[HOLD] Update SQLite to branch hctree-bedrock-lcd-ex

[flodnv]

Details

Fixed Issues

Redo of #2605 with a bug fix:

contains a fix to a problem that could cause certain kinds of UPDATE (and maybe other) statements to erroneously record full table scans. It also emits a log message each time a BEGIN CONCURRENT transaction records a full-table scan of an IPK table. Hopefully that won't be too many log messages..

Needs #2619 which was broken out in another PR.

Tests

---

Internal Testing Reminder: when changing bedrock, please compile auth against your new changes

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 7c74be1d1f

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Stop replaying writes after an error

During a BEGIN CONCURRENT logical commit, this block replays every recorded write but the inner loop keeps running after btreeCursorOpen(), btreeMoveto(), or sqlite3BtreeInsert() returns an error, and the delete path drops the sqlite3BtreeDelete() return value entirely. If any replayed write hits an IO/OOM/corruption error while a later write succeeds, rc can be overwritten back to SQLITE_OK and the commit can complete after applying only part of the transaction's write set.

Useful? React with 👍 / 👎.