Here is how I used to use the querystring attribute. I have a file where the datasets for the whole application are declared, each with the querystring attribute holding the method that the dataset is calling. >From there the application would fill in additional query parameters from the user interaction with the setQueryParams call. We then call a method in the application that adds the user authentication with setQueryParams right before we doRequest. From what has been described below it sounds like even if the querystring bug is fixed this feature of using querystring and then setQueryParams will no longer be supported. This workflow theory is the foundation for every database call in our application, how would you suggest fitting this theory into with the new way that the dataset class is being layed out. For right now I am thinking of a duct tape solution where oninit of the application I setQueryParams with the method for each dataset I am using.
David -----Original Message----- From: Henry Minsky [mailto:[EMAIL PROTECTED] Sent: Friday, September 07, 2007 2:03 PM To: P T Withington Cc: David Buckler; Laszlo Subject: Re: [Laszlo-user] Querystring Dataset Attribute Yeah, that's it. The issue was whether the API should attempt to merge all the query info together: from the original URL, from the query params table, and from the query string, and that just seemed even more confusing if you considered all the possible permutations On 9/7/07, P T Withington <[EMAIL PROTECTED]> wrote: > I think the intent was that set of src, query string or query params > all boil down to setting the query on the src url. The string and > params interfaces are convenience methods, but they should all be > affecting the same underlying url. > > On 2007-09-07, at 16:41 EDT, Henry Minsky wrote: > > > Yes, that is actually by design, setqueryparams blanks out the > > querystring attribute. > > I forgot why we made it do that, it does seem confusing now. > > > > > > > > On 9/7/07, David Buckler <[EMAIL PROTECTED]> wrote: > >> I did reproduce this and it does work. However in my case I need to > >> call setQueryParam/setQueryParams and when you do that it blanks > >> out the > >> querystring attribute. Let me know if you are noticing the same > >> behavior. > >> > >> David > >> > >> -----Original Message----- > >> From: Henry Minsky [mailto:[EMAIL PROTECTED] > >> Sent: Thursday, September 06, 2007 3:41 PM > >> To: David Buckler > >> Cc: [email protected] > >> Subject: Re: [Laszlo-user] Querystring Dataset Attribute > >> > >> Are you sure about that? If I run this example > >> > >> <dataset name="ds_test" request="false" queuerequests="true" > >> type="http" > >> src="https://testing.laszlo.com/test.cfc?method=baz"; > >> secure="true" > >> /> > >> > >> And then ask for ds_test.querystring in the debugger, I get the > >> 'method=baz' string as the value > >> > >> lzx> ds_test.querystring > >> 'method=baz' > >> > >> > >> > >> On 9/6/07, David Buckler <[EMAIL PROTECTED]> wrote: > >>> In addition when I add "?method=foo" to the src url, where it should > >>> parse out and set the querystring, it doesn't behave correctly > >>> either. > >>> > >>> Its almost seems like the setSrc function is being called twice; the > >>> first time when it gets parsed correctly and the "?method=foo" gets > >>> stripped from the src attribute by: > >>> > >>> url.query = null; > >>> this.src = url.toString(); > >>> > >>> then it gets run again and the querystring gets blanked out, just > >>> like > >>> when you set the querystring attribute like I have. > >>> > >>> Just a thought. > >>> David > >>> > >>> -----Original Message----- > >>> From: Henry Minsky [mailto:[EMAIL PROTECTED] > >>> Sent: Thursday, September 06, 2007 3:02 PM > >>> To: David Buckler > >>> Cc: [email protected] > >>> Subject: Re: [Laszlo-user] Querystring Dataset Attribute > >>> > >>> Yes, that is a bug. I will file a bug report. > >>> > >>> The setter for the "src" attribute in LzDataset.lzs is nulling > >>> out the > >>> querystring arg, because it tries to parse it out of the src url. > >>> > >>> LzDataset.lzs: > >>> function setSrc( src ) { > >>> var url = new LzURL(src); > >>> this.querystring = url.query; > >>> url.query = null; > >>> this.src = url.toString(); > >>> if ( this.autorequest ){ > >>> this.doRequest() ; > >>> } > >>> } > >>> > >>> I guess it ought to check if the user supplied the argument during > >>> init. The question is what to do if there is a conflict between the > >>> src url and the querystring, like > >>> > >>> <dataset src="http:foo.php?method=foo" querystring="method=bar"/> > >>> > >>> I guess a runtime warning would be the right thing to do. > >>> > >>> The workaround for you would be to call setQueryString manually > >>> before > >>> making the request. > >>> > >>> > >>> > >>> > >>> > >>> On 9/6/07, David Buckler <[EMAIL PROTECTED]> wrote: > >>>> > >>>> Thanks for checking this out Henry, here is a test line, obviously > >> the > >>>> cfc that this points to isn't real. Also in our situation we add > >>> query > >>>> parameters with the setQueryParam method. When I watch the HTTP > >>> requests > >>>> I notice that the querystring parameter is never added but the > >>>> parameters that are with setQueryParam are. I also tested by > >>>> adding > >>> the > >>>> method=testme to the end of the src and that didn't work either. > >>>> However from what I understand that is just pulled off the end of > >> the > >>>> src and is stored in the querystring attribute. Note that other > >>>> datasets we are using where we set the method by means of the > >>>> setQueryParam work fine. > >>>> > >>>> David > >>>> > >>>> > >>>> <dataset name="ds_test" request="false" queuerequests="true" > >>> type="http" > >>>> src="https://testing.laszlo.com/test.cfc"; secure="true" > >>>> querystring="method=testme"/> > >>>> > >>>> -----Original Message----- > >>>> From: Henry Minsky [mailto:[EMAIL PROTECTED] > >>>> Sent: Thursday, September 06, 2007 1:42 PM > >>>> To: David Buckler > >>>> Cc: [email protected] > >>>> Subject: Re: [Laszlo-user] Querystring Dataset Attribute > >>>> > >>>> You might want to compare the behavior in one of the nightly builds > >> of > >>>> LPS 4, to see if there is an issue that may have been recently > >> fixed, > >>>> as well. > >>>> > >>>> > >>>> > >>>> On 9/6/07, Henry Minsky <[EMAIL PROTECTED]> wrote: > >>>>> Do you have a small test case you can send me? > >>>>> > >>>>> > >>>>> > >>>>> On 9/6/07, David Buckler <[EMAIL PROTECTED]> wrote: > >>>>>> > >>>>>> > >>>>>> > >>>>>> > >>>>>> I just installed 4.0.3 and have begun moving my code base from > >>> 3.3.3 > >>>> over to > >>>>>> the new version. I am having an issue with the querystring > >>>> attribute in the > >>>>>> dataset object. It appears that Laszlo isn't even paying > >>> attention > >>>> to it. > >>>>>> Has anyone had a similar problem or know anything about this? > >>>>> > >>>>> > >>>>> -- > >>>>> Henry Minsky > >>>>> Software Architect > >>>>> [EMAIL PROTECTED] > >>>>> > >>>> > >>>> > >>>> -- > >>>> Henry Minsky > >>>> Software Architect > >>>> [EMAIL PROTECTED] > >>>> > >>> > >>> > >>> -- > >>> Henry Minsky > >>> Software Architect > >>> [EMAIL PROTECTED] > >>> > >> > >> > >> -- > >> Henry Minsky > >> Software Architect > >> [EMAIL PROTECTED] > >> > > > > > > -- > > Henry Minsky > > Software Architect > > [EMAIL PROTECTED] > > -- Henry Minsky Software Architect [EMAIL PROTECTED]
