I've been testing it for similar cloud automation work over the last week. Your YAML concern is valid, it frequently disrupts formatting in Docker Compose and Ansible files, requiring a manual IDE reformat. For version control, it creates commits automatically but the messages lack context, like "Update deployment.yml". You'll need to edit them for any meaningful history.
Where it does help is the initial boilerplate for Terraform modules or generating a Python script from a prompt. However, the integration feels shallow. It doesn't seem to leverage the IDE's own formatting APIs or understand project-specific commit templates, which leads to those extra cleanup steps. For rapid prototyping it's okay, but for iterative edits on existing files, the friction adds up quickly.
Oh, that's interesting about the Kubernetes ConfigMap. I haven't tried it on anything like that yet, just some basic Python. So it's actually putting tabs in when YAML needs spaces? That's a pretty basic mistake, yikes. 😬
It makes me wonder if the plugin has different modes for different file types, or if it's just treating everything the same. If it can't get the formatting right for config files, maybe I should hold off on trying it with my Ansible stuff.
You said it works for Terraform HCL though. Is that because it's less picky about whitespace, or does it just handle it better for some reason?
Yeah, I've been trying it out for a similar workflow. Your YAML concerns are spot-on, I've had to hit the reformat shortcut after every edit for CloudFormation files, which gets old fast.
It does automatically create commits, but the generic messages are a real pain. If you're making a lot of small changes, you'll spend more time editing those messages than you saved. For scaffolding new Terraform modules, it's a decent time-saver. For tweaking existing Python scripts, the extra cleanup steps make it feel slower than just doing it myself right now.
data over opinions