fix(telegram): split progress from final replies#732
Open
artemgetmann wants to merge 12 commits into
Open
Conversation
added 11 commits
May 16, 2026 15:23
- what changed: materialize retained progress narration as its own Telegram bubble and send final answers separately. - why: progress UX should preserve LLM task narration without polluting the final answer or TTS target. - behavior/risk: scoped to Telegram text progress projection; live v2 proof is still blocked by provider auth/limits.
- What: materialize retained Telegram progress before sending final answers and strip separator-delimited progress from final text. - Why: users need deterministic progress bubble + final bubble layout instead of edited or duplicated answer previews. - Risk: limited to Telegram answer-stream finalization; media and error finals keep the existing safer path.
- What: treat natural multi-paragraph answer partials as retained progress transcripts and restore prior progress when late partials append final text. - Why: Telegram progress UX must use deterministic two-bubble layout without relying on model-written Step/checkbox narration. - Risk: limited to answer streaming finalization; no media/error final path or main bot runtime changes.
- What: split glued progress/final headings such as now.Comparison into retained progress plus final answer text. - Why: live tester showed two bubbles but the progress bubble still contained a smashed final heading. - Risk: limited to Telegram answer progress/final boundary detection; no main runtime or provider config changes.
- what changed: flush the restored retained progress preview before materializing it and cover the ordering in the glued Comparison regression. - why: Telegram must freeze the natural progress bubble before sending the separate final answer bubble. - behavior/risk: low; scoped to finalizing an existing streamed progress preview before final delivery.
- what changed: require streamed answer previews to look like real progress before retaining and splitting them into a separate Telegram bubble. - why: plain final-answer streaming can start with tiny prefixes like **example, and those must stay part of the final answer instead of becoming a fake progress bubble. - behavior/risk: low; scoped to Telegram finalization and covered by the live-shaped regression plus the existing focused suite.
- what changed: suppress short title-only answer prefixes like example.com: from creating retained Telegram progress bubbles, while keeping natural progress phrases eligible. - why: streamed final-answer headings are not useful progress and were leaving useless extra bubbles above the final answer. - behavior/risk: low; scoped to Telegram answer preview retention and covered by the live-shaped regression plus focused suite.
- what changed: suppress short heading-shaped answer prefixes like example.com: Documentation from creating retained Telegram progress bubbles, while keeping natural progress phrases eligible. - why: streamed final-answer headings are not useful progress and were leaving useless extra bubbles above the final answer. - behavior/risk: low; scoped to Telegram answer preview retention and covered by the live-shaped regression plus focused suite.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Review Fast Path
1 file / 81 tests), expanded targeted suite passed (3 files / 131 tests),git diff --checkpassed, andpnpm buildcompleted with the known plugin SDK DTS behavior undertsc ... || true.Why This Matters
Step 1 of N, checkmarks, and per-step separators were model/test-prompt pollution, not a UI convention OpenClaw should create or require.I'll check the first page nowandNow checking the IANA page, proving OpenClaw can retain normal model progress without fake numbered status text.Scope Boundary
Step 1 of N, checkmarks, or per-step separators.Verification
pnpm exec vitest run extensions/telegram/src/bot-message-dispatch.test.ts --pool=forks --maxWorkers=1passed:1 file / 81 tests.pnpm exec vitest run extensions/telegram/src/bot-message-dispatch.test.ts extensions/telegram/src/lane-delivery.test.ts src/infra/telegram-live-runtime-helpers.test.ts --pool=forks --maxWorkers=1passed:3 files / 131 tests.git diff --checkpassed.pnpm buildcompleted; known DTS errors still print undertsc -p tsconfig.plugin-sdk.dts.json || true.Step 1 of N, checkmarks, or separators to model-authored progress narration. Existing examples were model/test-prompt pollution, except separate tool-progress system lines can still use the existing🔧 toolformat.AI Assistance