Preserve captures when the writer queue fills

[LightCyan01]

Combat observations could be discarded when the writer reached its queue limit. The writer now waits for space, uses faster gzip compression during recording, and releases waiting callers if output fails. Capture construction is serialized before copying bodies or snapshots so concurrent callbacks do not accumulate detached data while waiting.

Preserve a single oversized record when the budget is empty, document the remaining storage and snapshot limits, and prepare version 1.3.2. The ZIP format and final compression remain compatible with existing recordings.

[greptile-apps]

Confidence Score: 5/5

[High impact] The PR appears safe to merge; the previous memory-budget finding is addressed.

Summary

The PR preserves captures when the writer queue fills and prepares version 1.3.2.

• Capture records wait for space instead of disappearing.
• Capture callbacks build records one at a time.
• Recording compresses gzip blocks faster.
• The capture plugin is labeled version 1.3.2.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[Callback arrives] --> B[Wait for observation gate]
    B --> C[Copy body or build snapshot]
    C --> D[Wait for writer space]
    D --> E[Queue record and release gate]
    E --> F[Writer saves record]
    F --> G[Release retained bytes]
    G --> D

Loading

_{Reviews (2) · Last reviewed commit: "Bound capture construction across concur..." · Reviewed by Greptile}