language-servers/lsp-server-analysis
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
No description
19 commits 1 branch 0 tags 216 KiB
Find a file
Exact
Josh 1ccc35b5ed add harper
2025-05-12 12:43:56 -05:00
.claude/commands fix path references 2025-05-12 11:03:12 -05:00
servers add harper 2025-05-12 12:43:56 -05:00
.gitignore add local settings to gitignore 2025-05-12 10:10:25 -05:00
README.md split paragraph up 2025-05-12 11:03:42 -05:00
template.md fix path references 2025-05-12 11:03:12 -05:00

README.md

lsp-server-analysis

Use LLMs to query and explore how different LSP implementations manage their state. Contains a template for the LLM to fill out and (hopefully) keep it on track.

Take all of these analyses with a grain-of-salt obviously. LLMs lie/hallucinate/make shit up constantly – even with the source available.

My hope with the template and the command prompt is to minimize this as much as possible, as well as just the fact that this is an exploration and explaination of existing code – not writing anything new. Still, LLMs be LLMs and it's just a part of how they are built that they are going to lie their ass off and not bat an eye.

These are purely meant for educational and research purposes. I may be tilting at windmills hoping for more than your run-of-the-mill AI slop, but I have found it useful to at least getting a starting point for my own exploration.

Structure

  • .claude/commands/lsp_analysis.md - A Claude Code command to kick off or resume an LSP server exploration
  • servers/ - Analysis of individual LSP server implementations
  • template.md - Template for LLM exploration of LSP server implementations

Questions

  1. How is state structured and organized?
  2. What specific types of data are stored in the server state?
  3. How is thread safety handled?
  4. How is document state separated from project/workspace state?
  5. What patterns are used for immutability and concurrency?
  6. How is configuration managed and resolved?