2000ms timeout insufficient for dependency installs

[pythagora-io]

2000ms timeout insufficient for dependency installs

[nalbion]

The timeout is suggested by the LLM. We could retry with an longer timeout, but ideally the user would be able to provide free-form input for the LLM such as

My computer is slow, it's going to take 10 minutes.

or

Don't use @okta/okta-react, use @me/okta-react

[HIMA-MB02]

Hey @nalbion
What if we asked the LLM to use yarn to manage dependencies?
Installation of parallel dependencies will definitely improve install times.
Additionally, can we ask the LLM to batch installations of dependencies?
We could define a threshold for max number of dependencies that can be installed in one command.

I am interested in picking this up, happy to discuss further

[nalbion]

@HIMA-MB02 is Yarn still considered to be superior to NPM? or PNPM? or Bun?

...In any case, these are all very specific to the JavaScript world. People are using GPT Pilot with Unity, Python, Kotlin and potentially the next big language/framework.

You could probably include "use yarn for dependency management" in the original prompt, or add the --advanced arg when running and it will ask for further discussion on the technologies proposed.

There's a lot of room for improvement in "advanced" mode. I think that it should (be able to) explain why it's chosen each technology and allow you to have a discussion and either say "ok" or insist on a different choice/s.

#78

[HIMA-MB02]

Yarn is still quite superior than NPM, but Bun is the new kid on the block. It's very promising, but it is still too early to say. They don't have complete 100% support for Node.js yet, as we speak today.

Nevertheless, this is case specific, and I believe the solution you've provided should work.

Alternatively, what do you think about a "de-bugger" persona, that can possibly handle such decisions automatically, without having the developer involved?
Just one additional layer of checks, before the LLM requests for the developer to get involved.

[nalbion]

Yes, we can add other personas, but I think for this time out we could feedback to the LLM that the process timed out. Maybe it will try again with a longer timeout