Announcement

**Isomorphic** · 22 Jan 2013, 14:06

Callbacks do not fire for download operations because there is no way for the browser to be notified of completion. You will need to rearrange your logic to avoid needing a callback.

**RickBollinger** · 23 Jan 2013, 09:45

Yipes! this aspect of downloadResult is even in the docs.

Rearranging my logic is easier said than done. But, I am trying. I have successfully made the callback fire. And, it works great. It opens a new window and populates its form fields from the response headers.

Here is the frustration: I can see the PDF file contents in the response httpResponseText and the intended file name in the Content-Disposition header. So, I have been looking how to implement my own hiddenFrame transport and, duh, save the contents in a new file with that name.

Seems the original intent of the world-wide web to download files is not natively implemented in browers in an independent way. HTML5 has a saveAs that might work if I could find a way to access the browser window as an object.

I imagine if this was easy, downloadResult would work with callback.

Any ideas?

Rick

**Isomorphic** · 23 Jan 2013, 10:05

It's very very hard, and generally speaking, cannot be achieved purely browser side, and must be implemented by having the server monitor the streaming of bytes to the client (tricky in itself) and polling the server to check for download completion while the download occurs.

In other words, much harder than rearranging your logic and/or screen flow to avoid relying on a callback.

**RickBollinger** · 24 Jan 2013, 14:11

Actually, I solved it mostly on the browser side. All I did on the server was take the data that I was passing in the response headers and insert it into the already open database in a separate table. I used a unique random key calculated in the browser code and passed in the request as a parameter.

Then, after the request returns, and waiting for the result to be downloaded, I recursively call setTimeout and keep checking to see if the database record had been inserted yet.

The attempt that first finds the record present then opens its data in a new window for the user.

Code:

	var form = MeetingForm;
	
	var oldPrintOptions = form.getValue("AgendaPrintOptions"); 		
	var vmOptions = meetingTypeVM.getValues().Agenda.Options; 
	var vmNewOptions = vmOptions.Agenda.Options;	
	var vmTopic = meetingTypeVM.getValues().Agenda.Topic;
	var vmNewTopic = vmTopic.Agenda.Topic;
	var vmIssue = meetingTypeVM.getValues().Agenda.Issue;
	var vmNewIssue = vmIssue.Agenda.Issue;	
	var vmTask = meetingTypeVM.getValues().Agenda.Task;
	var vmNewTask = vmTask.Agenda.Task;	
	var vmSpecial = meetingTypeVM.getValues().Agenda.Special;
	var vmNewSpecial = vmSpecial.Agenda.Special;	
	var vmNewPrintOptions = { Options: vmNewOptions, Topic: vmNewTopic, Issue: vmNewIssue, Task: vmNewTask, Special: vmNewSpecial  };
	var newPrintOptions = isc.JSON.encode(vmNewPrintOptions);		
	if (oldPrintOptions != newPrintOptions) { 
		form.changesPending = true;
		form.setValue("AgendaPrintOptions", newPrintOptions);
	}

	var meetingDate;
	meetingDate = String(form.getValue("MeetingDate"));
	if (meetingDate == "undefined") meetingDate = ""; 
	var meetingTime;
	meetingTime = String(form.getValue("MeetingTime"));
	if (meetingTime == "undefined") meetingTime = ""; 
	
	var d = new Date();
	var minOffset = d.getTimezoneOffset();
	var s;
	if (minOffset < 0) {
		s = "+"; minOffset = -1 * minOffset;
	} else {
		s = "-";
	}
	var hrs = 100 + (minOffset / 60);
	var mins = 100 + minOffset - ( (hrs - 100) * 60);
	var timeZoneOffset = s + hrs.toString().substring(1) + ":" + mins.toString().substring(1);
	
	var emailsKey = Math.floor((Math.random()*1000000000)+1);
	
	form.setValue("emailsKey", emailsKey);
	
	RPCManager.sendRequest({
		httpMethod: "POST",
		actionURL: "/servlet/MeetingAgendaServlet11",
		params: {	meetingDate:		meetingDate,
					meetingTime:		meetingTime,
					timeZoneOffset:		timeZoneOffset,
					emailsKey:			String(emailsKey),			
					meetingTypeID:		String(Application.currentMeetingTypeID),
					agendaPrintOptions: String(form.getValue("AgendaPrintOptions")),
			},
		downloadResult: true, 
		downloadToNewWindow:false,
		useSimpleHttp: true,			
	});

	var attempts = 0;  //will be used to max out attempts
	var timeout = 250; //milliseconds that will double with each attempt
	
	Log.logDebug("***EmailsExchange record fetch attempt " + attempts);
	
	function tryAgain() {
		attempts = attempts + 1;
		EmailsExchange.fetchRecord( emailsKey, 
			function (dsResponse, data, dsRequest) {
				if (dsResponse.status >= 0) {
					if (dsResponse.totalRows == 1) {
						var rec = {};
						rec = data[0]; // first and only record
						EmailForm.editRecord(rec);
						EmailWindow.show();
						form.getItem("GetEmailsButton").setDisabled(false);	
					} else { 
						if (attempts < 10) {
							timeout = timeout * 2;
							setTimeout(tryAgain, timeout);
						} else {
							isc.say("Maximum attempts reached for fetching emails.");
						}
					}		
				} else { 
					if (attempts < 10) {
						timeout = timeout * 2;
						setTimeout(tryAgain, timeout);
					} else {
						isc.say("Maximum attempts reached for fetching emails.");
					}
				}
			} 	// end of fetchRecord function
		); 	// end of fetchRecord call
	} // end of tryAgain
	
	tryAgain();

If you never are able to get downloadresult to work with callbacks, I will understand. Such a callback would have to wait a long time for big files to download.

Thanks for all the quick replies and helpful comments. This application is getting bigger and more complex. The sooner I can abandon the XML, the better. Still some technology to implement yet.

Rick

Announcement

RPCManager.sendRequest callback does not fire

Comment

Comment

Comment

Comment