test: consolidate polling into a single shared helper and deflake tests

[vdusek]

Summary

Started as a generalization of the two integration polling helpers, and grew into the SDK counterpart of apify/apify-client-python#844: a single shared polling helper used across the whole test suite.

Changes:

• Add a shared tests/_utils.py with a single poll_until_condition helper — identical to the one in apify-client-python (chore(deps): update codecov/codecov-action action to v6 #844): polls a sync-or-async callable at a constant interval until a condition holds or a wall-clock timeout expires, with an optional backoff_factor that multiplies the interval after each poll. This subsumes the former call_with_exp_backoff (backoff_factor=2) and the "call once" single-RQ-mode case (timeout=0). The duplicated generate_unique_resource_name (integration + e2e copies) moves there too; tests/integration/_utils.py and tests/e2e/_utils.py are removed.
• Migrate all ~50 call_with_exp_backoff call sites in test_request_queue.py to poll_until_condition (timeout=30, backoff_factor=2 for shared mode ≈ the former exponential backoff; timeout=0 for single mode). Pre-existing poll_until_condition call sites get their former defaults (timeout=60, poll_interval=5) passed explicitly.
• Fix the flaky test_actor_adds_webhook_and_receives_event e2e test (failing CI run): the client Actor exited ~170 ms after registering the ad-hoc webhook, so the platform could process the run-succeeded event before the webhook propagated, leaving the server Actor blocked until its run timeout. The client now stays alive 5 s after add_webhook, and the unbounded INITIALIZED startup polling loop is replaced with poll_until_condition (timeout=300, growing interval). Replaces test: fix flaky webhook E2E test #930.
• Replace the hand-rolled retry_counter loops in test_actor_charge.py (4×) with poll_until_condition.
• Replace fixed sleeps in unit tests (test_actor_lifecycle, test_actor_key_value_store, test_apify_event_manager) with condition polling where the test waits for something to happen — faster on the happy path and robust under xdist load.

Fixed sleeps that are semantically required are intentionally left as-is: negative checks ("event must NOT fire"), mtime-granularity checks in test_apify_storages, simulated latency, and sleeps inside Actor main_func bodies (which are serialized and deployed to the platform, where the helper is not available).

Verification: lint, type-check, and the full unit suite pass; the three touched unit test files pass 10/10 repeated runs. Integration/e2e tests require a platform token and will be verified by CI.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 86.94%. Comparing base (695500e) to head (fffc7eb).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #920   +/-   ##
=======================================
  Coverage   86.94%   86.94%           
=======================================
  Files          48       48           
  Lines        2942     2942           
=======================================
  Hits         2558     2558           
  Misses        384      384           

Flag	Coverage Δ
e2e	37.55% <ø> (ø)
integration	58.90% <ø> (ø)
unit	75.69% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.