XML Support is provided in the library...

You are not providing JSON.stringify, JSON.parse etc inside the library, right? You expect users to make sure that JSON support is present.

Not sure what percentage of users require XML support. I would dare to assume that the percentage is on the low side. Do you have any data on this?

IMHO, XML support should be removed from the library. Users who want to work with XML should be expected to provide endode/decode methods. (Also, if somebody is already working with XML, it can be safely assumed that they already have some methods defined to help with this.)

Even if the bytes saved are not huge, there is no need to support this out-of-the box IMHO. Should be at least moved out to a plugin.

Thoughts?

Hi! Npm contains 0.4.8 version with ambiguous UNLICENSED license in package.json. Do you have an ability to upload latest version there? I am happy to asssit if you can share the ownership in npm.

Forgive me if I have overlooked this somewhere, but what is the scope of the stored data? Can I set data in one page (example.com/foo) and retrieve it in another page on the same domain (example.com/bar)?

To be clear, I definitely don't want this - I need to know if this is something I deliberately need to avoid by namespacing my keys in some way. I want to store separate values for the key baz on the two pages example.com/foo and example.com/bar.

https://github.com/andris9/jStorage/blob/5f4075a7e1736d5c7d65375288154a73ac895818/jstorage.js#L459-L470

Looking at this code, if JStorage runs out of room, it will assume IE7 and kill the exception. This should be changed to at least display a warning in the console that JStorage was unable to save because local storage is full. This would save a lot of time.