Ayuriel/openclaw
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: address remaining review feedback on followup media

Browse Source 
- Scope file-block stripping to body region only, preserving file blocks
  in quoted/replied text and thread history above the body
- Gate file-extraction skip on mutation evidence (Body differs from
  resolved original) instead of raw '<file name=' text matching to avoid
  false-positives on user messages containing literal XML
- Document collect-mode scope limitation in applyDeferredMediaUnderstandingToQueuedRun
- Rename duplicate test description to distinguish body-alongside-text case
- Prefer updated prompt over original summaryLine in overflow summaries
  so captioned voice note transcripts are surfaced
This commit is contained in:
Joey Krug 2026-03-14 23:51:22 -04:00
parent ad0a1bdc5e
commit f890bc75dc
2 changed files with 23 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
src/auto-reply/reply/followup-media.ts
View File

@ -92,13 +92,21 @@ function rebuildQueuedPromptWithMediaUnderstanding(params: {
stripped = stripLeadingMediaReplyHint(stripped);
}
// Strip pre-existing file blocks from the prompt when the updated body
// Strip pre-existing file blocks from the body region when the updated body
// contains new file blocks. Mixed messages (audio + PDF) can arrive with
// file extraction already applied in the primary path; without this strip
// the old block stays in the prompt while the updated body adds a new one,
// duplicating potentially large file payloads.
// Scope stripping to the body segment so quoted/replied text and thread
// history above the body retain any legitimate <file> blocks.
if (params.updatedBody && FILE_BLOCK_RE.test(params.updatedBody)) {
stripped = stripExistingFileBlocks(stripped);
const bodyTarget = params.originalBody?.trim();
const bodyIdx = bodyTarget ? stripped.lastIndexOf(bodyTarget) : -1;
if (bodyIdx >= 0) {
stripped = stripped.slice(0, bodyIdx) + stripExistingFileBlocks(stripped.slice(bodyIdx));
} else {
stripped = stripExistingFileBlocks(stripped);
}
}
const updatedBody = normalizeUpdatedBody({
@ -176,6 +184,9 @@ export async function applyDeferredMediaUnderstandingToQueuedRun(
queued: FollowupRun,
params: { logLabel?: string } = {},
): Promise<void> {
// NOTE: collect-mode and overflow-summary queue drains create synthetic
// followup runs without mediaContext — those paths are not covered here
// and rely on their own prompt-building logic in queue/drain.ts.
const mediaContext = queued.mediaContext;
if (!mediaContext || mediaContext.DeferredMediaApplied) {
return;
@ -191,10 +202,13 @@ export async function applyDeferredMediaUnderstandingToQueuedRun(
const resolvedOriginalBody =
mediaContext.CommandBody ?? mediaContext.RawBody ?? mediaContext.Body;
const bodyAlreadyHasFileBlock =
FILE_BLOCK_RE.test(resolvedOriginalBody ?? "") || FILE_BLOCK_RE.test(mediaContext.Body ?? "");
// Detect file extraction from the primary path by checking whether Body was
// mutated AND contains file blocks. This avoids false-positives when a user
// message literally contains "<file name=" text.
const fileExtractionAlreadyApplied =
mediaContext.Body !== resolvedOriginalBody && FILE_BLOCK_RE.test(mediaContext.Body ?? "");
if (bodyAlreadyHasFileBlock && hasOnlyFileLikeAttachments(mediaContext)) {
if (fileExtractionAlreadyApplied && hasOnlyFileLikeAttachments(mediaContext)) {
mediaContext.DeferredMediaApplied = true;
return;
}
@ -227,7 +241,7 @@ export async function applyDeferredMediaUnderstandingToQueuedRun(
muResult.appliedAudio ||
muResult.appliedImage ||
muResult.appliedVideo ||
(muResult.appliedFile && !bodyAlreadyHasFileBlock);
(muResult.appliedFile && !fileExtractionAlreadyApplied);
if (shouldRebuildPrompt) {
const newMediaNote = buildInboundMediaNote(mediaCtx);

4
src/auto-reply/reply/queue/drain.ts
View File

@ -85,7 +85,9 @@ async function resolveSummaryLines(items: FollowupRun[]): Promise<string[]> {
const summaryLines: string[] = [];
for (const item of items) {
await applyDeferredMediaUnderstandingToQueuedRun(item, { logLabel: "followup queue" });
summaryLines.push(buildQueueSummaryLine(item.summaryLine?.trim() || item.prompt.trim()));
// After deferred media, prefer the updated prompt (which includes transcripts)
// over the original summaryLine (which may just be the caption text).
summaryLines.push(buildQueueSummaryLine(item.prompt.trim() || item.summaryLine?.trim() || ""));
}
return summaryLines;
}