Bug: Claude API Change?

[TWDickson]

Looks like Anthropic has updated Claude's API, this extension and a few other similar command line plugins are failing for me,

Loading organization and project from config...
Error in sync files: API request failed: getaddrinfo ENOTFOUND api.claude.ai (unknown status)
Stack trace: Error: API request failed: getaddrinfo ENOTFOUND api.claude.ai (unknown status)
    at Pa.makeRequest (/Users/MYUSERNAME/.vscode/extensions/rexdotsh.claudesync-0.1.7/dist/extension.js:15:6572)
    at processTicksAndRejections (node:internal/process/task_queues:95:5)
    at Pa.getOrganizations (/Users/MYUSERNAME/.vscode/extensions/rexdotsh.claudesync-0.1.7/dist/extension.js:15:6810)
    at Aa.ensureProjectAndOrg (/Users/MYUSERNAME/.vscode/extensions/rexdotsh.claudesync-0.1.7/dist/extension.js:15:10201)
    at /Users/MYUSERNAME/.vscode/extensions/rexdotsh.claudesync-0.1.7/dist/extension.js:15:13423
    at Aa.handleError (/Users/MYUSERNAME/.vscode/extensions/rexdotsh.claudesync-0.1.7/dist/extension.js:15:9423)
    at /Users/MYUSERNAME/.vscode/extensions/rexdotsh.claudesync-0.1.7/dist/extension.js:15:21491
Failed to sync files: Failed to sync files: API request failed: getaddrinfo ENOTFOUND api.claude.ai (unknown status)

[rexdotsh]

Yes, please try updating to the latest version that I released yesterday :) v0.1.8

[TWDickson]

That worked, thanks for the quick response. Do you have any idea about implementing file packing like the original command line tool?

[rexdotsh]

Looked into this a little, I see the original CLI lets you set a compression algorithm and uploads files like that.
I'm curious as to how that works with claude, are you sure it can still process the compressed content properly and answer based on that?

[TWDickson]

I've only used the 'pack' option which basically merges everything into a single file with text indications for the starting and ending boundary of a file with its relative path. Claude seems to interpret this fine and responds with specific file only updates.

[rexdotsh]

Hmmm, what's the benefit of this though? Can't quite understand, does it take less space in the project that way?

[TWDickson]

I believe it does, I tested it on my one project and reduced the project size by a few percentage points. Likely due to the reduced indexing of multiple files. One thing I've noticed is that it seems there's a difference between how it interprets the files. With multiple files Claude searches but I think with the packed version it's forced to interpret the entire file and digest the entire project.