Specify that runtime attributes can be declared as inputs #313
Conversation
|
+1 for this PR :) |
| @@ -2647,6 +2648,23 @@ In JSON, the inputs to the workflow in the previous section might be: | |||
|
|
|||
| It's important to note that the type in JSON must be coercible to the WDL type. For example `wf.int_val` expects an integer, but if we specified it in JSON as `"wf.int_val": "three"`, this coercion from string to integer is not valid and would result in a coercion error. See the section on [Type Coercion](#type-coercion) for more details. | |||
|
|
|||
| ## Specifying / Overriding Runtime Attributes in JSON | |||
|
One thing worth making explicit one way or the other, do we want to consider these to be providable as call inputs too? ie something like... call foo { input:
runtime.docker: "ubuntu: latest"
}My gut feeling is yes but I could see an argument for no |
I think it should be supported, in the interest of keeping the WDL spec consistent. |
|
In light of #315, I wonder if we should revist this? |
Unless I'm misunderstanding your comment, I don't think it's required. There will be cases when PR overlap, but I don't think it's necessary to block or update one based on the other, since neither are yet part of the spec. In this case I think the PRs are orthogonal anyway; e.g. regardless of what changes are made to the contents of the |
|
Sorry this is what being out of town gets you. On the flip side, you should see the view I have while typing this :) @patmagee I was going to suggest folding in the spirit of my PR into this, if that's not already done. |
|
Closing this as per @geoffjentry suggestion. Changes have been folded into the #315 |
This PR replaces #301. The primary difference is that this PR only refers to things which are already part of the spec, i.e. references to input json format remain while references to input json files have been elided