platform · inferred from evidence
GitHub
A platform for hosting and collaborating on code repositories.
- Kestra: every script execution fails with Docker runner (script file not accessible)
Teaches the verified fix from a real kestra support thread: kestra: every script execution fails with Docker runner (script file not accessible)
- Latest nixpkgs breaks cbcli build: undefined method 'wait_readable' for TCPSocket
The nix build broke because the flake pulled Crystal 1.15, which removed something the old vendored ssh2 code relied on, hence the undefined method wait_readable for TCPSocket. The maintainer said locking the flake to Crystal below 1.14 was
- imgix.js ix-path attributes ignored in Vue.js apps (use the official Vue library)
Teaches the verified fix from a real imgix support thread: imgix.js ix-path attributes ignored in Vue.js apps (use the official Vue library)
- Drone/Harness runner fails on Docker 29: client version 1.41 is too old, minimum 1.44
Teaches the verified fix from a real harness support thread: drone/Harness runner fails on Docker 29: client version 1.41 is too old, minimum 1.44
- Cloud SQL Proxy: "Error during createEphemeral ... Error 403: The client is not authorized to make this request"
This is a known IAM propagation quirk, not a proxy bug: re-grant the role to force it to take effect. Remove the Cloud SQL Client role from the failing service account in IAM & Admin > IAM (uncheck it and Save), verify the account disappear
- Firecrawl fullPage screenshots randomly capture only the top of the page
Teaches the verified fix from a real firecrawl support thread: firecrawl fullPage screenshots randomly capture only the top of the page
- Fauna FetchError: socket hang up on requests to db.fauna.com (intermittent)
Teaches the verified fix from a real fauna support thread: fauna FetchError: socket hang up on requests to db.fauna.com (intermittent)
- KEY ERROR 1 while using classification preset in Evidently
Teaches the verified fix from a real evidently-ai support thread: kEY ERROR 1 while using classification preset in Evidently
- SandboxError "Invalid redirect value, must be one of follow or manual" running E2B on Vercel Edge
Teaches the verified fix from a real e2b support thread: sandboxError "Invalid redirect value, must be one of follow or manual" running E2B on Vercel Edge
- sync-wasm connect() fails with "unexpected path ... .db-log: files must be created in advance for OPFS IO"
Fixed in PR #8524, which registers the MVCC log path with the OPFS worker before the sync engine connects. Use `@tursodatabase/sync-wasm` 0.8.0-pre.9 or newer.
- Access a private GitHub repository from a Cloud Build step (glide install fails)
How to fix: Access a private GitHub repository from a Cloud Build step (glide install fails). The verified fix from the maintainer thread, distilled into reusable steps.
- docusign-php-client: ApiException#getResponseObject returns serialized PHP string instead of object
Upgrade to v5.5, where the maintainer fixed it: composer require docusign/esign-client:^5.5 (the fix was first available in the 5.5.0-RC). The DocuSign team confirmed the fix on the issue before closing it.
- doppler run no longer overrides local environment variables regression
Teaches the verified fix from a real doppler support thread: doppler run no longer overrides local environment variables regression
- atlassian-python-api `delete_space` on a missing space raises generic ApiError instead of ApiNotFoundError
You were right: the library was swallowing the 404 into a generic `ApiError`. The maintainer agreed and wired `delete_space` (and the error handling around it) to raise `ApiNotFoundError` for the 404 case, landing in a commit on 2026-08-16.
- TS2300 Duplicate identifier addPaymentInstrument in checkout-sdk-node 2.1.0
A contributor PR removed the duplicate declaration and it was merged and released; upgrade checkout-sdk-node past 2.1.0 to pick up the fix. The maintainer confirmed the merge and release in the thread.
- received 1011 (internal error): Deepgram did not receive audio data within the timeout window
Teaches the verified fix from a real deepgram support thread: received 1011 (internal error): Deepgram did not receive audio data within the timeout window
- Azure App Configuration throttled HTTP 429 even though hourly request count is far below limit
Throttling is not only hourly: you can get 429s from exceeding the daily limit on the Free tier, the hourly limit on Standard, momentary throttling from a large burst of requests, excessive bandwidth, or hitting the storage quota. A burst o
- Hasura: metadata inconsistency "Found conflicting definitions for GraphQL type" after upgrade to v2.49.5
This was a v2.49.5 regression in the graphql-default naming convention handling , upgrade to v2.50.0, where the reporter confirmed the behavior is gone (maintainer hgiasac suggested exactly that and the reporter closed the issue as resolved
- Codefresh CLI: log following breaks after node version bump
Teaches the verified fix from a real codefresh support thread: codefresh CLI: log following breaks after node version bump
- Firebase deploy fails: "Cloud Runtime Config is currently experiencing issues, which is preventing your functions from
The root cause is server-side on Google's end: firebase deploy calls the Cloud Runtime Configuration API (runtimeconfig.googleapis.com) during functions deploys, and when that API 500s the deploy halts.