Don't update inProgress (and the PR) if the codeflow had no updates#6122
Don't update inProgress (and the PR) if the codeflow had no updates#6122dkurepa wants to merge 1 commit intodotnet:mainfrom
Conversation
|
|
||
| result = result with | ||
| { | ||
| // For backflow, we want to always push the changes (even if only the <Source> tag changed) as it contains important information about where the flow came from |
There was a problem hiding this comment.
Somehow I don't understand how this explanation does not apply for forward flow too. Maybe it can be reworded to explain better why FF don't have it?
There was a problem hiding this comment.
I'm not sure too honestly. In FF, we'd consider this a meaningless change and not flow it. I'm not sure why backflows would want to treat this differently. I'd vote for removing this, pushing empty changes essentially.
@mmitche any thoughts on this? Is there any reason we'd want to update a backflow PR with a newer VMR build that doesn't actually bring anything else in (code or dependencies)
There was a problem hiding this comment.
If nothing is brought in, no dependencies and no code, then all we're doing is updating the latest backflow and that is just for tracking purposes at this point. That in itself may be pretty valuable.. I think this is pretty rare though. Maybe we see this with some tooling repos?
There was a problem hiding this comment.
yes, I think it's mostly repos like arcade and diagnostics. In that case it's probably fine to keep this. I was worried that if this was happening more often it might lead to us retriggering pipelines for no reason and just spending cores, but it sounds like it's just a rare thing.
I think the reason we don't do it for forward flows is because it'd eat up the pool fast, that's why we have the meaningless flows
There was a problem hiding this comment.
My comment was rather supposed to read like:
The way it's written right now in the comment, it does not have a lot of value because it does not explain why. Forward flows also carry important information about where the flow came from.
So if there's no updates to the PR, maybe we shouldn't update it in neither direction?
There was a problem hiding this comment.
I was gonna say that a new backflow update always brings in a new commit sha, even if there are no code contents (because other repos can have them), and that a FF would have the same sha if there are no content updates, but that's not really true is it, cloaked file changes would change the sha and there'd be no code update
#6063