Skip to content

perf(flowchat): Keep long thinking lightweight - #3204

Merged
wsp1911 merged 1 commit into
GCWing:mainfrom
wsp1911:dev-omen
Sep 22, 2026
Merged

wsp1911 merged 1 commit into
GCWing:mainfrom
wsp1911:dev-omen

Conversation

@wsp1911

@wsp1911 wsp1911 commented Sep 22, 2026 •

Copy link
Copy Markdown
Collaborator

Use Markstream for thinking Markdown while preserving the response
renderer and existing safe HTML, math, image, and navigation behavior.
Reuse unchanged parsed trees so settled thinking blocks skip renders.

Remove thinking arrival fading because its full-document DOM scans
cause frame drops on long streams. Keep thinking code fences as plain
source permanently, including Mermaid and completed/reopened content,
to avoid bulk highlighting stalls. Preserve line numbers and copying.
Document both restrictions to prevent reintroducing these costs.

Use Markstream for thinking Markdown while preserving the response
renderer and existing safe HTML, math, image, and navigation behavior.
Reuse unchanged parsed trees so settled thinking blocks skip renders.

Remove thinking arrival fading because its full-document DOM scans
cause frame drops on long streams. Keep thinking code fences as plain
source permanently, including Mermaid and completed/reopened content,
to avoid bulk highlighting stalls. Preserve line numbers and copying.
Document both restrictions to prevent reintroducing these costs.

Remove temporary performance probes after runtime verification.

Validation: 105 focused tests, check:web, ESLint, and diff checks pass.
@wsp1911
wsp1911 merged commit 5e5c4b4 into GCWing:main Sep 22, 2026
13 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant