main.js · part 16

60 text occurrences from desktop/Cursor.app/Contents/Resources/app/extensions/cursor-retrieval/dist/main.js, part 16. Every entry preserves the shipped literal and its saved verdict or selection reason.

File contents and all parts · All files

Received network error or non-101 status code.

Server did not respond with sent protocols.

Server did not set Upgrade header to "websocket".

Server did not set Connection header to "upgrade".

Incorrect hash received in Sec-WebSocket-Accept header.

Received different permessage-deflate than the one set.

Protocol was not set in the opening handshake.

use strict

cannot get uid or gid

use strict

use strict

Every keystroke ships the file body, filename and edit history to Cursor's Tab model.

Sends the instruction, selection, file and reranked context to Cursor-managed models.

Sends the first user message to Cursor; chats keep the local fallback name.

Sends the staged diff to Cursor.

Sends the diff or prompt to Cursor.

Uploads the full conversation for analytics.

The overview is generated server-side from the conversation text.

Streams the working-tree diff to Cursor.

Uploads file contents and paths to Cursor's embedding service.

Uploads a local agent run (every model invocation) to Cursor.

Publishes the transcript to Cursor.

Knowledge base items and agent memories are stored on Cursor's servers.

The agent's web fetch tool sends the URL to Cursor; the private-inference tool client refuses it.

The agent's image generation tool sends the description and reference images to Cursor; the private-inference tool client refuses it.

Pull-request metadata, diffs, checks, comments and merge actions go through Cursor's SCM service, as does the GitHub/GitLab connection probe.

After an agent turn edits files, uploads each file's path and diff to Cursor, plus the added lines and original contents when the server config asks for them.

Uploads git packfiles of the workspace's repositories to Cursor and reports which repositories the workspace contains.

Pull-request, issue and commit context is fetched from GitHub with the user's GitHub token, proxied through Cursor when the gate is on; all GitHub egress stays off until it is decided.

Private method is not writable

Private accessor was defined without a setter

Cannot write private member to an object whose class did not declare it

Private accessor was defined without a getter

Cannot read private member from an object whose class did not declare it

[${n}] ${a}: ${e}

AbortController is not defined. If using lru-cache in node 14, load an AbortController polyfill from the `node-abort-controller` package. A minimal polyfill is provided for use by LRUCache.fetch(), but it should not be relied upon in other contexts (eg, passing it to other APIs that use AbortController/AbortSignal might have undesirable effects). You may disable this with LRU_CACHE_IGNORE_AC_WARNING=1 in the env.

instantiate Stack using Stack.create(n)

cannot set size without setting maxSize or maxEntrySize on cache

perf option must have a now() method if specified

max option must be a nonnegative integer

invalid max value: ${a}

cannot set sizeCalculation without setting maxSize or maxEntrySize

sizeCalculation set to non-function

memoMethod must be a function if defined

fetchMethod must be a function if specified

maxSize must be a positive integer if specified

maxEntrySize must be a positive integer if specified

ttl must be a positive integer if specified

At least one of max, maxSize, or ttl is required

TTL caching without ttlAutopurge, max, or maxSize can result in unbounded memory consumption.

invalid size value (must be positive integer). When maxSize or maxEntrySize is used, sizeCalculation or size must be set.

sizeCalculation must be a function

sizeCalculation return invalid (expect positive integer)

fetch() returned undefined

no memoMethod provided to constructor

 ${e.name}

Invalid Date

${i} ${n.dim("=>")} ${l}

... ${r.length-i} more

${n.cyan(e)}: ${i}