You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This commit was created on GitHub.com and signed with GitHub’s verified signature.
Patch Changes
46ce85b: agent get --follow and agent send --follow now wait on the QA Wolf API for news instead of asking again every few seconds: one session waits with a held agent.get from where the last read ended, and --sessions waits on every session in one held activity.wait. runner events --follow and runner run --follow read every second while entries arrive and slow down to every five seconds while nothing does. --sessions now takes at most 50 session ids.
620f38b: Generated commands can now take a top-level input that holds objects and can't be split into flags as one --<name> <json> flag, instead of failing to build. The flag also takes @<path> to read the JSON from a file, or - to read it from stdin. The value is checked against the command's input before anything is sent. No command has such an input with @qawolf/api-contracts 0.116.0, so no command's flags change in this release.
c2a6262: npm install @qawolf/cli and npx @qawolf/cli finish again instead of running forever at full CPU. The CLI no longer installs @qawolf/flows, @qawolf/testkit and Playwright next to itself; it already installed its own copies before each run.
b81035a: When nothing was sent because the connection to the QA Wolf API never opened, the CLI now retries the request, including actions like runner act.