As anybody else experience buggy behavior around using the Get Data from URL parameters expression and getting either blank values or incorrectly decoded values?
I first experienced it about a month ago with one specific instance of using a special character + which resulted in the Get Data from URL returning empty values, since the encoded values sent into the URL from the Go to Page action were improperly encoded. Mostly what I saw was that in front of the encoded value for a space, there was the numbers 25. When I removed the special character and replaced it with the word andthe issue went away, and I thought I could just chop it up to the special character.
Now for the past 2 - 3 weeks there has been an increasingly consistent bug causing the URL parameter values to be encoded improperly via the Go to Page action, which therefore has the knock on effect of having the Get Data from URL values returning as empty (when setting the type to something other than text).
This has meant a live application in production with real users and real business operating on it, getting bookings data submitted with no dates, or other important details like allergies omitted.
So far Bubble has stated that it was reported by other users and they felt it might have been a âblipâ and that other users reported it seemed to have gone away. Apparently they are saying it likely has something to do with the library they use for encoding values.
Has anybody experienced this and have a consistent result of the issue with a simple test page?
We were eventually able to track down exactly what it was doing by using WireMock and repointing our API settings to that. In our case, instead of converting âcompany incâ to âcompany%20incâ, it was converting it to âcompany%2520â inc. Almost as if it is encoding it twice. Started morning of August 26th, then ended at 4:30pm ET the following day. Everything worked fine again until Wednesday of this week when the behavior came back, and it hasnât been fixed yet. All support will tell us is they are working on it.
Hope it gets fixed soon, because our app is dead in the water due to our heavy reliance on APIâs out to third-party providers. To make it worse, it happened literally two weeks after we launched with much fanfare our new app to replace our old-outdated one on a different platform. Our users are quite angry at this point.
IMHO, if I was advising investors in a future Bubble IPO, I would say Bubble needs to start thinking about scalability itself and stop making so many minor code updates and go to a more traditional release schedule with better QA. You knowâŚsomething like âwe push updates on Tuesday mornings.â That way, if something that was working on Monday stops working on Tuesday, at least your clients have a good idea it was Bubble and not something else, and can let Bubble know right away. If it doesnât do that, the increased diversity of its applications will eventually mean more and more stuff breaks, and what were once edge-cases are now a real problem in terms of support/fix costs.
This is fixed. Bubble reply with details for what caused the issue to begin with.
This was a regression from a recent code change that inadvertently affected how we encode and decode different pieces in our language. In the re-writing of some of our code, an extra decoding was lost, leading to the resulting problems. This is why very niche manual workarounds may have been able to resolve the behavior, but weâve fixed it on our end entirely so this shouldnât occur again going forward.
