fix(log): redirect the run log when the first log stream comes back empty

[vdusek]

call({ log }) and run.getStreamedLog() sometimes redirected no log lines at all. That made the integration test call({ log: "default" }) streams the run log to the console while the run executes fail now and then in CI (4 times so far, each green on retry).

The API answers GET /logs/:id?stream=1 with an empty 200 and closes it right away when the run hasn't written its first line yet (apify-core src/api/src/routes/logs/log.ts). StreamedLog opened the stream once, right after start(), so it took that empty response as the whole log. Hitting the API directly reproduced it in 1 of 20 freshly started apify/hello-world runs.

StreamedLog now reopens a stream that ended without any bytes, after a 500 ms pause that stop() cuts short. If stop() comes before anything has arrived, including while a reopened stream is still connecting, it reads the log once with a plain get(). A run that finishes quickly still gets its whole log redirected, and a stream opened during stop() can't cut it off after the first chunk.

Tests cover five cases, each one also in apify/apify-client-python#1084:

• the first stream is empty, so the log has to come through the reopened one (getStreamedLog())
• every stream stays empty until the run finishes, so stop() reads the whole log (call())
• stop() lands while the reopened stream is connecting
• a missing log is neither reopened nor read
• a failing one-shot read is logged and stop() still resolves

The first three fail without the fix. The live integration test passed 30 of 30 runs afterwards.

The Python client had the same bug: apify/apify-client-python#1084.

✍️ Drafted by Claude Code

[github-actions]

⚠️ There are broken links in the documentation.

See more at https://github.com/apify/apify-client-js/actions/runs/37321221161#summary-111800489211