fix(cli): preserve table ANSI color across wrapped lines by chiga0 · Pull Request #4050 · QwenLM/qwen-code

chiga0 · 2026-05-11T08:30:36Z

Summary

Related issue: Fixes Table inline-code highlight is lost after wrapping in narrow markdown tables #4052.
What changed: Preserves ANSI foreground color when markdown table cell content wraps across lines in the terminal UI.
Why it changed: Long inline-code table names could lose their code highlight on continuation lines after a narrow-width table wrap.
Reviewer focus: Please focus on ANSI state handling at table wrap boundaries and the deterministic TUI regression evidence.

Validation

Commands run:

cd packages/cli && npx vitest run src/ui/utils/TableRenderer.test.tsx

npm run lint

npm run build && npm run typecheck && npm run bundle

QWEN_TUI_E2E_OUT=/tmp/qwen-table-wrap-ansi/fixed \
  npx tsx integration-tests/terminal-capture/table-inline-code-wrap-regression.ts

QWEN_TUI_E2E_REPO=/Users/gawain/Documents/codebase/opensource/qwen-code-table-wrap-ansi-highlight-base \
QWEN_TUI_E2E_OUT=/tmp/qwen-table-wrap-ansi/base \
QWEN_TUI_E2E_EXPECT_PASS=false \
  npx tsx integration-tests/terminal-capture/table-inline-code-wrap-regression.ts

Prompts / inputs used: The E2E test uses a local fake OpenAI-compatible server that returns a fixed markdown table containing a long inline-code table name.
Expected result: In a narrow terminal, the table name actually wraps and every wrapped continuation keeps the active code foreground color.
Observed result: Base reproduced the missing color on the continuation; this branch keeps the continuation colored.
Quickest reviewer verification path: Run the table renderer unit test, then run the terminal-capture regression script once on this branch.
Evidence (output, logs, screenshots, video, JSON, before/after, etc.):
- Unit: TableRenderer.test.tsx passed with 47 tests.
- Uploaded E2E screenshots and summaries are in this PR comment: fix(cli): preserve table ANSI color across wrapped lines #4050 (comment)
- Artifact commit: chiga0/qwen-code@e8a0a725053314bf5e3399368c1efc134b99ae15 on fix/table-wrap-ansi-highlight-artifacts.
- Base summary: https://raw.githubusercontent.com/chiga0/qwen-code/e8a0a725053314bf5e3399368c1efc134b99ae15/table-wrap-ansi-highlight/base/summary.json reported finalScreenWrappedTableName: true, coloredContinuationOccurrences: 0, uncoloredContinuationOccurrences: 1, pass: false.
- Fixed summary: https://raw.githubusercontent.com/chiga0/qwen-code/e8a0a725053314bf5e3399368c1efc134b99ae15/table-wrap-ansi-highlight/fixed/summary.json reported finalScreenWrappedTableName: true, coloredContinuationOccurrences: 1, uncoloredContinuationOccurrences: 0, pass: true.

Scope / Risk

Main risk or tradeoff: The change adds foreground SGR carry logic around table cell wrapping. It intentionally only tracks foreground color, not full terminal styling state.
Not covered / not validated: Non-table inline code and fenced code blocks are not affected by this path; they use Ink text color rendering rather than table ANSI-string wrapping.
Breaking changes / migration notes: None.

Testing Matrix

	🍏	🪟	🐧
npm run	✅	⚠️	⚠️
npx	⚠️	⚠️	⚠️
Docker	⚠️	N/A	N/A
Podman	⚠️	N/A	N/A
Seatbelt	⚠️	N/A	N/A

Testing matrix notes:

Validated locally on macOS with npm scripts and the terminal-capture E2E harness. Other platforms were not run locally.

Linked Issues / Bugs

Fixes #4052

chiga0 · 2026-05-11T08:42:08Z

E2E report for table inline-code wrap ANSI regression:

Harness: integration-tests/terminal-capture/table-inline-code-wrap-regression.ts
Scenario: local fake OpenAI-compatible server returns a fixed markdown table with a long inline-code table name; real dist/cli.js renders it in a 100x32 terminal.
Artifact commit: chiga0/qwen-code@e8a0a725053314bf5e3399368c1efc134b99ae15 on fix/table-wrap-ansi-highlight-artifacts.

Base worktree (origin/main) reproduced the bug: finalScreenWrappedTableName: true, coloredContinuationOccurrences: 0, uncoloredContinuationOccurrences: 1, pass: false with QWEN_TUI_E2E_EXPECT_PASS=false.

Fixed branch passed: finalScreenWrappedTableName: true, coloredContinuationOccurrences: 1, uncoloredContinuationOccurrences: 0, pass: true.

wenshao · 2026-05-11T16:04:15Z

-  const lines = wrapped.split('\n');
+  const lines = preserveForegroundAcrossLineBreaks(wrapped).split('\n');
  // Trim trailing empty lines (wrap-ansi artifacts) but preserve internal ones
  while (lines.length > 1 && lines[lines.length - 1]!.length === 0) {


[Suggestion] The trailing-empty-line trim here can be defeated by preserveForegroundAcrossLineBreaks.

When a foreground is active and wrapAnsi produces a trailing empty line, preserveForegroundAcrossLineBreaks injects \x1b[39m + the active foreground escape around the trailing \n. The resulting line contains only ANSI codes (e.g. \x1b[38;2;R;G;Bm\x1b[39m), so its .length is non-zero and the while-loop doesn't trim it.

This inflates the line count returned by wrapText, which feeds into calculateMaxRowLines. An inflated count can cause tables to switch from horizontal to vertical format unnecessarily.

stripAnsi is already imported, so the fix is minimal:

Suggested change

while (lines.length > 1 && lines[lines.length - 1]!.length === 0) {

while (lines.length > 1 && stripAnsi(lines[lines.length - 1]!).length === 0) {

— glm-5.1 via Qwen Code /review

Applied as suggested in b5f93c2. Verification trace: when an inline-code span is active at the end of a row, wrapAnsi can emit a trailing \n while the foreground is still open. Walking through preserveForegroundAcrossLineBreaks: the final segment after the last SGR contains text\n, and because activeForeground is set, the \n is rewritten to \x1b[39m\n\${activeForeground}. After split('\n') the trailing element is \${activeForeground} (e.g. \x1b[38;2;R;G;Bm) — .length is 7+, so the old length === 0 check left it in place and calculateMaxRowLines saw an extra line. Switched to stripAnsi(lines[lines.length - 1]!).length === 0; all 47 TableRenderer.test.tsx cases still pass.

wenshao · 2026-05-11T16:04:16Z

+QWEN_TUI_E2E_OUT=/tmp/qwen-table-wrap-ansi/fixed \
+  npx tsx integration-tests/terminal-capture/table-inline-code-wrap-regression.ts
+
+QWEN_TUI_E2E_REPO=/Users/gawain/Documents/codebase/opensource/qwen-code-table-wrap-ansi-highlight-base \


[Suggestion] The QWEN_TUI_E2E_REPO path is hardcoded to a developer-local directory. Other contributors cannot reproduce these commands as-is.

Consider replacing with a relative path or a placeholder like <path-to-base-worktree> so the E2E test plan is reproducible by anyone.

— glm-5.1 via Qwen Code /review

Fixed in b5f93c2. The Commands block now starts with FIX_REPO=<path-to-fix-worktree> / BASE_REPO=<path-to-base-worktree> placeholders and the QWEN_TUI_E2E_REPO, cd, and inner cd packages/cli invocations all consume those vars, so anyone with two checkouts (one on this branch, one pinned at the merge-base) can run the plan verbatim.

tanzhenxin

Review

Well-scoped fix at the right seam. The underlying gap is actually upstream in wrap-ansi — its SGR-tracking regex only matches single-parameter codes, so truecolor and 256-color forms aren't re-emitted after a line break. Patching this in the table renderer is correct because Ink doesn't help here (table cells render as flat strings). The post-pass is idempotent with wrap-ansi's native single-parameter carry and composes cleanly with the OSC 8 envelopes in PR #4037. Foreground-only scope is the right call — wrap-ansi already carries bold/italic/underline natively.

Verdict

APPROVE — solid fix with strong evidence.

The initial revert downgraded wrap-ansi to 9.0.2 (the pre-PR-#3860 state). After rebasing onto current main, PR #4050 (preserve table ANSI color across wrapped lines) brought in a new test ("does not preserve foreground after an explicit foreground reset") whose wrap point depends on ink 7's <Text> wrapping behavior. Two-part fix: 1. Restore wrap-ansi to 10 (cli direct dep). The wrap-ansi version is independent of the ink regression we're reverting — wrap-ansi 10 has no peer-dep tie to ink 7 — and #4050's TableRenderer code on main already assumes wrap-ansi 10. Keeping the wrap-ansi bump removes the root override for wrap-ansi (was forcing all transitives to 9.0.2) so cli's TableRenderer gets the wrap-ansi 10 it expects, while ink 6's transitive wrap-ansi naturally resolves to 9 (its own declared range) — no conflict. 2. Skip the one new test that asserts a specific wrap position. The other assertions in that test (foreground cleared, equal visible widths) still pass on ink 6 — only `expectWrappedContinuation` is ink-7-specific. The sibling test 'does not preserve foreground after an explicit reset' (using \\u001b[0m instead of \\u001b[39m) still passes unmodified on ink 6, so the ANSI-handling logic itself is verified end-to-end. The TODO marker references the re-upgrade path. Local verification: - TableRenderer.test.tsx: 54/54 pass + 1 skipped - AppContainer.test.tsx: 61/61 pass - MainContent.test.tsx: 6/6 pass - clearCommand.test.ts: 13/13 pass - npm run typecheck --workspace=packages/cli: clean - npm ls ink → single ink@6.8.0 - npm ls wrap-ansi → cli direct: 10.0.0; ink 6 transitive: 9.0.2 (no conflict, no override) Generated with AI Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>

…n from #3860 (#4083) * revert(deps): downgrade ink 7.0.2 → 6.x to fix Static-remount regression from #3860 PR #3860 upgraded ink 6.2.3 → 7.0.2 with the claim of "no business code changes." In production this turns out to break the TUI: - After `/clear`, the next user message and AI response do not render to the static history area — only the dynamic spinner/input area is visible (#3860 + chore/upgrade-ink-7 branch reproduce this). - After Ctrl+O (TOGGLE_COMPACT_MODE), the screen is cleared and stays blank. - Any `refreshStatic()` call path (auth refresh, model change, render- mode switch, /clear, Ctrl+O) puts the UI into the same "muted" state. Root cause is an ink 7 regression: when `<Static>` is remounted by changing its `key` prop, the new instance's items are never written to stdout. A 30-line minimal repro (pure ink + Static + key++) confirms this independently of qwen-code. Closest upstream issue: vadimdemedes/ink#773 (useLayoutEffect-driven child stripping in <Static>). PR #905 ("Fix dangling staticNode reference") merged into ink 7 fixed the unmount-OOM path but not this remount path. No upstream issue yet matches the "remount loses content" case — we should file one and ship a re-upgrade once it is resolved. Scope of this revert (intentional partial revert of #3860): - ink ^7.0.2 → ^6.2.3 (cli + root hoist) - react / react-dom 19.2.4 pin → ^19.1.0 (cli direct, root overrides removed) - wrap-ansi ^10.0.0 → 9.0.2 (cli direct, root override restored) - react-devtools-core kept at ^6.1.5 (still ink-6 compatible — ink 6.8.0's peerOptional requires >=6.1.2; downgrading to 4.x would re-introduce a conflict) - @vitest/eslint-plugin pin "1.3.4" → "^1.3.4" - "@types/node" override removed (was only needed for ink 7's Node 22 type drift) What this revert keeps: - Node engines >=22 across root / cli / core / sdk / web-templates and the matching Dockerfile / .nvmrc / CI matrix work. PR #1876 followed up by adding Node 24 support to the matrix, and rolling those back would conflict with that work. The visible bug is the ink runtime regression, not the engine bump. - doctorChecks.ts MIN_NODE_MAJOR = 22 (matches engines). - The test gating that #3860 added for ink-7 input throttle (AuthDialog / AskUserQuestionDialog / InputPrompt). With ink 6 these tests would pass un-gated, but leaving the gate in place is harmless and a follow-up can un-gate them. Keeping this revert minimal. Verification (local, ink 6.8.0 single instance): - npm ls ink → single ink@6.8.0 - npm ls react → single react@19.2.4 (kept by vscode-ide-companion workspace pin; ink 6 is fine on 19.2) - npm run typecheck --workspace=packages/cli → clean - AppContainer.test.tsx 61/61 pass - MainContent.test.tsx 6/6 pass - clearCommand.test.ts 13/13 pass Re-upgrade path: once ink ships a fix for the Static-remount regression, redo this upgrade behind the feat/virtual-viewport-on-ink7 branch where the `<Static>` + clearTerminal combo is replaced by an overflowY=hidden self-managed viewport. Generated with AI Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com> * ci(fix): keep wrap-ansi 10 + skip 1 ink-7-specific TableRenderer test The initial revert downgraded wrap-ansi to 9.0.2 (the pre-PR-#3860 state). After rebasing onto current main, PR #4050 (preserve table ANSI color across wrapped lines) brought in a new test ("does not preserve foreground after an explicit foreground reset") whose wrap point depends on ink 7's <Text> wrapping behavior. Two-part fix: 1. Restore wrap-ansi to 10 (cli direct dep). The wrap-ansi version is independent of the ink regression we're reverting — wrap-ansi 10 has no peer-dep tie to ink 7 — and #4050's TableRenderer code on main already assumes wrap-ansi 10. Keeping the wrap-ansi bump removes the root override for wrap-ansi (was forcing all transitives to 9.0.2) so cli's TableRenderer gets the wrap-ansi 10 it expects, while ink 6's transitive wrap-ansi naturally resolves to 9 (its own declared range) — no conflict. 2. Skip the one new test that asserts a specific wrap position. The other assertions in that test (foreground cleared, equal visible widths) still pass on ink 6 — only `expectWrappedContinuation` is ink-7-specific. The sibling test 'does not preserve foreground after an explicit reset' (using \\u001b[0m instead of \\u001b[39m) still passes unmodified on ink 6, so the ANSI-handling logic itself is verified end-to-end. The TODO marker references the re-upgrade path. Local verification: - TableRenderer.test.tsx: 54/54 pass + 1 skipped - AppContainer.test.tsx: 61/61 pass - MainContent.test.tsx: 6/6 pass - clearCommand.test.ts: 13/13 pass - npm run typecheck --workspace=packages/cli: clean - npm ls ink → single ink@6.8.0 - npm ls wrap-ansi → cli direct: 10.0.0; ink 6 transitive: 9.0.2 (no conflict, no override) Generated with AI Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com> --------- Co-authored-by: 秦奇 <gary.gq@alibaba-inc.com> Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>

wenshao · 2026-05-12T10:19:44Z

 }

+function updateActiveForeground(
+  activeForeground: string,


[Suggestion] updateActiveForeground only tracks SGR foreground color codes (38, 30-37, 90-97). Format attribute codes produced by renderMarkdownToAnsi — bold (1/22), italic (3/23), underline (4/24), strikethrough (9/29) — are silently ignored. When wrap-ansi breaks a line inside a formatted span (e.g., \x1b[1mBold text that wraps\x1b[22m), updateActiveForeground returns '', so preserveForegroundAcrossLineBreaks does not re-apply bold on the continuation line.

Impact: AI responses with bold, italic, underline, or strikethrough markdown in narrow table cells will lose formatting on wrapped continuation lines. The test suite never exercises this path (all **bold** tests use cell widths that don't trigger wrapping).

Suggested change

activeForeground: string,

// Extend updateActiveForeground to also track format attributes (codes 1-9, 22-29),

// or document its foreground-only scope. At minimum, add a test with narrow-width

// bold content verifying bold persists on continuation lines.

— DeepSeek/deepseek-v4-pro via Qwen Code /review

wenshao · 2026-05-12T10:19:44Z

+  const foregroundAtText = (
+    output: string,
+    text: string,
+  ): string | undefined => {


[Suggestion] The new wrap-preservation tests only cover truecolor (38;2) and 256-color (38;5) foreground preservation. Standard ANSI color codes 30-37 (red, green, yellow, blue, magenta, cyan, white) and bright foreground codes 90-97 are not tested for wrap preservation, despite being the most commonly used ANSI colors.

A regression in the standard-color branch of updateActiveForeground would not be caught by any test.

Suggested change

): string | undefined => {

// Add wrap-preservation tests for standard and bright foreground codes:

// - `\x1b[31mABCDEFGHIJKLMNOPQRSTUVWXYZ\x1b[39m` with narrow width

// - `\x1b[91mABCDEFGHIJKLMNOPQRSTUVWXYZ\x1b[39m` with narrow width

// Assert foregroundAtText on continuation lines returns '31' / '91'.

— DeepSeek/deepseek-v4-pro via Qwen Code /review

wenshao · 2026-05-12T10:19:44Z

@@ -245,7 +338,7 @@ function wrapText(
    trim: false,


[Suggestion] renderRowLines composes three ANSI layers: (1) preserveForegroundAcrossLineBreaks inside wrapText ensures inline colors survive wrapping, (2) recolorAfterResets re-applies base cell color after internal resets, (3) applyColor wraps the entire cell in theme color. The interaction between these — where recolorAfterResets depends on \x1b[39m inserted by preserveForegroundAcrossLineBreaks — is undocumented. A future maintainer who reorders or merges these calls would break ANSI layering with no clear signal why.

Suggested change

trim: false,

// Add a comment explaining the three-layer ANSI composition and why order matters:

// 1. preserveForegroundAcrossLineBreaks → preserves inline colors across \n

// 2. recolorAfterResets → restores base color after internal \x1b[39m

// 3. applyColor → wraps entire line in theme text color

— DeepSeek/deepseek-v4-pro via Qwen Code /review

… [upstream cherry-pick] Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Co-authored-by: 秦奇 <gary.gq@alibaba-inc.com>

…n from #3860 (#4083) * revert(deps): downgrade ink 7.0.2 → 6.x to fix Static-remount regression from #3860 PR #3860 upgraded ink 6.2.3 → 7.0.2 with the claim of "no business code changes." In production this turns out to break the TUI: - After `/clear`, the next user message and AI response do not render to the static history area — only the dynamic spinner/input area is visible (#3860 + chore/upgrade-ink-7 branch reproduce this). - After Ctrl+O (TOGGLE_COMPACT_MODE), the screen is cleared and stays blank. - Any `refreshStatic()` call path (auth refresh, model change, render- mode switch, /clear, Ctrl+O) puts the UI into the same "muted" state. Root cause is an ink 7 regression: when `<Static>` is remounted by changing its `key` prop, the new instance's items are never written to stdout. A 30-line minimal repro (pure ink + Static + key++) confirms this independently of qwen-code. Closest upstream issue: vadimdemedes/ink#773 (useLayoutEffect-driven child stripping in <Static>). PR #905 ("Fix dangling staticNode reference") merged into ink 7 fixed the unmount-OOM path but not this remount path. No upstream issue yet matches the "remount loses content" case — we should file one and ship a re-upgrade once it is resolved. Scope of this revert (intentional partial revert of #3860): - ink ^7.0.2 → ^6.2.3 (cli + root hoist) - react / react-dom 19.2.4 pin → ^19.1.0 (cli direct, root overrides removed) - wrap-ansi ^10.0.0 → 9.0.2 (cli direct, root override restored) - react-devtools-core kept at ^6.1.5 (still ink-6 compatible — ink 6.8.0's peerOptional requires >=6.1.2; downgrading to 4.x would re-introduce a conflict) - @vitest/eslint-plugin pin "1.3.4" → "^1.3.4" - "@types/node" override removed (was only needed for ink 7's Node 22 type drift) What this revert keeps: - Node engines >=22 across root / cli / core / sdk / web-templates and the matching Dockerfile / .nvmrc / CI matrix work. PR #1876 followed up by adding Node 24 support to the matrix, and rolling those back would conflict with that work. The visible bug is the ink runtime regression, not the engine bump. - doctorChecks.ts MIN_NODE_MAJOR = 22 (matches engines). - The test gating that #3860 added for ink-7 input throttle (AuthDialog / AskUserQuestionDialog / InputPrompt). With ink 6 these tests would pass un-gated, but leaving the gate in place is harmless and a follow-up can un-gate them. Keeping this revert minimal. Verification (local, ink 6.8.0 single instance): - npm ls ink → single ink@6.8.0 - npm ls react → single react@19.2.4 (kept by vscode-ide-companion workspace pin; ink 6 is fine on 19.2) - npm run typecheck --workspace=packages/cli → clean - AppContainer.test.tsx 61/61 pass - MainContent.test.tsx 6/6 pass - clearCommand.test.ts 13/13 pass Re-upgrade path: once ink ships a fix for the Static-remount regression, redo this upgrade behind the feat/virtual-viewport-on-ink7 branch where the `<Static>` + clearTerminal combo is replaced by an overflowY=hidden self-managed viewport. Generated with AI Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com> * ci(fix): keep wrap-ansi 10 + skip 1 ink-7-specific TableRenderer test The initial revert downgraded wrap-ansi to 9.0.2 (the pre-PR-#3860 state). After rebasing onto current main, PR #4050 (preserve table ANSI color across wrapped lines) brought in a new test ("does not preserve foreground after an explicit foreground reset") whose wrap point depends on ink 7's <Text> wrapping behavior. Two-part fix: 1. Restore wrap-ansi to 10 (cli direct dep). The wrap-ansi version is independent of the ink regression we're reverting — wrap-ansi 10 has no peer-dep tie to ink 7 — and #4050's TableRenderer code on main already assumes wrap-ansi 10. Keeping the wrap-ansi bump removes the root override for wrap-ansi (was forcing all transitives to 9.0.2) so cli's TableRenderer gets the wrap-ansi 10 it expects, while ink 6's transitive wrap-ansi naturally resolves to 9 (its own declared range) — no conflict. 2. Skip the one new test that asserts a specific wrap position. The other assertions in that test (foreground cleared, equal visible widths) still pass on ink 6 — only `expectWrappedContinuation` is ink-7-specific. The sibling test 'does not preserve foreground after an explicit reset' (using \\u001b[0m instead of \\u001b[39m) still passes unmodified on ink 6, so the ANSI-handling logic itself is verified end-to-end. The TODO marker references the re-upgrade path. Local verification: - TableRenderer.test.tsx: 54/54 pass + 1 skipped - AppContainer.test.tsx: 61/61 pass - MainContent.test.tsx: 6/6 pass - clearCommand.test.ts: 13/13 pass - npm run typecheck --workspace=packages/cli: clean - npm ls ink → single ink@6.8.0 - npm ls wrap-ansi → cli direct: 10.0.0; ink 6 transitive: 9.0.2 (no conflict, no override) Generated with AI Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com> --------- Co-authored-by: 秦奇 <gary.gq@alibaba-inc.com> Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>

fix(cli): preserve table ANSI color across wrapped lines

1c31a13

chiga0 force-pushed the fix/table-wrap-ansi-highlight branch from 67031df to 1c31a13 Compare May 11, 2026 08:41

chiga0 mentioned this pull request May 11, 2026

Table inline-code highlight is lost after wrapping in narrow markdown tables #4052

Closed

chiga0 requested review from tanzhenxin, wenshao and yiliang114 May 11, 2026 09:06

wenshao reviewed May 11, 2026

View reviewed changes

chiga0 requested a review from wenshao May 12, 2026 07:10

tanzhenxin added the type/bug Something isn't working as expected label May 12, 2026

tanzhenxin approved these changes May 12, 2026

View reviewed changes

tanzhenxin merged commit dc7a90c into QwenLM:main May 12, 2026
8 checks passed

wenshao reviewed May 12, 2026

View reviewed changes

TaimoorSiddiquiOfficial mentioned this pull request May 13, 2026

Merge/upstream safe commits TaimoorSiddiquiOfficial/HopCode#24

Merged

wenshao pushed a commit that referenced this pull request May 17, 2026

fix(cli): preserve table ANSI color across wrapped lines (#4050)

42ae7ce

Co-authored-by: 秦奇 <gary.gq@alibaba-inc.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(cli): preserve table ANSI color across wrapped lines#4050

fix(cli): preserve table ANSI color across wrapped lines#4050
tanzhenxin merged 1 commit into
QwenLM:mainfrom
chiga0:fix/table-wrap-ansi-highlight

chiga0 commented May 11, 2026 •

edited

Loading

Uh oh!

chiga0 commented May 11, 2026 •

edited

Loading

Uh oh!

wenshao May 11, 2026

Uh oh!

chiga0 May 12, 2026

Uh oh!

wenshao May 11, 2026

Uh oh!

chiga0 May 12, 2026

Uh oh!

tanzhenxin left a comment

Uh oh!

Uh oh!

wenshao May 12, 2026

Uh oh!

wenshao May 12, 2026

Uh oh!

wenshao May 12, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

	while (lines.length > 1 && lines[lines.length - 1]!.length === 0) {
	while (lines.length > 1 && stripAnsi(lines[lines.length - 1]!).length === 0) {

-  activeForeground: string,
+// Extend updateActiveForeground to also track format attributes (codes 1-9, 22-29),
+// or document its foreground-only scope. At minimum, add a test with narrow-width
+// bold content verifying bold persists on continuation lines.

-  ): string | undefined => {
+// Add wrap-preservation tests for standard and bright foreground codes:
+// - `\x1b[31mABCDEFGHIJKLMNOPQRSTUVWXYZ\x1b[39m` with narrow width
+// - `\x1b[91mABCDEFGHIJKLMNOPQRSTUVWXYZ\x1b[39m` with narrow width
+// Assert foregroundAtText on continuation lines returns '31' / '91'.

-    trim: false,
+// Add a comment explaining the three-layer ANSI composition and why order matters:
+// 1. preserveForegroundAcrossLineBreaks → preserves inline colors across \n
+// 2. recolorAfterResets → restores base color after internal \x1b[39m
+// 3. applyColor → wraps entire line in theme text color

Conversation

chiga0 commented May 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Validation

Scope / Risk

Testing Matrix

Linked Issues / Bugs

Uh oh!

chiga0 commented May 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

wenshao May 11, 2026

Choose a reason for hiding this comment

Uh oh!

chiga0 May 12, 2026

Choose a reason for hiding this comment

Uh oh!

wenshao May 11, 2026

Choose a reason for hiding this comment

Uh oh!

chiga0 May 12, 2026

Choose a reason for hiding this comment

Uh oh!

tanzhenxin left a comment

Choose a reason for hiding this comment

Review

Verdict

Uh oh!

Uh oh!

wenshao May 12, 2026

Choose a reason for hiding this comment

Uh oh!

wenshao May 12, 2026

Choose a reason for hiding this comment

Uh oh!

wenshao May 12, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

chiga0 commented May 11, 2026 •

edited

Loading

chiga0 commented May 11, 2026 •

edited

Loading