Hi Forest guys.
I’m absolutely sure that today at 15.30 (CET) the following was true (since years):
in a smart action, if we have a text field which:
- is left empty
- is not “touched” by the user
when the smart action is submitted, the field value receieved in the backend was
As far as I can understand, I can see that at about 17.30 (CET), in the very same conditions, the returned value is not
undefined any more and
null is returned.
At the moment I’m dealing with a number of regressions in my smart actions. So now I’m fixing on my side, managing this new
null value too.
But in the mean time, may you please double check if something changed on your side in the Forest front end?
Yes good catch there is a tiny change.
null value are not converted to
undefined anymore. It makes more sense cause
undefined means the field does not exist while null means it exists but it’s empty.
I hope it’s not too much to fix for you. Let me know otherwise I’ll do a quick “fix”.
No problem, we already fixed. And I understand your logic: I think that the actual solution with
null instead of
undefined is better. So, no need for a revert on your side.
The fix on my side was just a few lines of code…
The real hassle was finding out the problem, because this situation was unexpected and a change in Forest UI was the very last option we considered. In the afternoon a number of smart actions on our side stoped working. So this slight change created a huge incident in our business.
As far as I know, there’s no public release notes for your releases in the forest UI. So It’s very difficult to me to match the problems we experience with the changes on your side. We all agree that you are expected not to introduce regressions or breaking changes, but I understand it’s impossible to consider all the possible scenarios, especially with slight edge changes like this one. Nevertheless, we need to prevent this kind of situations.
In my opinion, it would be nice to have some kind of release notes whenever you have a new release. This could enable us to contribute better, while spending less time in diagnosing edge situations like this one. Why not?
Thank you in advance,
Hi @Matteo Thank you for your report.
We are totally agree with you and your feedback brings us to think about this type of changes.
We apologize for the time you have lost to find out where the problem came from.
Thank you again, i’m gonna share your feedback to our team.
For your information, we decided to rollback this behaviour as it might impact a lot of people without any prior notice (as it happens to you).
We’ll probably make it live again in the future, but after a great communication about it.
Thanks a lot for your precious feedbacks
Ok @adriguy thank you for letting me know.
I confirm that with our latest fix we can now deal with both
null, so no problem on my side.