Bump github.com/posthog/posthog-go from 0.0.0-20230801140217-d607812dee69 to 1.23.1

[dependabot]

Bumps github.com/posthog/posthog-go from 0.0.0-20230801140217-d607812dee69 to 1.23.1.

Release notes

Sourced from github.com/posthog/posthog-go's releases.

1.23.1

Unreleased

1.23.0

Unreleased

1.22.0

Unreleased

1.21.0

Unreleased

1.20.0

Unreleased

1.19.0

Unreleased

1.18.0

Unreleased

1.17.5

Unreleased

1.17.4

Unreleased

1.17.3

Unreleased

1.17.2

Unreleased

1.17.1

Unreleased

1.17.0

Unreleased

1.16.2

Unreleased

1.16.1

Unreleased

1.16.0

Unreleased

1.15.1

Unreleased

... (truncated)

Changelog

Sourced from github.com/posthog/posthog-go's changelog.

1.23.1

Patch Changes

• ec8f6c0: Normalize event timestamps to the equivalent UTC instant before serializing legacy batch and Capture V1 payloads.

1.23.0

Minor Changes

• 941d9ca: Support the starts_with, not_starts_with, ends_with, and not_ends_with property filter operators in feature flag local evaluation. Matching is case-insensitive and mirrors icontains, so flags using these operators no longer fall back to remote evaluation.

1.22.0

Minor Changes

• f102f70: Error tracking stack traces now support server-side symbolication. When the running executable's identity can be determined (GNU build id on Linux, LC_UUID on macOS), the default extractor emits frames with raw instruction addresses and client-expanded inline groups, and exceptions carry a $debug_images property, so PostHog can re-symbolicate against debug symbols uploaded with posthog-cli symbol-sets upload — including exact inline expansion and source context. Without uploaded symbols the runtime-resolved frames are kept as-is, and when the executable can't be identified the frames keep their previous plain shape.

  Requires a PostHog backend with grouped inline frame support (PostHog Cloud, or self-hosted ≥ July 2026). Older backends resolve each group member independently, which duplicates inline expansions when debug symbols are uploaded. To keep the previous plain frames, use a custom StackTraceExtractor.

1.21.0

Minor Changes

• 9e9ad63: Send minimal $feature_flag_called events when the server enables it. When the /flags?v=2 response carries minimalFlagCalledEvents: true (remote evaluation) or the local-evaluation definitions payload carries minimal_flag_called_events: true, and the evaluated flag has has_experiment: false, the event keeps only a strict allowlist of evaluation properties ($feature_flag, $feature_flag_response, $feature_flag_has_experiment, $feature_flag_id, $feature_flag_version, $feature_flag_reason, $feature_flag_request_id, $feature_flag_evaluated_at, $feature_flag_error, locally_evaluated, $groups, $process_person_profile, $geoip_disable, $is_server, $session_id, $window_id, $device_id, $lib, $lib_version) plus static system context ($os, $os_version, $os_distro, $go_version); everything else — including Config.DefaultEventProperties and the snapshot path's $feature/<key> and $feature_flag_payload — is stripped. Any missing signal (gate absent, has_experiment unknown, experiment-linked flag, legacy response shapes) keeps today's full event shape. The gate is server-controlled per project; no SDK configuration is added.

1.20.0

Minor Changes

• 3ea8ce6: Error tracking stack frames are now sent in the canonical cross-SDK wire order: the entry point is first and the crash/capture site is last. Previously frames were sent innermost-first. Coordinated rollout: the ingestion pipeline gates on $lib_version, so this ships as a minor release.

1.19.0

Minor Changes

• 154fd08: Add a $feature_flag_has_experiment boolean property to $feature_flag_called events. The value comes from the has_experiment field the server reports on flag metadata (/flags?v=2) and on local-evaluation flag definitions. The property is only sent when the server explicitly reported has_experiment (true or false); it is omitted when unknown (older deployments, v3 responses, or flags missing from the response).

1.18.0

Minor Changes

• 39c72dc: Drop the newest event instead of blocking when the in-memory queue is full, and make the queue size configurable.

  • Enqueue now returns ErrQueueFull when the queue is full, dropping the newest message rather than blocking the caller until space frees up. This matches posthog-python and posthog-rs. The drop is reported only through the returned error, not through Callback.Failure, so it stays cheap and off the callback goroutines under sustained overload. Backfill/bulk callers: check the error returned by Enqueue for ErrQueueFull and throttle or retry (or raise MaxQueueSize); otherwise events that overflow the queue are dropped, not delayed.
  • Add Config.MaxQueueSize (default DefaultMaxQueueSize = 10000) to control the in-memory message queue capacity independently of BatchSize. It is clamped up to BatchSize so the queue always holds at least one full batch. This replaces the previous hardcoded BatchSize * 10 sizing.
  • Change the default BatchSize (DefaultBatchSize) from 250 to 100, aligning with posthog-python, posthog-node, and posthog-rs. Callers that set BatchSize explicitly are unaffected.

1.17.5

... (truncated)

Commits

• See full diff in compare view

[snyk-io]

✅ Snyk checks have passed. No issues have been found so far.

Status	Scan Engine	Critical	High	Medium	Low	Total (0)
🔚	Open Source Security	0	0	0	0	0 issues
✅	Code Security	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.