Wait did he say you can adjust the context passed to a prompt? Is this a manual process where I could specify the line numbers in specific files? Or do I have to include that in my prompt? And if it's by prompt, can I just include file paths and line numbers in my prompt?
@workspace when activated does not need to paste code snippet just highlight the code in the file such as app.js and tell workspace that code is highlighted. This is enough for the workspace to know the context of the code and you will receive relevant code suggestions/explanations to your questions. I have experimented this and would like to know anything otherwise. thanks
How does this work when, inside of the project, you have put something like a TON of json files or something like that and increase the context size? For example: Someone is testing a machine learning model and just puts the data in the main project folders.
@SnowyMango He explained that. The reason is about the context. You do not want to feed a natural language processing system the entire project if you just have a small question about the code you are working with. Currently the Copilot uses what is selected (on the current file) or the selections around your cursor. The feeding of the entire project takes much longer (especially on huge projects) and could be very helpful sometimes but not others. It is all about context and realizing what context you are asking your question in.
Wait did he say you can adjust the context passed to a prompt? Is this a manual process where I could specify the line numbers in specific files? Or do I have to include that in my prompt? And if it's by prompt, can I just include file paths and line numbers in my prompt?
@workspace when activated does not need to paste code snippet just highlight the code in the file such as app.js and tell workspace that code is highlighted. This is enough for the workspace to know the context of the code and you will receive relevant code suggestions/explanations to your questions. I have experimented this and would like to know anything otherwise. thanks
How does this work when, inside of the project, you have put something like a TON of json files or something like that and increase the context size? For example: Someone is testing a machine learning model and just puts the data in the main project folders.
is this feature opensource? or do we need to pay for it?
Why make the developer type @workspace. Why not just do it automatically in the background.
@SnowyMango He explained that. The reason is about the context. You do not want to feed a natural language processing system the entire project if you just have a small question about the code you are working with. Currently the Copilot uses what is selected (on the current file) or the selections around your cursor. The feeding of the entire project takes much longer (especially on huge projects) and could be very helpful sometimes but not others. It is all about context and realizing what context you are asking your question in.
A video about workspaces that most developers don’t have access to
It's about "@workspace" the tag in normal copilot. It's not about the new feature that many don't have yet.