Settings and activity
6 results found
-
7 votes
Ron Wang
supported this idea
·
An error occurred while saving the comment -
21 votesRon Wang
supported this idea
·
An error occurred while saving the comment Ron Wang
commented
This can be a bit mitigated by using ask-cli. https://www.npmjs.com/package/ask-cli.
If you use ask-cli, like the "ask dialog" command to simulate with you developed skill, and if this message happens, you can run "ask lambda log" to immediately retrieve the recent cloudwatch logs, and you can very likely see the reason why this request failed. Hope this will help!
-
27 votesRon Wang
supported this idea
·
-
30 votesRon Wang
supported this idea
·
-
4 votes
An error occurred while saving the comment Ron Wang
commented
Hey Smeech,
Currently, it's true there is no way to local test your skill. We will definitely absorb this feature request!
In the meantime, you can actually achieve the goal of deploying to separate stack by using profiles. Basically you need to do the following steps to keep developing skills but not influence your live one: (Supposing you always use your "default" profile and use at least CLI v1.5)
1. Run "ask init --profile test". I'd suggest you use separate account/email for the [test] profile, so that you can have all stacks with the same name but different ID. (If you want to use same AWS account this can also be doable by using the "merge" field to overwrite the uri for the [test] profile)
2. Deploy using "-p test" when you want to keep developing, when your default profile is blocked by review, or the Lambda is used by live.We will also have this documented in our official docs in few months.
Thanks for your suggestion,
Ron -
4 votesRon Wang
supported this idea
·
An error occurred while saving the comment Ron Wang
commented
Hey all,
Thanks for the great suggestions and we have already had this feature implemented. The release of ask-cli v1.4 (June 21, 2018) starts to support skill project with eTag.
A new command "ask diff" is introduced to explicitly check the difference between local and remote resources.
For "ask deploy", once eTag mismatch happens, CLI will also prevent users from uploading.
For now you can try to use the "hook" feature that CLI newly brings in ask-cli@v1.5 to realize your goal. If you look at the python pre-deploy script example https://developer.amazon.com/docs/smapi/ask-cli-intro.html#hooks, you will find that we modify the sourceDir to a lambda_upload folder for python skills. And you can totally do this using some task manager tools like gulp (if you are nodejs developer) to decide what you want to deploy. This pre-deploy hook will be triggered before each deploy.