[codex] fix diagnostic cache key collisions

[Sean-Kenneth-Doherty]

Summary

• key the hover Markdown cache by diagnostic message, code, range, and related information instead of message alone
• use the same key for the sidebar HTML cache so symbol links and action ranges stay tied to the correct diagnostic
• add regression coverage for duplicate messages with different symbol locations and ranges

Fixes #143

Validation

• npx prettier --write apps/vscode-extension/src/diagnosticCacheKey.ts apps/vscode-extension/src/diagnostics.ts apps/vscode-extension/src/provider/webviewViewProvider.ts apps/vscode-extension/src/test/suite/diagnosticCacheKey.test.ts
• npm run lint --workspace pretty-ts-errors
• npm run compile-tests --workspace pretty-ts-errors
• npx mocha apps/vscode-extension/out/test/suite/diagnosticCacheKey.test.js --ui tdd
• npm run compile --workspace pretty-ts-errors
• git diff --check

AI assistance: prepared with OpenAI Codex.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[kevinramharak]

Ill check this one out, I think it solves the issue, just want to check it locally for edge cases