Is anyone else having basic usage problems with Fleet Preview?
I have experienced the following problems with Fleet:
Extraordinarily slow typing and response rates both for code editing and AI Assistant prompt editing and response generation. Seconds of delay per character at times, and this is apparently consistent across times of day etc.
Loss of code completion and elementary refactoring (like renaming a class) after a mandatory upgrade to a more recent release
Loss of all AI Assistant history after a mandatory upgrade to a more recent release
Documentation that is not up to date with the releases
A promise of free AI Assistant during the whole Preview period, followed by a demand to pay for a subscription to continue using it with the Preview of Fleet
Faiiure to recognize my paid-for licence for AI Assistant, so I had to start a new Trial … which blitzed my history, again.
I was doing a tiny project with a few dozen classes and a couple of modules in Python. I've abandoned this for PyCharm which is wonderful (but for how long, if this is typical of modern JetBrains products?)
I do not expect great things of a free Preview, but I do expect to get capabilities I have paid for. And I would have thought that the Preview would be minimally usable for the most basic tasks.
I have submitted tickets or feedback for these items: I see no sign of change with the latest release.
Please sign in to leave a comment.
Thank you for your valuable feedback! We truly appreciate it and use it to improve our product!
Could you please clarify which version of Fleet you're currently using? Additionally, could you let us know the type of project you're working on and the languages involved, except for Python?
Unfortunately, we are aware of an issue where AI history from pre-release versions may be incompatible with newer versions, potentially leading to history loss. However backward compatibility of AI history is expected for the versions after the official launch of Fleet.
Could you also let us know if there are any other topics, besides the AI assistant trial license, that you believe need to be covered in more detail in our documentation?
Thank you again for your input!
Yes, it's terribly slow, and the bugs are embarrassingly bad.
For example, in Swift, I do a lot of switch'ing on enums, which I always start with “add missing cases”. The thing is, it's *always* inserting the cases *after* the body of the switch, instead of inside of it… I just got used to undoing that and executing “add missing cases” a 2nd time, where it actually works.
A lot of completions are like this, out of place, sometimes in the middle of other code etc.
It's very much a broken product (but still better than Xcode in some aspects so…)
Hello,
Thank you for the feedback!
Please tell if the issue is reproduced for you with the code sample as following:
If no, please share the code sample to reproduce on our side.
Thank you