[Laszlo-dev] For Review: Change 20070803-hqm-0 Summary: preliminary implementation of dataprovider API

[P T Withington]

On 2007-08-05, at 14:02 EDT, Henry Minsky wrote:

> On 8/5/07, P T Withington <ptw at openlaszlo.org> wrote:
>> Questions:
>>
>> 1. Should there be an abstract `dataprovider` class?
>>
>
> There was the doRequest method on the DataProvider interface, so I  
> didn't think
> it was worth making a base class for it.

But it would allow you to document dataprovider, test instanceof (and  
eventually type declare things to be datproviders).

>> 2. If we had a dataprovider class, we could avoid global pollution by
>> having defaultdataprovider be a property of that.
>
> I'll try some things with Pablo, and see what seems to be the most  
> expedient
> and easy to deal with when he's implmenting the webtop dataprovider
> class, and then
> we'll make a proposal for what we find works out most smoothly.

Ok.

>>
>> 3. I don't recall `abortLoadForRequest` being part of the
>> dataprovider API, did I miss that?
>
> We didn't put that in the  API, but LzDataset supports an abort()
> method, which LzHTTPDataSource supported, so I need to support that
> somehow. The implementation in LzHTTPDataProvider is not working yet,
> but I guess we should see if we want to add it to the abstract
> DataProvider API.

Yeah, maybe generify it to simply abortRequest?