fix(matrix): truncate thread starter body on code-point boundaries by ly-wang19 · Pull Request #96407 · openclaw/openclaw

ly-wang19 · 2026-06-24T12:36:13Z

What Problem This Solves

truncateThreadStarterBody in extensions/matrix/src/matrix/monitor/thread-context.ts (reached via the exported summarizeMatrixThreadStarterEvent) trims an over-length Matrix thread-root body with a raw UTF-16 slice:

return `${value.slice(0, MAX_THREAD_STARTER_BODY_LENGTH - 3)}...`;

String.prototype.slice operates on UTF-16 code units, so when the 500-char boundary lands in the middle of an astral character (emoji, many CJK extensions, etc.), the surrogate pair is cut in half. The result is a lone, unpaired high surrogate immediately before the ellipsis, which renders as a broken/replacement glyph (U+FFFD).

Repro — a thread-root m.room.message whose content.body is 496 a characters, then 😀 (U+1F600 — a surrogate pair at UTF-16 indices 496–497), then bcd (total length 501):

summarizeMatrixThreadStarterEvent({
  type: "m.room.message",
  content: { msgtype: "m.text", body: "a".repeat(496) + "😀" + "bcd" },
})

Before (wrong): slice(0, 497) cuts the emoji in half, leaving a lone high surrogate: "a".repeat(496) + "\uD83D" + "..." — a broken glyph followed by the ellipsis.
After (fixed): truncation stops on a code-point boundary, dropping the half emoji entirely: "a".repeat(496) + "..." — 496 as then the ellipsis, no orphaned surrogate.

The fix swaps the raw slice for the existing SDK helper sliceUtf16Safe (from openclaw/plugin-sdk/text-utility-runtime), which already backs the surrogate-safe truncation used by the Discord, Slack, Telegram, and other adapters. Behavior is unchanged for short bodies and for long bodies whose cut point is not inside a surrogate pair.

Evidence

A standalone Node script replicates the old (buggy) and new (fixed) truncateThreadStarterBody logic and asserts the red→green transition plus that unaffected inputs are unchanged:

PASS repro input length is 501
PASS RED: old code leaves a lone (orphaned) high surrogate
PASS RED: old code emits half-emoji + ellipsis
PASS GREEN: fixed code drops the half emoji entirely
PASS GREEN: fixed code has NO orphaned surrogate
PASS short body unchanged (old)
PASS short body unchanged (fixed)
PASS long ASCII truncation identical old vs fixed
PASS long ASCII fixed output is 497 x + ellipsis
PASS emoji before cut survives intact
PASS emoji-before-cut output has no lone surrogate

ALL CHECKS PASSED

RED confirms the pre-fix code emits "a".repeat(496) + "\uD83D" + "..." (a lone high surrogate).
GREEN confirms the fixed code emits "a".repeat(496) + "..." with no orphaned surrogate.
Unaffected inputs (short body, long ASCII body, an emoji that sits safely before the cut) are byte-for-byte unchanged.

A matching unit test was added to extensions/matrix/src/matrix/monitor/thread-context.test.ts that calls the exported summarizeMatrixThreadStarterEvent directly with the repro body and asserts the surrogate-safe output ("a".repeat(496) + "..." with no surrogate code unit surviving).

truncateThreadStarterBody (reached via summarizeMatrixThreadStarterEvent) truncated the thread-root body with a raw `value.slice(0, 497)`. When the 500-char boundary falls in the middle of an astral character (emoji etc.), the raw slice cuts the surrogate pair in half and leaves a lone high surrogate before the ellipsis, rendering as a broken/replacement glyph. Use the SDK's surrogate-safe `sliceUtf16Safe` so truncation stops on a code-point boundary, dropping the half character entirely. Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

openclaw-barnacle · 2026-06-24T12:37:33Z

Closing this PR because the author has more than 20 active PRs in this repo. Please reduce the active PR queue and reopen or resubmit once it is back under the limit. You can close your own PRs to get back under the limit.

@ly-wang19

…97121) Matrix thread-starter previews truncated long bodies by raw UTF-16 slice, which could cut an astral character (e.g. emoji) and leave a lone surrogate, rendering mojibake in the agent's thread context. Reuse the existing sliceUtf16Safe helper so the cut backs up to a valid surrogate boundary, preserving the 500-code-unit limit and '...' suffix. Adds a regression test that fails against the raw-slice implementation. Salvages the original fix from #96407 (auto-closed by the active-PR queue cap). Preserves @ly-wang19's authorship; rebased clean onto main by @Bartok9. Co-authored-by: ly-wang19 <[email protected]>

vincentkoc · 2026-06-26T23:28:43Z

Superseded by the clean replacement PR #97121, now merged to main as 289865b39226c685af382a81da55c38ce3df9049.

The landed diff preserves the original Matrix fix and contributor credit, touching only extensions/matrix/src/matrix/monitor/thread-context.ts and its regression test. Exact-main proof passed: node scripts/run-vitest.mjs extensions/matrix/src/matrix/monitor/thread-context.test.ts (6/6), and git diff --check was clean.

@ly-wang19

…penclaw#97121) Matrix thread-starter previews truncated long bodies by raw UTF-16 slice, which could cut an astral character (e.g. emoji) and leave a lone surrogate, rendering mojibake in the agent's thread context. Reuse the existing sliceUtf16Safe helper so the cut backs up to a valid surrogate boundary, preserving the 500-code-unit limit and '...' suffix. Adds a regression test that fails against the raw-slice implementation. Salvages the original fix from openclaw#96407 (auto-closed by the active-PR queue cap). Preserves @ly-wang19's authorship; rebased clean onto main by @Bartok9. Co-authored-by: ly-wang19 <[email protected]>

@ly-wang19

…penclaw#97121) Matrix thread-starter previews truncated long bodies by raw UTF-16 slice, which could cut an astral character (e.g. emoji) and leave a lone surrogate, rendering mojibake in the agent's thread context. Reuse the existing sliceUtf16Safe helper so the cut backs up to a valid surrogate boundary, preserving the 500-code-unit limit and '...' suffix. Adds a regression test that fails against the raw-slice implementation. Salvages the original fix from openclaw#96407 (auto-closed by the active-PR queue cap). Preserves @ly-wang19's authorship; rebased clean onto main by @Bartok9. Co-authored-by: ly-wang19 <[email protected]>

@ly-wang19

…penclaw#97121) Matrix thread-starter previews truncated long bodies by raw UTF-16 slice, which could cut an astral character (e.g. emoji) and leave a lone surrogate, rendering mojibake in the agent's thread context. Reuse the existing sliceUtf16Safe helper so the cut backs up to a valid surrogate boundary, preserving the 500-code-unit limit and '...' suffix. Adds a regression test that fails against the raw-slice implementation. Salvages the original fix from openclaw#96407 (auto-closed by the active-PR queue cap). Preserves @ly-wang19's authorship; rebased clean onto main by @Bartok9. Co-authored-by: ly-wang19 <[email protected]>

@ly-wang19

…penclaw#97121) Matrix thread-starter previews truncated long bodies by raw UTF-16 slice, which could cut an astral character (e.g. emoji) and leave a lone surrogate, rendering mojibake in the agent's thread context. Reuse the existing sliceUtf16Safe helper so the cut backs up to a valid surrogate boundary, preserving the 500-code-unit limit and '...' suffix. Adds a regression test that fails against the raw-slice implementation. Salvages the original fix from openclaw#96407 (auto-closed by the active-PR queue cap). Preserves @ly-wang19's authorship; rebased clean onto main by @Bartok9. Co-authored-by: ly-wang19 <[email protected]>

openclaw-barnacle Bot added channel: matrix Channel integration: matrix size: XS r: too-many-prs Auto-close: author has more than twenty active PRs. labels Jun 24, 2026

openclaw-barnacle Bot closed this Jun 24, 2026

This was referenced Jun 26, 2026

fix(matrix): truncate thread starter body on code-point boundaries #96961

Closed

fix(matrix): truncate thread starter body on code-point boundaries #97121

Merged

clawsweeper Bot mentioned this pull request Jun 29, 2026

fix(matrix): truncate thread starter body on UTF-16 boundary #97596

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(matrix): truncate thread starter body on code-point boundaries#96407

fix(matrix): truncate thread starter body on code-point boundaries#96407
ly-wang19 wants to merge 1 commit into
openclaw:mainfrom
ly-wang19:fix/matrix-thread-starter-surrogate-safe

ly-wang19 commented Jun 24, 2026

Uh oh!

openclaw-barnacle Bot commented Jun 24, 2026

Uh oh!

vincentkoc commented Jun 26, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

ly-wang19 commented Jun 24, 2026

What Problem This Solves

Evidence

Uh oh!

openclaw-barnacle Bot commented Jun 24, 2026

Uh oh!

vincentkoc commented Jun 26, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants