Is V2 Really a 'Next Generation' of V1?

This topic was automatically generated from Slack. You can find the original thread here.

I must say it - As much as I am trying to adopt the workflows V2, I keep going back to V1. I keep unused V1 workflows around and repurpose them, just so I dont have to use V2.
V2 is IMO way less user friendly than V1, for so many reasons. It is so frustrating to work with V2, really. I simply don’t understand how V2 is a next generation of V1…
Am I the only one feeling that way?

Hi Haim, appreciate the feedback. Are you able to articulate exactly what you like about v1 / where you think the v2 user experience is lacking?

Sure, here are some examples:
• Use case - Editing a workflow after investigating a specific event. With V1, it is straight forward. With V2 - you must enter edit mode, and you lose the context of the event. It also refresh the page, so you loose the point of code that you wanted to edit. This is really the biggest pain point I have with v2, and really dont understand how is it better than v1… V1 is so simple to use - and this is exactly why I use pipedream. V2 adds more friction (IMO) and I just dont see the benefit
• When ending a workflow in V1, we can write a custom text, and it will show on the left side pane. Super helpful. I didnt found a way to do that with V2
• I have cases where a v2 workflow has an error, but I simply can’t understand why the workflow fails. I see the error, see the event, really no explanation. This never happens with V1. I can provide an example from today
I could go on but these are from the top of my head. The first point (edit/view mode) is the biggest downside from my POV

That’s all great feedback. We consider these bugs, and I’m raising with the team again. We’re tracking #1 here and #2 here. Can you share what you’re seeing re: #3? The error / stack should be no different in v2, so it’s possible that’s also a bug we can address.

We primarily moved to the view / edit model to let you test your workflow before deploying it to production. In v1, there’s no mechanism to do that — you must deploy changes directly to prod, and can’t test individual steps. For real production workflows, this caused major issues for customers. That does naturally add some friction, so I sympathize with that. I think we can absolutely improve the UX to make it “feel” more like v1 in terms of speed of editing, while retaining the ability to test the workflow before deploying.

Thanks, glad to hear that some of these are work in progress!
I can think of other ways to support a dev/prod environment, dont see how the edit mode from V2 helps in that.
Sorry for the harsh feedback, I am kind of a Pipedream ambassador actually, promoting it everywhere haha, just want it to be the best…

yeah we love feedback! I agree that there are other ways we could approach the dev / prod environment. I also think we can make the v2 editor more like v1 without substantial changes to how we handle environments… so I think there are a lot of ways to approach the core issue. Will share all of this with the team!

Another downside to V2 - Why doesnt it tell me which line has an error? I need to do guessing work

do you know if we have a public issue tracking this?

Please keep raising these Haim — these are likely just bugs. There’s no product reason we hide the stack, it looks like a regression

I see. I think I just found another bug - flow.exit is not exiting the workflow, I couldnt understand why I had a failure. Is that expected?

it’s subtle, but you’ll actually need to return $.flow.exit() : Writing Node.js in Steps

> do you know if we have a public issue tracking this?
not that I’m aware of

tracking #3 here: [BUG] Errors in inspect mode don't show stack trace · Issue #4173 · PipedreamHQ/pipedream · GitHub

we addressed #1 and #3! Let me know if that works better for you