[CODEX-SYNC] deferred 4 Codex

Quick fix

Upgrade cc-switch to v3.20.3 to fix Windows session files permanently deferred when mtime does not change.

Symptom

Error outputCopy

[CODEX-SYNC] 同步完成: 导入 0 条, 跳过 0 条, deferred 4 个, 扫描 87 个文件

English translation

[CODEX-SYNC] Sync complete: imported 0 entries, skipped 0 entries, deferred 4, scanned 87 files

Cause

On Windows, Codex session files being written may not continuously update mtime. The synchronizer records the mtime and offset before a file is finished; if mtime remains unchanged after completion, later runs treat it as unchanged and skip it, permanently preventing deferred files and subsequent usage from being imported.

This is not triggered by file size or model name. The fixed version uses a persistent byte cursor and determines file changes from both mtime and file size.

Fix

1. 1

   Upgrade cc-switch to v3.20.3.

2. 2

   After upgrading, wait for Codex usage sync to rescan the previously stuck rollout session files; the file cursor has been verified to catch up and restore missed usage.

Affected Versions

ToolCodex

Versionv3.19.1（受影响，v3.20.3 已修复）

PlatformsWindows

Source Issues

This page is distilled from 1 real issues

• #6023Codex 用量统计停止记录：4 个会话文件永久卡在 deferred，同步一直导入 0 条（v3.19.1 / Windows）

FAQ

(No FAQ yet)

Related problems

Codex session import ignores cache_write_input_tokensUpgrade cc-switch or wait for an official fix: Codex session log import misses cache_write_input_tokens, so cache writes are counted as ordinary input and…1 issues mergedCost always 0: GPT-6 Sol usage statistics in CodexEnable “Model Price Sync” in Settings to synchronize pricing data for gpt-6-sol.1 issues mergedcc-switch Usage Statistics Differ from DeepSeek Official Site / Stop UpdatingUpgrade to v3.19.2 or later, then run “Rebuild Codex Usage” once on the “Usage Statistics” page to correct historical overcounts or missing records.22 issues merged