Content not refreshing in Google assistant, but works fine in Dialog Flow - actions-on-google

I have been searching so long for this, and couldn't find a solution.
So, I have developed a pretty basic dialog flow project and deployed it onto firebase using inline editor.
Now the 1st deployment works perfectly fine with the simulator.
However, I changed some content to include suggestion chips and modify the responses as well; this gets reflected fine in dialog flow. And I can see the changes I have made.
But when I update the Draft and try to test it again in simulator/google assistant device, it doesn't load the newer changes. I am not sure if what I am doing is correct or not.
Please help :(
Just now found this - https://discuss.api.ai/t/actions-on-google-integration-agent-updates-do-not-always-publish-correctly/7483
It seems many users are facing this issue.. :/

Dialogflow had a bug causing this issue which should now be fixed.

Related

Google Assistant Hello World Draft Project not updating

Very new to Google Actions. Testing out tutorial stuff.
I have tried this in a couple of test projects just to double check. After initial run of any project I do not get any updates on draft projects. No changes show up in draft projects for me for both simulator and real device.
Started new project
Even blank project has basic conversation telling you where to add things next.
Change text.
Notice prompt does not change in testing environment.
In below pictures I have changed the words "hello world" with "Hey Dude" for both fulfillment and console output. I would expect Testing Prompt to respond with "Hey Dude from fulfillment and Hey Dude from the console!" But it does not. Instead it does not reflect any recent changes.
I think there may be two slightly different (but sometimes related) issues going on here.
The first is that there are known problems with the simulator being slow to pick up on updates, or them not seeming to show up. The second has to do with making sure you're deploying changes from the build-in code editor.
I don't have a clear answer to the first problem, although I know they're looking into it. I find that I can make some changes and they may not be noticed, but I know they have been picked up if I see the "Your preview is being updated..." spinner appear. There are other spinners that sometimes appear, but unless it explicitly says that it is being updated - the updates aren't always picked up. (Sometimes they are, however.)
Usually, if I don't see this, I'll go back and force an apparent change (delete a character from a webhook handler name, then add it back) and go back to the simulator. In general, this time it will say it is updating.
If you're using the Cloud Functions editor, you need to do three things:
Save the changes. You'll do this by clicking the "Save Fulfillment" button, but this only saves it so you can leave the editor. It doesn't mean that the simulator has access to it yet.
Deploy the changes. This deploys your code to Cloud Functions so they can be run. Note in the illustration that it says the code is saved, but not yet deployed.
Wait till the changes are fully deployed. Deploying takes time, and until it is completed, it won't be available in the simulator. While deploying, it lets you know.
Once it has deployed, however, the message changes, and you the impacts should be available through the simulator (although you may still need to see the "being updated" message to be sure).
Remember, however, that you don't need to use the "Cloud Functions editor" in order to deploy a webhook. You can deploy a webhook on any web server where
The host is public (so has a public IP address that Google can reach)
It can handle HTTPS with a non-self-signed certificate
You can even deploy yourself to Cloud Functions for Firebase, which is the same service that the Actions Builder uses. This way you have the URL set once in the Actions Builder and, once it is set, you won't need to change it.
But you'll still be able to change your code by managing your own deployment separate from Actions Builder.

Branch link opens Ionic 3 app, which immediately closes before loading

I'm configuring Branch and seeing what I think is unexpected behavior upon trying to test my setup. I have an Ionic 3.7 app and am new to Ionic/iOS/Android development- so I've been following the Branch docs to try and get this configured (just for iOS right now).
What I've done so far:
I started with this guide: https://dev.branch.io/getting-started/sdk-integration-guide/guide/cordova/
The pre-requisite is that Universal Links are setup, which makes sense. In order to get that working, I clicking the link in the guide above, which took me to this guide: https://dev.branch.io/getting-started/universal-app-links/guide/cordova/
The Universal Link steps have gone fine until #3, where I'm trying to verify that my setup is correct. I created a Quick Link and texted it to my phone (which had a build of my app that I ran after adding the config in step 2).
Clicking the link starts to open the app (I see the splash screen), but it closes before loading is complete (I see a black screen momentarily). When I open this build of the app normally, it loads and runs normally
As I mentioned, I'm new to this- do I need to follow the Apple Universal Link setup (from here) with the apple-app-site-association file before starting the Branch config? I didn't see any mention of that in the guides I followed- however, it may be something so basic that it's assumed the user has done it already.
Regardless, I did configure that file and make it available as the Apple guide says. I'm still seeing the same behavior when I click my Branch link.
The 2nd Branch guide I posted only mentions that the app opens immediately without routing through Safari as a result of the test- is what I'm seeing actually correct? I don't have the SDK installed to handle the links, so I could see that being the case too.
UPDATE
I went ahead and installed the SDK, following the instructions here. It was somewhat confusing, as I wasn't sure if I should be supplementing the additions to config.xml from my step 1) above, or if my branch-config section should only have the keys from the SDK installation example.
I ended up commenting out the changes to config.xml from step 1) and it looks like it worked, as I now see the deeplink data when opening the app via the Quick Link I created before. I had previously added the domains to my entitlements file directly in Xcode before, though, so I'm now wondering if I need those extra config keys in order to persist the entitlements changes should I need to remove/readd the iOS platform.
Amruta from Branch.io here:
You can follow the iconic integration guide available on Github here.You need to update your config.xml with the Branch key assigned to your App, which can be found on the Account Settings page of the Branch dashboard. The URI scheme is the one you have defined in the Link Settings page of the Branch dashboard and the link domain for your Branch links is also defined on the Link Settings Page.
If you still have issues, I would recommend writing to integrations#branch.io.

WF 4.5 Breakpoints and Visual Tracking issue

I am using the code from Kushal's blog: http://blogs.msdn.com/b/kushals/archive/2009/12/29/visualworkflowtrackingwithstepservice.aspx
It works as expected when i load a workflow, set some breakpoints and hit run.
Visual feedback works fine too.
The problems show up when i add a new activity to the designer. For example, i've added a new WriteLine. Plain and simple.
I am not sure why this happens, why do i get those inconsistent data? My breakpoint is drawn, but never hit. Is there a way to refresh somehow the designer without reloading it?
SourceLocationProvider.CollectMapping(documentRootElement, documentRootElement, designerSourceLocationMapping,
this.wd.Context.Items.GetValue<WorkflowFileItem>().LoadedFile);
The designerSourceLocationMapping is populated as above.
I am not sure what details should i mention, the full code example can be found at the link provided - on Kushal's blog. Has anyone encountered this scenario until now and found a solution?

long script error internet explorer

Please can someone help?
I've had facebook integration scripts running on my website for many years already with no problem. However, I am now being flooded with complaints from customers who use INTERNET EXPLORER saying that since an update to IE a week or so ago they are constantly getting the long-running script errors...
"Stop Running This Script.
A script on this page is causing Internet Explorer to run slowly.
If it continues to run, your computer might become unresponsive."
I've worked it out (by disabling scripts) that the problem is with the Facebook scripts.
Nothing has been changed in months on the website so it's not something I can figure out how to fix.
Scripts: Invite your friends, Like, Share page, etc
Please, please help
Thanks
Matt
I've ran into the same issue. I can reproduce the issue not just on my site, but on others as well when using IE 9 or IE 10 (have not tested with older IE versions). The issue seems to be a result of IE's compatibility view feature. If that is enabled for a website, which also has the facebook like button on the page of the site you're trying to load, a long running script error is produced (error from a function called _t). When I disable compatibility view for that website, the error goes away. I'm not sure if the issue is with the IE feature, or with facebook's button...I'm not getting same error from other scripts, nor other social buttons like google+, which makes me think the issue is on Facebook's end. I've been trying different IE setting configs for days now...so far, the only fix I can come up with is turn compatibility view off. And of course resolve issues so that compatibility view are not needed. Though telling your users they have to turn off compatibility view doesn't seem like an appropriate response...so I hope a better fix is found.

Umbraco - Unable to navigate to sections other than Content after upgrade to 4.5.2

I've tried posting this on the Umbraco forums to no avail. Hoping to find some help here.
I thought I upgraded successfully from 4.0.4.2 to 4.5.2 (on my way to 4.8...) as I received no errors and everything seemed to go smoothly. However, in the backoffice, when I click on any section other than Content, I get redirected back to Content. I can manually reach each section by typing in the name of it after the # sign like: #media, but as soon as I click on a tree item, it redirects me back to Content. Any ideas would be appreciated.
2 thoughts:
You're cacheing some of the old javascript in your browser. Try clearing your cache. IE can be stubborn in this regard; preferably try in Chrome.
Personally, if your site is working under 4.5.2, I wouldn't sweat the back office issue and just continue with the upgrade to 4.8.1. There will likely be new problems to solve there anyway :) and the backoffice issue may go away.
Updating the icons via the following script fixed this problem. Links must run off the appIcon field. Not sure why this is not executed when the database is updated via the 4.5.2 upgrade.
update umbracoApp
set appIcon = '.tray' + appAlias
where appAlias IN ('content','media','users','settings','developer','member')