@ecruzado The action is published to Alpha and is available on any of my devices using a "Talk To" invocation. I'm also seeing that there seems to be some issue maintaining context between commands - all stuff that was working before. Trying to see if I can glean more info.
hi @whetstonesanj_twitter , in general, to be able to test a google action with our virtual devices you need to be part of the team of the Google Action project, your google action has to be on Test mode by enabling "Testing on Device" and you have to make sure that your virtual device has been created with the same google account that has access to your google action
if it still doesn't work, maybe we can have a small call to troubleshoot the issue, just pick a time slot here: https://calendly.com/bespoken-diego/30min and Edgar and I will be glad to help
@dmarvp as regards general followup, because of the configuration I need to use with our customer, the only way I can test on a device is to publish to Alpha/Beta - that is working fine on physical devices the test account is associated with - since we just became a paying customer, I got an email asking to have a discussion next week, so I went ahead and requested times for a call that way where I can better describe the whys of our customer configuration. As regards the context issues - turns out that looks like an issue on our end that bst helped uncover - so thanks for that :-).
I've been using the command line
ask dialog --replay replayscript.json to run some of my testing, but wanted to move over use bespoken cli. Are the replay scripts from the ask cli mapped better to the end-to-end tests or unit tests? Also is there a conversion script to let me just run those simple json files as bst test files?
hi @TomV, as the ask command interacts with a deployed skill, it would map better to our e2e tests. We don't have a conversion tool for it, but by looking at an example file it should be pretty easy to convert them. You can read more about our e2e tests here: https://read.bespoken.io/end-to-end/getting-started/
Thanks @dmarvp ! When I tried running a simple test, I'm getting an odd response, and I can't seem to find anything in the documentation to explain the output. Here's what I see: \n
143 files checked.
testMatch: - 0 matches
testPathIgnorePatterns: /coverage/, /locales/, /lambda/, /test_output/ - 41 matches
testRegex: lambda/test/unit/index.test.yml - 1 match
Pattern: - 0 matches
(I can't seem to enter CR so that the markdown properly renders..)
Where are the
testPathIgnorePatterns getting set, and do I need to move my test OUTSIDE of the lambda folder to get it to run?
I find no results when searching the Bespoken docs for "testPathIgnorePatterns".. seems like it's not documented. :-( Trying
bst test -h does not yield any insight either, except the expected value after the test param is a regex, not just a path to the test(s). Escaping regex chars did not make a difference (e.g., the "/" and "." characters.)
Summary - I'm having a surprising amount of trouble just "getting out of the gate" test the bespoke cli tool - the proxy stuff works smooth as butter though! (seems preferable to ngrok.)
Hi @TomV sorry to hear about your issues, our tool use Jest in order to run the tests, that's why the testPathIgnorePatterns is not in our documentation, we do setup "/coverage/, /locales/, /lambda/, /test_output/" folders as ignored and it seems you have included your tests inside the lambda folder (usually the lambda zone includes only what you want to deploy to lambda and our tests are not part of that), could you try again setting the test folder outside the lambda folder one?
Now I have an e2e test that appears to run all the way to the end without errors, but I get a bad gateway errror (as an HTML page) response in the error. It's pretty consistent. How can I trouble shoot that error?
<meta http-equiv="content-type" content="text/html;charset=utf-8">
<title>502 Server Error</title>
</head>
<body text=#000000 bgcolor=#ffffff>
<h1>Error: Server Error</h1>
<h2>The server encountered a temporary error and could not complete your request.<p>Please try again in 30 seconds.</h2>
<h2></h2>
</body></html>
Timestamp:
2019-12-30T16:38:12.879
(not from my code, but from a failure reaching the lambda function.. which I guess would also be from the Alexa service..)
Hi @TomV , we usually recommend running them with a CI solution like Circle Ci or Travis, to run when you commit something instead of when the file changes.
In order to run jest on file changes usually, you use the --watch option, but we haven't implemented directly that option on bst test. We will continue looking into it to see if it's something you could add from the configuration or if we are able to add that as a feature in the near future.
In order to run jest on file changes usually, you use the --watch option, but we haven't implemented directly that option on bst test. We will continue looking into it to see if it's something you could add from the configuration or if we are able to add that as a feature in the near future.
Any ideas about the html response with a
502 error? Is that something other people have reported? I assume when I start on a new tool, I'm making the mistakes everyone else makes, so usually it's been documented somewhere.. :-)
One quick follow up regarding tests not running if inside of
/lambda folder (which does make sense..) The video from Sept 2019 has Emerson Sklar demoing command line but tests are in /lambda/custom, and they run fine. Seems like that would no longer work. ( https://youtu.be/YR7y2PA0ekI )
We do recommend keep them outside of the lambda folder, typically as a top-level directory called tests
With regard to the 502 error, that is not something that has been reported before
If you share more about the specific tests we might have some ideas, but just from what you have shared, nothing comes to mind
Is there any info on how and where jest is configured? (And Happy New Year to the Bespoken team 🎉🥳🎉 )
But you can override any of them by setting a "jest" property in your testing.json - then set the configuration to whatever you like
For example:
const request = alexa.request()
.intent("MyIntentName")
.slot("SlotName", "SlotValue")
.set("forceError", true) // Setting additional property
And then my code that reads it and throws an error:
https://github.com/jkelvie/ClassicMovieMusts/blob/master/lib/index.js#L16
https://github.com/jkelvie/ClassicMovieMusts/blob/master/lib/index.js#L16
how do i revoke virtual device access to my alexa developer account?
With regard to removing access, we are listed as a device like any other and can be removed in the same way
Such as explained here: https://www.cnet.com/how-to/how-to-remove-alexa-devices-from-your-account/
How can I assign a4b room and profile with the virtual device?
Are there other ways to associate a4b information during integration testing?
1) Deploy the skill as a normal skill in development mode and test that way