[mistral] Publish field in workflow tasks
Hello,
Currently, there are two "publish" fields, one in the task(regular "publish")-the scope is branch and not global, and another under "on-success", “on-error” or “on-complete”.
In the current behavior, regular "publish" is ignored if there is "publish" under "on-success", “on-error” or “on-complete” [1].
For example:- (a) version: '2.0' wf1: tasks: t1: publish: res_x1: 1 on-success: publish: branch: res_x2: 2
(b) version: '2.0' wf2: tasks: t1: publish: res_x1: 1
"res_x1" won't be published in (a), but it will in (b).
We can either:-
1) Invalidate such syntax. 2) Merge the two publishes together and if there are duplicate keys, there are two options:- a) What takes priority is what's in publish under "on-success" or “on-error” or “on-complete. b) Not allow having a duplicate.
What is your opinion? And please tell us if you have other suggestions.
Ali,
I’m for the option 2.a because it’s not so difficult to implement but it’ll be the best effort to handle a situation more gracefully if someone puts “publish” in both places (old syntax and advanced syntax). Over time we’ll deprecate the old “publish” completely though.
Thanks
Renat Akhmerov @Nokia On 28 Aug 2019, 15:37 +0700, Ali Abdelal maxget7@gmail.com, wrote:
Hello,
Currently, there are two "publish" fields, one in the task(regular "publish")-the scope is branch and not global, and another under "on-success", “on-error” or “on-complete”.
In the current behavior, regular "publish" is ignored if there is "publish" under "on-success", “on-error” or “on-complete” [1].
For example:- (a) version: '2.0' wf1: tasks: t1: publish: res_x1: 1 on-success: publish: branch: res_x2: 2
(b) version: '2.0' wf2: tasks: t1: publish: res_x1: 1
"res_x1" won't be published in (a), but it will in (b).
We can either:-
- Invalidate such syntax.
- Merge the two publishes together and if there are duplicate keys, there are two options:-
a) What takes priority is what's in publish under "on-success" or “on-error” or “on-complete. b) Not allow having a duplicate.
What is your opinion? And please tell us if you have other suggestions.
participants (2)
-
Ali Abdelal
-
Renat Akhmerov