Skip to content

css: eliminate all unreachable!() via type-level hardening - #33332

Draft
alii wants to merge 16 commits into
mainfrom
claude/css-remove-unreachable
Draft

alii wants to merge 16 commits into
mainfrom
claude/css-remove-unreachable

Conversation

@alii

@alii alii commented Jul 4, 2026 •

Copy link
Copy Markdown
Member

What does this PR do?

Removes all 67 unreachable!() sites from src/css/ by making the invalid states compile errors instead of runtime panics.

Each site was traced to its invariant and cross-referenced against Lightning CSS. The fixes fall into a handful of patterns:

Type restructuring — the invalid state is no longer representable:

  • ColorFallbackKind bitflags was used both as a set and as a single discriminant → new ColorFallback enum with three variants; get_fallback matches exhaustively.
  • CssColor::to_light_dark() returned a CssColor that was always the LightDark variant, forcing every caller to let-else unreachable!() → now returns a (Box<CssColor>, Box<CssColor>) tuple.
  • parse_qualified_name(in_attr_selector: bool) returned a 6-variant QNamePrefix where each of its two callers panicked on the other's variants → split into two return types.
  • margin_padding extract_* fns panicked on wrong variant → return Option<&T>; callers fuse the tag guard.
  • QueryCondition::extra_to_css had a panicking default → made a required trait method.
  • Features::from_compat returns Option<Features> (was Features with an empty() fallthrough); should_compile_same handles None by falling back to !is_compatible() — semantically correct instead of the always-true contains(empty()).

Stale borrowck workarounds — matches!() guard followed by re-matching the same value with _ => unreachable!() → collapsed to direct by-value match / or-patterns (calc.rs, selector.rs, custom.rs, css_parser.rs).

Dead code deleted — variants and helpers with zero construction sites: AnimationTimeline::Scroll/View and their four payload types (ScrollTimeline, ViewTimeline, Scroller, ScrollAxis), AtRulePrelude::FontFeatureValues, Token::raw(), VendorPrefix::from_name_str, four #[deprecated] to_css poison-pill shims.

Internal contracts — cssparser parse_until_before delimiter guarantees and similar invariants that can't be encoded in types → debug_assert! with a safe fallback (return error / emit nothing) instead of a release-build panic.

How did you verify your code works?

  • rg 'unreachable!' src/css/ → 0 (was 67 including one added by css: scope animation-name to its @keyframes hash in CSS modules (#18921) #33322 while this branch was open).
  • cargo clippy -p bun_css --no-deps clean.
  • bun bd test test/js/bun/css/css.test.ts test/bundler/css/ test/bundler/esbuild/css.test.ts — all pass.
  • Added QNamePrefix selector-matrix tests covering [svg|attr], |foo, |*, svg|*, and the [*] error path.

@robobun

robobun commented Jul 4, 2026 •

Copy link
Copy Markdown
Collaborator
Updated 9:52 AM PT - Jul 6th, 2026

❌ @alii, your commit 46e30e8 has some failures in Build #68646 (All Failures)


🧪   To try this PR locally:

bunx bun-pr 33332

That installs a local version of the PR into your bun-33332 executable, so you can run:

bun-33332 --bun

@coderabbitai

coderabbitai Bot commented Jul 4, 2026 •

Copy link
Copy Markdown
Contributor

Review Change Stack

Walkthrough

This PR replaces several CSS parser and serialization panic paths with debug assertions or fallback behavior, changes color fallback APIs from bitflags to enum values, refactors margin/padding extraction to return options, removes unused animation timeline variants, reshapes selector parsing/serialization, and updates syntax repetition handling and tests.

Changes

Panic hardening and API refactors

Layer / File(s) Summary
Parser and media-query contract hardening
src/css/css_parser.rs, src/css/lib.rs, src/css/media_query.rs
At-rule parsing, comma-list exhaustion, nested-rule composition state, and media-query serialization replace unreachable branches with debug assertions or explicit errors; AtRulePrelude::FontFeatureValues, Token::raw, and VendorPrefix::from_name_str are removed; QueryCondition::extra_to_css becomes required.
AnimationTimeline Scroll/View variant removal
src/css/properties/animation.rs
AnimationTimeline::{Scroll, View} and the ScrollTimeline/ViewTimeline structs are removed, and to_css/deep_clone/PartialEq are reduced to the remaining variants.
ColorFallback enum and helper changes
src/css/small_list.rs, src/css/values/color.rs, src/css/values/image.rs, src/css/values/gradient.rs
Shared fallback traits and value helpers switch from ColorFallbackKind to ColorFallback; CssColor::to_light_dark, CssColor::interpolate, and fallback generation in images and gradients are updated accordingly.
Color fallback callsites in properties
src/css/properties/background.rs, src/css/properties/border.rs, src/css/properties/custom.rs
Background, border, and custom-property fallback callsites are updated to pass concrete ColorFallback values and propagate the new fallback type through token lists, variables, environment variables, and functions.
Option-based margin/padding extraction
src/css/properties/margin_padding.rs
SizeHandlerSpec extractors and the generated projection macro return Option<&...> instead of panicking, and handle_property/flush/prop logic is rewritten around the new contract.
Rule and property serialization fixes
src/css/properties/properties_generated.rs, src/css/properties/size.rs, src/css/rules/container.rs, src/css/rules/mod.rs
Property::Custom serialization and stretch serialization for Size and MaxSize no longer hit unreachable paths; StyleQuery and ContainerCondition define extra_to_css; minify_style_arm operates on StyleRule directly and rebuilds emptied rules explicitly.
Selector qualified-name and serialization refactor
src/css/selectors/parser.rs, src/css/selectors/selector.rs
Type-selector and attribute-selector qualified-name parsing are split apart, deprecated selector to_css methods are removed, and selector serialization/specificity branches are rewritten with direct arms and debug assertions.
Feature mapping, calc arithmetic, syntax separators, and tests
src/css/targets.rs, src/css/values/calc.rs, src/css/values/length.rs, src/css/values/percentage.rs, src/css/values/easing.rs, src/css/values/syntax.rs, test/js/bun/css/css.test.ts
Feature mapping becomes exhaustive, calc and length addition logic is rewritten with direct matches, easing keywords serialize explicitly, repeated syntax uses a separator enum, and CSS tests cover selector QName splitting and attribute selector errors.

Possibly related PRs

  • oven-sh/bun#31920: Both PRs modify src/css/rules/mod.rs’s CssRuleList::minify and minify_style_arm logic around style-rule merging and minification behavior.
🚥 Pre-merge checks | ✅ 4
✅ Passed checks (4 passed)
Check name Status Explanation
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
Title check ✅ Passed The title clearly summarizes the main change: removing unreachable!() by hardening CSS types and invariants.
Description check ✅ Passed The description follows the required template and includes both the change summary and verification steps.

Comment @coderabbitai help to get the list of available commands.

Comment thread src/css/values/gradient.rs Resolved
Comment thread src/css/values/syntax.rs Resolved
@alii
alii force-pushed the claude/css-remove-unreachable branch from 2bb0c82 to f557d51 Compare July 5, 2026 20:34

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 3

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (2)
src/css/properties/animation.rs (1)

154-267: 📐 Maintainability & Code Quality | 🟠 Major | ⚡ Quick win

Extract the repeated name-disambiguation check.

direction, fill_mode, and play_state each repeat the same css::parse_utility::parse_string::<T>(dest.arena, n, T::parse).is_ok() pattern against name_str. Per this repo's own review rule, a pattern repeated a third time in the same diff should be pulled into a shared helper.

♻️ Proposed helper
+fn name_disambiguates<T: css::EnumProperty>(dest: &Printer, name_str: Option<&[u8]>) -> bool {
+    name_str.is_some_and(|n| {
+        css::parse_utility::parse_string::<T>(dest.arena, n, T::parse).is_ok()
+    })
+}

Then call name_disambiguates::<AnimationDirection>(dest, name_str), etc., at each site (keeping the extra !eql_case_insensitive_ascii(n, b"none", true) guard specific to fill_mode).

As per coding guidelines, "The second time a multi-line block appears in your diff, extract a named helper and use it at EVERY parallel site."

🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/css/properties/animation.rs` around lines 154 - 267, The
name-disambiguation logic in AnimationName::to_css is duplicated across
direction, fill_mode, and play_state, so extract it into a shared helper and
reuse it at each site. Add a small helper like name_disambiguates::<T>(dest,
name_str) that wraps the css::parse_utility::parse_string::<T>(dest.arena, n,
T::parse).is_ok() check, then call it for AnimationDirection, AnimationFillMode,
and AnimationPlayState. Keep the fill_mode-specific
!strings::eql_case_insensitive_ascii(n, b"none", true) guard alongside the
helper call.

Source: Coding guidelines

src/css/values/easing.rs (1)

131-196: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Dead-but-safe fallback arms are fine.

Lines 188-192 duplicate the outer match's Linear/Ease/EaseIn/EaseOut/EaseInOut arms and are unreachable in practice (already excluded by the outer _ branch), but writing explicit keyword strings instead of unreachable!() is a safe, low-risk way to satisfy exhaustiveness per the PR's hardening goal. Optionally, this could be flattened into a single match to avoid the duplicated arms, but that's a pre-existing structural choice, not introduced by this diff.

🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/css/values/easing.rs` around lines 131 - 196, Simplify
EasingFunction::to_css by flattening the nested match so the
Linear/Ease/EaseIn/EaseOut/EaseInOut cases are handled once instead of being
duplicated in the fallback arm. Keep the existing special-case handling for
CubicBezier and Steps, but remove the redundant keyword-string arms in the inner
match to make the control flow clearer and easier to maintain.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@src/css/properties/properties_generated.rs`:
- Around line 2806-2827: The allowlists in the property parsing branch for
animation properties incorrectly include VendorPrefix::MS, which lets
unsupported -ms-animation and -ms-animation-name through. Update the matching
logic in properties_generated.rs for the PropertyId::AnimationName and
PropertyId::Animation cases to remove VendorPrefix::MS from the allowed prefixes
so they only accept prefixes actually emitted by PrefixFeature::Animation and
PrefixFeature::AnimationName.

In `@src/css/rules/mod.rs`:
- Around line 852-861: The placeholder StyleRule created in the code path around
core::mem::replace should use Location::dummy() instead of Location::default()
so synthesized rules are marked consistently with the rest of the codebase.
Update the style::StyleRule initialization in this block to use the dummy source
location sentinel, keeping the existing selectors, vendor_prefix, declarations,
and rules unchanged.

In `@src/css/selectors/parser.rs`:
- Around line 3555-3565: The attribute selector parsing in parser logic is
dropping namespaces for lowercase namespaced attributes because the
value-selector path only falls back to AttributeOther when namespace.is_some()
and the local name is not ASCII lowercase. Update the selector handling around
AttrQName::Specific/Any and the attribute value branch so the optimized
no-namespace variant is used only when namespace.is_none(), and keep namespaced
selectors like svg|href routed through the namespaced path even for lowercase
local names.

---

Outside diff comments:
In `@src/css/properties/animation.rs`:
- Around line 154-267: The name-disambiguation logic in AnimationName::to_css is
duplicated across direction, fill_mode, and play_state, so extract it into a
shared helper and reuse it at each site. Add a small helper like
name_disambiguates::<T>(dest, name_str) that wraps the
css::parse_utility::parse_string::<T>(dest.arena, n, T::parse).is_ok() check,
then call it for AnimationDirection, AnimationFillMode, and AnimationPlayState.
Keep the fill_mode-specific !strings::eql_case_insensitive_ascii(n, b"none",
true) guard alongside the helper call.

In `@src/css/values/easing.rs`:
- Around line 131-196: Simplify EasingFunction::to_css by flattening the nested
match so the Linear/Ease/EaseIn/EaseOut/EaseInOut cases are handled once instead
of being duplicated in the fallback arm. Keep the existing special-case handling
for CubicBezier and Steps, but remove the redundant keyword-string arms in the
inner match to make the control flow clearer and easier to maintain.
🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

  • Push a commit to this branch (recommended)
  • Create a new PR with the fixes

ℹ️ Review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: c2de3420-ccc0-4dcc-be68-ec5b872ec581

📥 Commits

Reviewing files that changed from the base of the PR and between 2bb0c82 and f557d51.

📒 Files selected for processing (25)
  • src/css/css_parser.rs
  • src/css/lib.rs
  • src/css/media_query.rs
  • src/css/properties/animation.rs
  • src/css/properties/background.rs
  • src/css/properties/border.rs
  • src/css/properties/custom.rs
  • src/css/properties/margin_padding.rs
  • src/css/properties/properties_generated.rs
  • src/css/properties/size.rs
  • src/css/rules/container.rs
  • src/css/rules/mod.rs
  • src/css/selectors/parser.rs
  • src/css/selectors/selector.rs
  • src/css/small_list.rs
  • src/css/targets.rs
  • src/css/values/calc.rs
  • src/css/values/color.rs
  • src/css/values/easing.rs
  • src/css/values/gradient.rs
  • src/css/values/image.rs
  • src/css/values/length.rs
  • src/css/values/percentage.rs
  • src/css/values/syntax.rs
  • test/js/bun/css/css.test.ts
💤 Files with no reviewable changes (1)
  • src/css/lib.rs

@coderabbitai coderabbitai Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Caution

Inline review comments failed to post. This is likely due to GitHub's internal server error or limits when posting large numbers of comments. If you are seeing this consistently it is likely a permissions issue. Please check "Moderation" -> "Code review limits" under your organization settings.

Actionable comments posted: 3

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (2)
src/css/properties/animation.rs (1)

154-267: 📐 Maintainability & Code Quality | 🟠 Major | ⚡ Quick win

Extract the repeated name-disambiguation check.

direction, fill_mode, and play_state each repeat the same css::parse_utility::parse_string::<T>(dest.arena, n, T::parse).is_ok() pattern against name_str. Per this repo's own review rule, a pattern repeated a third time in the same diff should be pulled into a shared helper.

♻️ Proposed helper
+fn name_disambiguates<T: css::EnumProperty>(dest: &Printer, name_str: Option<&[u8]>) -> bool {
+    name_str.is_some_and(|n| {
+        css::parse_utility::parse_string::<T>(dest.arena, n, T::parse).is_ok()
+    })
+}

Then call name_disambiguates::<AnimationDirection>(dest, name_str), etc., at each site (keeping the extra !eql_case_insensitive_ascii(n, b"none", true) guard specific to fill_mode).

As per coding guidelines, "The second time a multi-line block appears in your diff, extract a named helper and use it at EVERY parallel site."

🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/css/properties/animation.rs` around lines 154 - 267, The
name-disambiguation logic in AnimationName::to_css is duplicated across
direction, fill_mode, and play_state, so extract it into a shared helper and
reuse it at each site. Add a small helper like name_disambiguates::<T>(dest,
name_str) that wraps the css::parse_utility::parse_string::<T>(dest.arena, n,
T::parse).is_ok() check, then call it for AnimationDirection, AnimationFillMode,
and AnimationPlayState. Keep the fill_mode-specific
!strings::eql_case_insensitive_ascii(n, b"none", true) guard alongside the
helper call.

Source: Coding guidelines

src/css/values/easing.rs (1)

131-196: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Dead-but-safe fallback arms are fine.

Lines 188-192 duplicate the outer match's Linear/Ease/EaseIn/EaseOut/EaseInOut arms and are unreachable in practice (already excluded by the outer _ branch), but writing explicit keyword strings instead of unreachable!() is a safe, low-risk way to satisfy exhaustiveness per the PR's hardening goal. Optionally, this could be flattened into a single match to avoid the duplicated arms, but that's a pre-existing structural choice, not introduced by this diff.

🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/css/values/easing.rs` around lines 131 - 196, Simplify
EasingFunction::to_css by flattening the nested match so the
Linear/Ease/EaseIn/EaseOut/EaseInOut cases are handled once instead of being
duplicated in the fallback arm. Keep the existing special-case handling for
CubicBezier and Steps, but remove the redundant keyword-string arms in the inner
match to make the control flow clearer and easier to maintain.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@src/css/properties/properties_generated.rs`:
- Around line 2806-2827: The allowlists in the property parsing branch for
animation properties incorrectly include VendorPrefix::MS, which lets
unsupported -ms-animation and -ms-animation-name through. Update the matching
logic in properties_generated.rs for the PropertyId::AnimationName and
PropertyId::Animation cases to remove VendorPrefix::MS from the allowed prefixes
so they only accept prefixes actually emitted by PrefixFeature::Animation and
PrefixFeature::AnimationName.

In `@src/css/rules/mod.rs`:
- Around line 852-861: The placeholder StyleRule created in the code path around
core::mem::replace should use Location::dummy() instead of Location::default()
so synthesized rules are marked consistently with the rest of the codebase.
Update the style::StyleRule initialization in this block to use the dummy source
location sentinel, keeping the existing selectors, vendor_prefix, declarations,
and rules unchanged.

In `@src/css/selectors/parser.rs`:
- Around line 3555-3565: The attribute selector parsing in parser logic is
dropping namespaces for lowercase namespaced attributes because the
value-selector path only falls back to AttributeOther when namespace.is_some()
and the local name is not ASCII lowercase. Update the selector handling around
AttrQName::Specific/Any and the attribute value branch so the optimized
no-namespace variant is used only when namespace.is_none(), and keep namespaced
selectors like svg|href routed through the namespaced path even for lowercase
local names.

---

Outside diff comments:
In `@src/css/properties/animation.rs`:
- Around line 154-267: The name-disambiguation logic in AnimationName::to_css is
duplicated across direction, fill_mode, and play_state, so extract it into a
shared helper and reuse it at each site. Add a small helper like
name_disambiguates::<T>(dest, name_str) that wraps the
css::parse_utility::parse_string::<T>(dest.arena, n, T::parse).is_ok() check,
then call it for AnimationDirection, AnimationFillMode, and AnimationPlayState.
Keep the fill_mode-specific !strings::eql_case_insensitive_ascii(n, b"none",
true) guard alongside the helper call.

In `@src/css/values/easing.rs`:
- Around line 131-196: Simplify EasingFunction::to_css by flattening the nested
match so the Linear/Ease/EaseIn/EaseOut/EaseInOut cases are handled once instead
of being duplicated in the fallback arm. Keep the existing special-case handling
for CubicBezier and Steps, but remove the redundant keyword-string arms in the
inner match to make the control flow clearer and easier to maintain.
🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

  • Push a commit to this branch (recommended)
  • Create a new PR with the fixes

ℹ️ Review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: c2de3420-ccc0-4dcc-be68-ec5b872ec581

📥 Commits

Reviewing files that changed from the base of the PR and between 2bb0c82 and f557d51.

📒 Files selected for processing (25)
  • src/css/css_parser.rs
  • src/css/lib.rs
  • src/css/media_query.rs
  • src/css/properties/animation.rs
  • src/css/properties/background.rs
  • src/css/properties/border.rs
  • src/css/properties/custom.rs
  • src/css/properties/margin_padding.rs
  • src/css/properties/properties_generated.rs
  • src/css/properties/size.rs
  • src/css/rules/container.rs
  • src/css/rules/mod.rs
  • src/css/selectors/parser.rs
  • src/css/selectors/selector.rs
  • src/css/small_list.rs
  • src/css/targets.rs
  • src/css/values/calc.rs
  • src/css/values/color.rs
  • src/css/values/easing.rs
  • src/css/values/gradient.rs
  • src/css/values/image.rs
  • src/css/values/length.rs
  • src/css/values/percentage.rs
  • src/css/values/syntax.rs
  • test/js/bun/css/css.test.ts
💤 Files with no reviewable changes (1)
  • src/css/lib.rs
🛑 Comments failed to post (3)
src/css/properties/properties_generated.rs (1)

2806-2827: 🗄️ Data Integrity & Integration | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
# Verify PrefixFeature has Animation/AnimationName variants and correct browser mappings
rg -n 'PrefixFeature' src/css/targets.rs | head -50
rg -n 'Animation' src/css/targets.rs

Repository: oven-sh/bun

Length of output: 203


🏁 Script executed:

#!/bin/bash
set -euo pipefail

echo '--- PrefixFeature enum and mappings ---'
rg -n 'enum PrefixFeature|PrefixFeature::Animation|AnimationName|AnimationTimelineShorthand|prefix_feature' src/css -g'*.rs' -g'*.ts'

echo '--- targets.rs around Feature mappings ---'
sed -n '430,540p' src/css/targets.rs

echo '--- properties_generated.rs around the cited lines ---'
sed -n '2788,2845p' src/css/properties/properties_generated.rs

Repository: oven-sh/bun

Length of output: 13439


🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Locate the PrefixFeature type and any Animation-related variants/usages.
rg -n 'PrefixFeature' src/css
rg -n 'AnimationName|AnimationTimelineShorthand|Animation\b' src/css/targets.rs src/css -g'*.rs'

Repository: oven-sh/bun

Length of output: 13883


🏁 Script executed:

#!/bin/bash
set -euo pipefail

sed -n '1,340p' src/css/prefixes.rs

Repository: oven-sh/bun

Length of output: 9825


🌐 Web query:

caniuse CSS animations vendor prefixes -webkit -moz -o -ms animation animation-name

💡 Result:

As of July 5, 2026, vendor prefixes for CSS animations—including -webkit-, -moz-, -o-, and -ms-—are not required for modern web development [1][2]. Standard, unprefixed CSS animation properties and the @keyframes at-rule are widely supported across all modern browsers, including Chrome, Edge, Firefox, Safari, and Opera [1][3]. The necessity for these prefixes was relegated to legacy browser versions that have long since been deprecated or phased out [1][4][2]. Key points regarding vendor prefixes for animations: - Unprefixed support: Modern browsers (e.g., Chrome 43+, Firefox 16+, Safari 9+, Edge 12+) support the standard, unprefixed animation syntax by default [1][3]. - Legacy usage: Prefixes like -webkit- or -moz- were used in early browser implementations (e.g., Firefox 5-15, Chrome 4-42) to provide experimental support, but these are no longer needed for current projects [1]. - No -ms- support: The -ms- prefix was never officially used for CSS animations or keyframes in Internet Explorer [5]. - Best practice: Instead of using vendor prefixes, you should use the standard, unprefixed properties [2]. If you need to detect support for older environments or specific features, use the @supports at-rule (e.g., @supports (animation-name: test)) or feature detection in JavaScript [6][7]. Including vendor prefixes in your CSS today serves no functional purpose for modern users and can even lead to bloated, harder-to-maintain code [5][4]. If you are working on a project that requires supporting extremely old, legacy browsers, you may choose to use build-time tools (such as Autoprefixer) to automatically manage any remaining necessary prefixes based on your specified target browser list [2].

Citations:


🌐 Web query:

caniuse animation -ms prefix @-ms-keyframes``

💡 Result:

There is no such thing as an @-ms-keyframes at-rule in CSS [1][2]. Internet Explorer 10 and later, which introduced support for CSS animations, implemented the standard unprefixed @keyframes syntax [2][3]. The -ms- prefix was never used for keyframes in any stable version of Internet Explorer or Microsoft Edge [1][4][2]. Any references to @-ms-keyframes often stem from misunderstandings or pre-release developer preview versions that are long obsolete [1][5]. Using an -ms- prefix for keyframes is not only unnecessary but can also cause the entire animation rule to be ignored by the browser [4]. To implement CSS animations, you should use only the standard, unprefixed @keyframes syntax [4][2][6].

Citations:


🏁 Script executed:

#!/bin/bash
set -euo pipefail

sed -n '1,220p' src/css/prefixes.rs
printf '\n--- property id prefix handling ---\n'
sed -n '1260,1375p' src/css/properties/properties_generated.rs
printf '\n--- feature/prefix application logic ---\n'
sed -n '1,120p' src/css/properties/properties_generated.rs

Repository: oven-sh/bun

Length of output: 12278


🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Find how VendorPrefix is converted/filtered during serialization.
rg -n 'intersects\(pre\)|VendorPrefix::MS|prefixes_for\(|prefix_feature\(\)' src/css -g'*.rs'

Repository: oven-sh/bun

Length of output: 29819


🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Inspect whether the parser intentionally allows prefixes that prefixes_for never emits.
sed -n '1,140p' src/css/prefixes.rs
printf '\n---\n'
sed -n '1260,1375p' src/css/properties/properties_generated.rs
printf '\n---\n'
sed -n '5800,5865p' src/css/properties/properties_generated.rs

Repository: oven-sh/bun

Length of output: 10260


🏁 Script executed:

#!/bin/bash
set -euo pipefail

sed -n '320,390p' src/css/prefixes.rs

Repository: oven-sh/bun

Length of output: 3223


🏁 Script executed:

#!/bin/bash
set -euo pipefail

python3 - <<'PY'
from pathlib import Path
text = Path('src/css/prefixes.rs').read_text().splitlines()
for start in [330]:
    for i in range(start-1, min(len(text), start+80)):
        print(f"{i+1}:{text[i]}")
PY

Repository: oven-sh/bun

Length of output: 3985


🏁 Script executed:

#!/bin/bash
set -euo pipefail

python3 - <<'PY'
from pathlib import Path
text = Path('src/css/prefixes.rs').read_text().splitlines()
for i in range(330-1, min(len(text), 410)):
    print(f"{i+1}:{text[i]}")
PY

Repository: oven-sh/bun

Length of output: 3985


Remove VendorPrefix::MS from the animation allowlists. PrefixFeature::Animation and PrefixFeature::AnimationName never emit MS, so -ms-animation / -ms-animation-name are accepted here even though the feature table doesn't support them.

🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/css/properties/properties_generated.rs` around lines 2806 - 2827, The
allowlists in the property parsing branch for animation properties incorrectly
include VendorPrefix::MS, which lets unsupported -ms-animation and
-ms-animation-name through. Update the matching logic in properties_generated.rs
for the PropertyId::AnimationName and PropertyId::Animation cases to remove
VendorPrefix::MS from the allowed prefixes so they only accept prefixes actually
emitted by PrefixFeature::Animation and PrefixFeature::AnimationName.
src/css/rules/mod.rs (1)

852-861: 🎯 Functional Correctness | 🔵 Trivial | 💤 Low value

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
# Check whether Location derives/implements Default, and how dummy()/default() are used elsewhere
rg -n -B3 -A3 'struct Location' src/css/lib.rs
rg -n -B2 -A8 'impl Location' src/css/lib.rs
rg -n 'Location::default\(\)|Location::dummy\(\)' src/css

Repository: oven-sh/bun

Length of output: 898


🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Inspect the surrounding code paths for the placeholder rule and nearby location handling.
sed -n '620,670p' src/css/rules/mod.rs
printf '\n---\n'
sed -n '840,870p' src/css/rules/mod.rs
printf '\n---\n'
sed -n '2548,2580p' src/css/css_parser.rs

Repository: oven-sh/bun

Length of output: 4762


🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Inspect the Location definition and all call sites to distinguish Default vs dummy semantics.
sed -n '320,350p' src/css/lib.rs
printf '\n---\n'
rg -n -A2 -B2 'Location::default\(\)|Location::dummy\(\)' src/css

Repository: oven-sh/bun

Length of output: 1995


Use Location::dummy() for the placeholder rule

Location::default() is fine here, but dummy() is the established sentinel for synthesized rules with no real source location, so this keeps placeholder handling consistent.

🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/css/rules/mod.rs` around lines 852 - 861, The placeholder StyleRule
created in the code path around core::mem::replace should use Location::dummy()
instead of Location::default() so synthesized rules are marked consistently with
the rest of the codebase. Update the style::StyleRule initialization in this
block to use the dummy source location sentinel, keeping the existing selectors,
vendor_prefix, declarations, and rules unchanged.
src/css/selectors/parser.rs (1)

3555-3565: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick win

Preserve namespaces for lowercase attribute selectors.

The AttrQName::Specific/Any mapping creates Some(namespace), but the value-selector path below only uses AttributeOther when namespace.is_some() && !local_name_is_ascii_lowercase. Lowercase namespaced attributes like [svg|href=x] can fall through to AttributeInNoNamespace and lose the namespace. Gate the optimized no-namespace variant on namespace.is_none() instead.

Proposed fix
-    if namespace.is_some() && !local_name_is_ascii_lowercase {
+    if namespace.is_some() || !local_name_is_ascii_lowercase {
📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

if namespace.is_some() || !local_name_is_ascii_lowercase {
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/css/selectors/parser.rs` around lines 3555 - 3565, The attribute selector
parsing in parser logic is dropping namespaces for lowercase namespaced
attributes because the value-selector path only falls back to AttributeOther
when namespace.is_some() and the local name is not ASCII lowercase. Update the
selector handling around AttrQName::Specific/Any and the attribute value branch
so the optimized no-namespace variant is used only when namespace.is_none(), and
keep namespaced selectors like svg|href routed through the namespaced path even
for lowercase local names.

@claude claude Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for addressing the earlier nits — both look fixed. No bugs found, but this is a ~25-file structural refactor of the CSS crate (selector qname parsing split, minify_style_arm placeholder change, calc/length add arm reshaping, new ColorFallback/as_single semantics), so it warrants a pass from someone who owns src/css/.

Extended reasoning...

Overview

This PR removes all 66 unreachable!() sites from src/css/ across 25 files. The changes fall into several categories: (1) type restructuring — new ColorFallback enum split from the ColorFallbackKind bitflags, to_light_dark() returning a tuple, parse_qualified_name split into type-selector and attribute-selector variants with distinct return types, extract_* returning Option<&T>; (2) dead-code deletion — AnimationTimeline::{Scroll,View} and payload types, AtRulePrelude::FontFeatureValues, Token::raw, VendorPrefix::from_name_str, four deprecated to_css shims; (3) borrowck-workaround collapse — matches!-then-rematch patterns folded into direct or-pattern matches in calc.rs, length.rs, percentage.rs, selector.rs, custom.rs; (4) debug_assert! + safe fallback for cssparser delimiter contracts. Regression tests were added for the two user-input-reachable sites.

Security risks

None identified. This is purely internal to the CSS parser/printer/minifier; no auth, crypto, network, filesystem, or FFI surface is touched. The changes replace release-build panics with either compile-time exhaustiveness or debug-assert-plus-graceful-fallback, which strictly reduces DoS surface on adversarial CSS input.

Level of scrutiny

High. While each individual pattern is defensible, several changes need equivalence verification by someone familiar with the crate:

  • The parse_qualified_name → parse_qualified_name + parse_attr_qualified_name split duplicates ~60 lines of parser control flow with subtly different token handling (the *-after-star and *-as-local-name paths diverge between the two).
  • minify_style_arm now takes &mut StyleRule<R> and leaves an empty StyleRule placeholder via mem::replace instead of CssRule::Ignored; whether any downstream code observes the placeholder needs confirmation.
  • The length.rs/percentage.rs add restructuring widens the first two match arms (dropping the matches!(*ca, Calc::Value(_)) guard) and handles the non-Value case inline; I traced this as equivalent to the old fallthrough via into_calc(), but it deserves a second pair of eyes.
  • ColorFallbackKind::as_single() returns None for the OKLAB bit (and any multi-bit set), where the old get_fallback path would have panicked — whether lowest() can ever yield OKLAB determines if this is a silent behavior change.
  • ColorFallback::supports_condition() gained an Rgb arm that the old bitflags version panicked on; it appears unreachable from current callers but is new API surface.

Other factors

Both of my earlier cosmetic comments (gradient.rs -> spacing, stale Multiplier comment in syntax.rs) were addressed in follow-up commits and the threads are resolved. The bug-hunting system found nothing this run. Test coverage is stated as 1328 passing across the CSS/bundler suites. The ~200-variant exhaustive match in targets.rs is verbose but does exactly what the description says (compile-time coverage of compat::Feature). Given the breadth of parser/serializer control-flow reshaping, this is the kind of PR a src/css/ owner should sign off on rather than a bot.

…lector tests

- targets.rs: from_compat returns Option<Features>; should_compile_same falls
  back to !is_compatible() for unmapped features instead of the always-true
  contains(empty()) path.
- properties_generated.rs: debug_assert on __to_css_helper Custom arm.
- css_parser.rs: debug_assert on the second parse_until_before contract site,
  matching its sibling.
- css.test.ts: drop the 6 tests that exercised unchanged arms; add QNamePrefix
  selector matrix ([svg|attr], |foo, |*, svg|*, [*] error).

@claude claude Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Both earlier nits are addressed and I didn't find any bugs this pass, but this is a ~25-file refactor that reshapes several public-ish contracts (ColorFallback enum, split qualified-name parsing, minify_style_arm signature, length/percentage add rewrites) — worth a human sanity-check on the behavior-preservation claims before merge.

Extended reasoning...

Overview

This PR eliminates all 67 unreachable!() sites from src/css/ across 25 files. The changes fall into four buckets: (1) type restructuring so invalid states are unrepresentable — new ColorFallback enum split from the ColorFallbackKind bitflags, CssColor::to_light_dark() returning a tuple, parse_qualified_name split into type-selector and attribute-selector variants with distinct return types, SizeHandlerSpec::extract_* returning Option<&T>, new RepeatSeparator enum; (2) collapsing stale borrowck-workaround matches!-then-rematch patterns into direct or-pattern matches (calc.rs, length.rs, percentage.rs, selector.rs, custom.rs); (3) deleting dead variants/helpers (AnimationTimeline::{Scroll,View} + payload types, AtRulePrelude::FontFeatureValues, Token::raw, VendorPrefix::from_name_str, four deprecated to_css shims); (4) converting cssparser-contract invariants to debug_assert! + safe fallback.

Security risks

None identified. This is entirely internal to the CSS parser/minifier — no auth, crypto, filesystem, network, or user-input validation surfaces are touched. The changes convert release-build panics into either compile errors or debug-asserts with graceful fallbacks, which is strictly a robustness improvement (a reachable unreachable!() on user CSS input was previously a DoS vector).

Level of scrutiny

Medium-high. While the intent (remove panics) is straightforward and the individual patterns are sound, several changes require careful behavior-preservation verification that I can't fully validate without running the suite:

  • length.rs/percentage.rs add rewrites restructure the match arms — the new nested-match shape must produce identical Calc::Sum trees for every (Calc-variant × non-Calc) combination as before.
  • minify_style_arm now takes &mut StyleRule and swaps in an empty StyleRule placeholder instead of CssRule::Ignored — the caller loop's subsequent handling of the swapped-out slot needs to be equivalent.
  • parse_attr_qualified_name is ~60 lines of freshly-written parsing logic split from the shared function; the added tests cover the matrix but a maintainer familiar with the Lightning CSS reference should confirm the error-kind choices match.
  • ColorFallback::Rgb.supports_condition() is a new code path (previously unreachable!()) — it's never called in practice per the callers, but the value chosen (rgb(0, 0, 0)) is a design decision.
  • The 200-variant exhaustive arm in Features::from_compat is mechanical but easy to typo.

Other factors

Both of my earlier cosmetic nits (gradient.rs -> spacing, syntax.rs stale comment) are resolved in the current diff. The bug-hunting system found nothing this run. Tests were added for the QNamePrefix split. The PR description is thorough and the commit log shows iterative review response. Given the scope (25 files, core CSS pipeline) and the number of independent refactors bundled together, a human maintainer sign-off is the right bar here — this is well beyond a mechanical change even though each individual pattern is defensible.

@alii
alii marked this pull request as draft July 6, 2026 19:59
Jarred-Sumner added a commit that referenced this pull request Aug 17, 2026
…ll_jsc, node-fallbacks, and misc crates (#39420)

Net -4,402 lines (172 files, +509 / -4,911). Everything removed has zero
references across `src/`, `scripts/`, `test/`, `packages/` and the
regenerated `build/debug/codegen/` output, and the removal builds:
removing a Rust or C++ definition that still had a caller fails to
compile or link, so a green build is the reference check for those; JS,
codegen and manifest removals were additionally grepped by name
(including `bun:internal-for-testing` consumers under `test/`).

Most of these were first identified by the sweeps that were closed
yesterday for merge conflicts (#37272, #37062, #38439, #37089, #38703).
This PR re-applies the subset that still applies on current main, minus
anything an open PR already deletes and minus small hunks in files that
change daily (see "Left out" below), plus a few new finds.

### react_compiler (-1,166)

- `validate_no_derived_computations_in_effects_exp` and its ~22
exclusive helpers/types (~1.1k lines in
`validation/validate_no_derived_computations_in_effects.rs`). The
pipeline only ever calls the non-`_exp` validation; the `_exp`
env-config flag was parsed from fixture pragmas and read by nothing.
`react-compiler-fixtures.test.ts` now lists the two pragmas as ignored
instead of handled; the fixture suite still passes.
- `SymbolHost` (back-compat alias of `Host`; `DESIGN.md` updated),
`HirBox`, `is_use_state_type`, `default_true`.

### C++ bindings (-1,350 across 64 files)

- `JSDOMConvertBufferSource.h`: the IDL typed-array specializations and
`toPossiblyShared*Array` helpers for every element type no binding
converts (only the Uint8Array/ArrayBuffer views are used).
- `JSDOMPromiseDeferred.h/.cpp`: `resolveWithJSValue`,
`resolveWithNewlyCreated`, `resolveCallbackValueWithNewlyCreated` and
friends.
- `IDLTypes.h`: the `IDLUnsupportedType` family, the IDB / WebGL /
`ScheduledAction` wrappers and stale forward declarations;
`JSDOMConvertDate.h/.cpp` (the only `IDLDate` converter) deleted along
with its two includes. The 8-line `IDLDate` struct itself stays for now
(see "Left out"). `JSDOMConvertScheduledAction.h`, orphaned in the same
way, is already deleted by #35775, so it is not touched here.
- `NetworkLoadMetrics.h`: WebKit networking-stack fields/accessors bun
never reads.
- Deleted files: `JSWorkerOptions.h/.cpp` (`JSWorker.cpp` builds
`WorkerOptions` by hand), `JSMIMEBindings.h/.cpp` (`createMIMEBinding`
had no callers; include dropped from `ZigGlobalObject.cpp`),
`JSDOMIterator.cpp` (`addValueIterableMethods`).
- `ncrypto.h/.cpp`: `peekError`, `BignumPointer::isZero`,
`EVPKeyCtxPointer::sign`, unused copy/move `operator=` overloads and an
`AsymmetricKeyEncodingConfig` constructor.
- Smaller: `JSDOMOperationReturningPromise.h`
(`call*ReturningOwnPromise`), `JSDOMAttribute.h`
(`setPassingPropertyName`, `setStatic`), `JSDOMGuardedObject`
(`DoNotRegisterWithGlobalObjectTag`), `Event`/`EventTarget`
(`resetBeforeDispatch`, default-handled flags, `isNode()`),
`EventEmitter::{eventTypes,eventListeners}`,
`HTTPHeaderIdentifiers::identifierFor`, `JSURLPatternResult`
`convertDictionary` stubs,
`PerformanceTiming::monotonicTimeToIntegerMilliseconds`,
`ContextDestructionObserver::protectedScriptExecutionContext`,
`expectedEnumerationValues<CryptoKeyUsage>`,
`BufferSource::mutableData`/`toBufferSource`, `BunString__toInt32`,
`BunString::utf8ByteLength`, the `Ref<StringImpl>` overload of
`toCrossThreadShareable`, `normalWorld()`, `TextEncoding(const
String&)`, `JSBuffer` `createBuffer`/`constructFromEncoding` overloads,
the `JSX509Certificate` `m_infoAccess` lazy property and the non-legacy
branch of `computeInfoAccess` (the prototype getter reads the view
directly; x509 tests pass),
`BakeAdditionsToGlobalObject::wrapComponent`,
`jsFunction_lsanDoLeakCheck`, and the dead `BunObject+exports.h` macro
entries (together with the Rust `BunObject_callback_nanoseconds` export
the `nanoseconds` entry declared; `Bun.nanoseconds` is
`functionBunNanoseconds` in `BunObject.cpp`).

### install_jsc / install_types / bun:internal-for-testing (-344)

- `install_jsc/dependency_jsc.rs` and `update_request_jsc.rs` deleted:
their only consumers were the `npa` / `npmTag` exports of
`bun:internal-for-testing`, which no test imports. The
`dispatch_js2native.rs` re-exports, the `generate-js2native.ts` file-map
entry and the `lsanDoLeakCheck` export (tests use `isASANEnabled`) go
with them.
- `install_types/lib.rs`: the `ExternalString` / `SlicedString` /
`SemverString` re-export modules; nothing names those paths (everything
imports the types from `bun_semver`). New in this PR.

### node-fallbacks (-400) and codegen (-392)

- `util.js`: a ~270 line commented-out `util.types` block.
- `package.json` / `bun.lock` / `tsconfig.json`: dependencies and path
mappings nothing imports (`esbuild`, `buffer`, `events`, `util`, `url`,
`process`, `path-browserify`, `os-browserify`, `timers-browserify`,
`tty-browserify`, `vm-browserify`, ...). `build-fallbacks.ts` marks
every builtin name external, and the remaining sources only import the
packages still listed; `bun install --frozen-lockfile` in the directory
is a no-op and `bundler_browser.test.ts` passes.
- `src/codegen/generate-unified-source-bundles.rb`: WebKit's Ruby
generator, superseded by `scripts/build/unified.ts`; nothing invokes it.

### Rust crates (-1,200 across bun_jsc, bun_runtime, css, uws_sys and
leaf crates)

- bun_jsc: `BuiltinName::get` + `BUILTIN_NAME_MAP`,
`MarkedArrayBuffer::to_js` (the `ArrayBuffer::alloc` Uint8Array arm and
its `Bun__allocUint8ArrayForCopy` binding are kept, so `alloc` keeps
mirroring `create`), the `__dangerouslySetPtr` wrapper
`js_class_module!` emitted into every class, `ErrorBuilder::new`,
`TopExceptionScope::new`, `Task::new`, `JSCell::to_js`,
`JSGlobalObject::{to_js,ref_,ctx}`, `job::{on_js_thread,off_thread}`,
`AbortReason` impl, `JSPromise` settle helpers, `UUID::ZERO`,
`TagPayload::get`.
- bun_runtime: the `target_os = "wasi"` directory-iterator backend in
`dir_iterator.rs` (no shipped target is wasi and the `bun_sys::wasi`
module it imports does not exist), the `Display` impls in
`assert/myers_diff.rs`, the non-unix stub and not-macos escapes in
`fs_events.rs` (the file is only compiled on macOS),
`ArrayBufferSink::to_js`, unused re-exports in `node.rs` /
`api/bun/spawn.rs` / `ffi/mod.rs` (with `abi_type` formatters narrowed
to `pub(crate)`), `Error::UnableToDecode`,
`MyersDiff::Error::OutOfMemory`. The first three are new in this PR.
- css: the inherent `eql` / `to_css` / `parse` forwarders whose callers
all go through the `CssEql` / `ToCss` / `Parse` trait impls
(`values/calc.rs` and friends), `generics::{implement_eql,parse}`,
`TokenList::parse_with_options`, `CssString::parse`.
- uws_sys: the `uws_loop_defer`, `uws_res_clear_corked_socket`,
`uws_ws_iterate_topics` and `uws_h3_req_get_parameter` C shims plus
their Rust declarations and wrappers
(`Loop::{uncork,wake,next_tick,run}`), `AnyResponse::init`,
`SocketGroup::is_empty`, `socket.rs` `group()` accessors and the
`SocketTcp`/`SocketTls` aliases, `Opcode::Close`, `WindowsLoop`.
- leaf crates: `windows_sys` constants and their `bun_sys::windows`
re-exports, `zlib`/`zlib_sys` declarations (`deflateInit_`,
`inflateInit_`, the `gz*` file API, legacy type aliases), `sha_hmac`
deprecated-API hashers (`SHA512` raw, `RIPEMD160`, `MD5_SHA1`, `Blake2`
evp), `wyhash` `HashInt` impls for u16/u64, `libarchive` commented-out
Zig-era callbacks, `bun_alloc` (`AllocError::name`, `usable_size`,
`BSSList::init`), `clap::Error::WriteFailed`, `csrf` error variants,
`pe::Error::{InputIsSigned,InsufficientSpace}`, `md` `Setextheader`,
`opaque_mut_nn`, `cares_sys` `AddrInfo_hints::is_empty`, `boringssl_sys`
constants, `errno` `Mode` re-exports, `bounded_array::get`,
`string::write::Result`, `SplitIterator::rest`, `OutOfRangeValue` impls,
`symbol::Map::init`, `sql_jsc` re-exports.

### src/js (-32)

- Unused REPL primordials entries in
`internal/repl/node-primordials.js`; with that gone `SafeWeakSet` had no
importer, so `internal/primordials.js` stops exporting it (new in this
PR). Unused export-object entries in `internal/fs/watch.ts` and
`internal/readline/interface.js`.

### scripts (-9)

- `glob-sources.ts` `src/*.c` pattern (matched nothing since
`asan-config.c` was deleted), the write-only `BUN_DEP_*` defines in
`depVersionsHeader.ts`, the unread `kqueue` config field.

### Verification

- `bun bd` (full debug build) passes.
- `bun run rust:check-all`: all 11 target triples ok (covers the
windows/darwin-only removals in `windows_sys`, `sys/windows`,
`zlib_sys/win32.rs`, `fs_events.rs`, `windows-shim`).
- `cargo fmt --check`, `cargo clippy --workspace`, clang-format on every
touched C++ file, prettier, and `bun run lint` are clean.
- All of `test/internal/source-lints/` (including the new
`dead-symbols-react-compiler-webcore-idl-misc.test.ts` that guards these
symbols, and `dead-code-escapes` against the updated
`dead-code-escape-limits.json`), `react-compiler-fixtures.test.ts`,
`bundler_browser.test.ts`, css, cryptohasher/hash, node:assert, zlib,
url, events, websocket, inspect and x509 tests pass. `serve.test.ts` has
the same 4 failures as the unmodified release build in this container
(IPv6 / root port range), nothing else.

### Left out on purpose (follow-up candidates)

- The pre-engine inbound path in `h2_frame_parser.rs` (~2.2k lines,
still dead, #37272's diff still applies cleanly): the file has had 15
commits in the last two weeks, so it is better landed on its own.
- Deletions already owned by open PRs: simdutf wrappers (#38958),
`getStackTraceForThrownValue` (#37450), `validateOneOf` (#38401), the
redis error variants (#34829), `URL::from_js` (#33889 / #34577),
`schema::api` re-exports (#37095), `FsPath` (#39327), `NodeJSFS` `Null`
impl (#38065), the deprecated selector `to_css` (#33332).
- Small hunks in high-churn files (`bindings.cpp`,
`ZigGlobalObject.cpp`, `Blob.rs`, `streams.rs`, `BunProcess.cpp`, the
`ManifestLoad::LoadFromMemory` parameter across `src/install`, the
watcher `loader` parameter), and the `#[no_mangle]` statics
(`Zig_ErrorCode*`, `Bun__versions_*`) nothing on the C++ side reads.
- `VM::has_termination_request` and its `JSC__VM__hasTerminationRequest`
shim in `bindings.cpp`: a dead pair, kept intact here because
`bindings.cpp` is the most actively edited file in the tree; both halves
go together in a follow-up.
- `IDLDate` in `IDLTypes.h`: its only converter is deleted here, but the
struct itself is left in place so this PR's file deletions stay
independent of the header edits; removing the struct is a one-hunk
follow-up once the converter files are gone.
- Found but not removed here: the windows shim's `ReadWithoutLaunch`
mode (~110 lines, overlaps #36200), `node_quic_binding.rs` constants JS
never destructures (~35 lines), the native `NodeJSFS.unwatchFile` /
`FSWatcher.hasRef` / `QuicSession.silentClose` / `QuicEndpoint.ref`
bindings JS never calls, never-constructed `bun_install::Error`
variants, and ~160 lines of `$`-declarations in `src/js/builtins.d.ts`
with no users.

<!-- robobun:evidence:begin -->

---

**[review]** gate passed · iteration 1 · 172 files touched

<details><summary>fails on main (without fix)</summary>

```console
ASAN without fix: 10 failed, 320 skipped
$ BUN_DEBUG_QUIET_LOGS=1 bun scripts/build.ts --profile=debug --quiet test "--reporter=junit" "--reporter-outfile=/tmp/mechgate.xml" test/bundler/transpiler/react-compiler-fixtures.test.ts test/internal/source-lints/dead-symbols-react-compiler-webcore-idl-misc.test.ts
bun test v1.4.0 (8326d1b)

test/internal/source-lints/dead-symbols-react-compiler-webcore-idl-misc.test.ts:
88 |     // Arena box alias with zero uses (HirVec is the one HIR actually uses).
89 |     ["src/react_compiler/hir/mod.rs", /\bHirBox\b/],
90 |     // Type predicate whose only callers were in the removed _exp validation.
91 |     ["src/react_compiler/hir/mod.rs", /\bis_use_state_type\b/],
92 |   ];
93 |   expect(resurrected(checks)).toEqual([]);
                                   ^
error: expect(received).toEqual(expected)

- []
+ [
+   "src/react_compiler/validation/validate_no_derived_computations_in_effects.rs: validate_no_derived_computations_in_effects_exp",
+   "src/react_compiler/hir/environment_config.rs: validate_no_derived_computations_in_effects_exp",
+   "src/react_compiler/program.rs: validate_no_derived_computations_in_effects_exp",
+   "src/react_compiler/program.rs: \bSymbolHo
... (truncated)

release without fix: 10 failed, 1146 skipped
bun test v1.4.0-canary.1 (21a4206)

test/internal/source-lints/dead-symbols-react-compiler-webcore-idl-misc.test.ts:
88 |     // Arena box alias with zero uses (HirVec is the one HIR actually uses).
89 |     ["src/react_compiler/hir/mod.rs", /\bHirBox\b/],
90 |     // Type predicate whose only callers were in the removed _exp validation.
91 |     ["src/react_compiler/hir/mod.rs", /\bis_use_state_type\b/],
92 |   ];
93 |   expect(resurrected(checks)).toEqual([]);
                                   ^
error: expect(received).toEqual(expected)

- []
+ [
+   "src/react_compiler/validation/validate_no_derived_computations_in_effects.rs: validate_no_derived_computations_in_effects_exp",
+   "src/react_compiler/hir/environment_config.rs: validate_no_derived_computations_in_effects_exp",
+   "src/react_compiler/program.rs: validate_no_derived_computations_in_effects_exp",
+   "src/react_compiler/program.rs: \bSymbolHost\b",
+   "src/react_compiler/lib.rs: \bSymbolHost\b",
+   "src/react_compiler/hir/mod.rs: \bHirBox\b",
+   "src/react_compiler/hir/mod.rs: \bis_use_state_type\b",
+ ]

- Expected  - 1
+ Received  + 9

      at <anonymous> (/workspace/bun/test/internal/source
... (truncated)
```

</details>

<details><summary>passes on PR (with fix)</summary>

```console
ASAN with fix: 320 skipped
$ BUN_DEBUG_QUIET_LOGS=1 bun scripts/build.ts --profile=debug --quiet test "--reporter=junit" "--reporter-outfile=/tmp/mechgate.xml" test/bundler/transpiler/react-compiler-fixtures.test.ts test/internal/source-lints/dead-symbols-react-compiler-webcore-idl-misc.test.ts
bun test v1.4.0 (8326d1b)

test/internal/source-lints/dead-symbols-react-compiler-webcore-idl-misc.test.ts:
(pass) dead react_compiler symbols do not reappear [21.23ms]
(pass) dead exe_format symbols do not reappear [2.63ms]
(pass) dead bun_core / bun_alloc / bun_ast / bun_ptr items do not reappear [18.66ms]
(pass) dead bun_css items do not reappear [21.91ms]
(pass) dead bun_jsc items do not reappear [20.67ms]
(pass) dead FFI-crate items do not reappear [30.81ms]
(pass) dead Rust symbols (install, webcore, jsc, leaf crates) do not reappear [18.03ms]
(pass) unused re-export names do not reappear [12.35ms]
(pass) stale build-script entries do not reappear [6.80ms]
(pass) dead Rust FFI wrappers and trait methods do not reappear [22.18ms]
(pass) dead C++ binding helpers do not reappear [99.82ms]
(pass) dead WebCore / IDL binding code does not reappear [114.33ms]
(pass) dead code in install_jsc, install_t
... (truncated)

release with fix: 1146 skipped
$ bun scripts/build.ts --profile=release
[configured] bun-profile → bun (stripped) in 641ms (unchanged)
ninja: Entering directory `/workspace/bun/build/release'
[1/130] install /workspace/bun/src/node-fallbacks
bun install v1.4.0-canary.1 (21a4206)

Checked 111 installs across 104 packages (no changes) [3.00ms]
[2/130] gen node-fallbacks/react-refresh.js
Bundled 1 module in 5ms

  react-refresh.js  4.81 KB  (entry point)

[3/130] gen JSBuffer.lut.h
Generating /workspace/bun/build/release/codegen/JSBuffer.lut.h from /workspace/bun/src/jsc/bindings/JSBuffer.cpp
[4/130] gen generated_host_exports.rs
generated_host_exports.rs: 92 exports (host=3, lazy=10, generic=79, rust=0); 241 extern-C blocks audited
[5/130] gen node-fallbacks/*.js
[6/130] gen cpp.rs (cppbind)
[7/130] gen JS modules (bundle-modules)
Preprocess modules (7887ms)
Bundle modules (42ms)
Postprocesss modules (259ms)
Bundle Functions (695ms)
Generate Code (32ms)

[8.94s] Bundled "src/js" for production
  2630 kb
  197 internal modules
  13 native modules
  91 internal functions across 17 files
[7/129] cargo bun_bin → libbun_rust.a (--target x86_64-unknown-linux-gnu)

  nightly-2026-07-20-x86_64-unknow
... (truncated)
```

</details>

<details><summary>diff hotspot</summary>

```
scripts/build/config.ts                            |    4 -
 scripts/build/depVersionsHeader.ts                 |    3 -
 scripts/build/source.ts                            |    2 +-
 scripts/glob-sources.ts                            |    1 -
 src/ast/symbol.rs                                  |    8 -
 src/boringssl_sys/boringssl.rs                     |    8 -
 src/bun_alloc/lib.rs                               |   29 +-
 src/bun_core/bounded_array.rs                      |    8 -
 src/bun_core/fmt.rs                                |   17 -
 src/bun_core/string/immutable.rs                   |    7 -
 src/bun_core/string/mod.rs                         |   11 -
 src/bun_core/string/write.rs                       |    3 -
 src/bun_core/windows_sys.rs                        |    2 +-
 src/cares_sys/c_ares.rs                            |    6 -
 src/clap/error.rs                                  |    9 -
 src/clap/lib.rs                                    |    5 -
 src/codegen/generate-js2native.ts                  |    1 -
 src/codegen/generate-unified-source-bundles.rb     |  392 -------
 src/csrf/lib.rs                                    |    3 -
 src/css/css_parser.rs                              |   19 +-
 src/css/generics.rs                                |   10 -
 src/css/lib.rs                                     |    2 +-
 src/css/properties/custom.rs                       |    8 +-
 src/css/rules/mod.rs                               |    7 +-
 src/css/values/angle.rs                            |    4 -
 src/css/values/calc.rs                             |  132 +--
 src/css/values/css_string.rs                       |    6 -
 src/css/values/time.rs                             |    7 -
 src/css_derive/lib.rs                              |    4 +-
 src/css_jsc/css_internals.rs                       |   11 -
 src/errno/darwin_errno.rs                          |    1 -
 src/errno/freebsd_errno.rs                         |    1 -
 src/errno/linux_
... (truncated)
```

</details>

**gate history** · 4 passed · 1 rejected · iteration 1

<details><summary>evidence per changed file</summary>

```
file                                reads  edits  tests
scripts/build/config.ts                 0      0      0
scripts/build/depVersionsHeader.ts      0      0      0
scripts/build/source.ts                 0      0      0
scripts/glob-sources.ts                 0      0      0
src/ast/symbol.rs                       0      0      0
src/boringssl_sys/boringssl.rs          0      0      0
src/bun_alloc/lib.rs                    0      0      0
src/bun_core/bounded_array.rs           0      0      0
src/bun_core/fmt.rs                     0      0      0
src/bun_core/string/immutable.rs        0      0      0
src/bun_core/string/mod.rs              0      0      0
src/bun_core/string/write.rs            0      0      0
src/bun_core/windows_sys.rs             0      0      0
src/cares_sys/c_ares.rs                 0      0      0
src/clap/error.rs                       0      0      0
src/clap/lib.rs                         0      0      0
(+ 156 more files)
```

</details>

<!-- robobun:evidence:end -->

---------

Co-authored-by: Jarred Sumner <jarred@jarredsumner.com>
Jarred-Sumner added a commit that referenced this pull request Oct 27, 2026
### What does this PR do?

Consolidates the open pull requests on the CSS parser (`src/css`:
parsing, minifying, lowering, printing, CSS modules) into one.

179 open pull requests were in scope. Each was judged by one rule: **the
output must mean to a browser what the input meant.** Every claim was
run through a build of `main`, a build with the change, lightningcss
1.33 and Chromium 156 (CSSOM, computed values, painted pixels for colors
and gradients), and counted against 10,116 real stylesheets: the
published CSS of about 190 popular npm packages, plus stylesheets from
open source repositories. 107 land here, whole or in part, most of them
cut down. The source diff is net negative.

Of the 10,116, 9,509 are stylesheets that someone ships. The rest are
test fixtures of other tools (521), Tailwind v4 source that needs its
plugin (59) and SCSS, PostCSS or template source (27).

| 9,509 shipped stylesheets | `main` | this branch |
|---|---|---|
| fail to build | 71 | 35 |
| ... of those, valid CSS that Bun rejects | | 0 |
| minified output reads differently in Chromium than the input | 552 | 2
|
| CSS modules with a name that is referenced but not defined, or
exported but not defined | 166 | 0 |

| all 10,116 | `main` | this branch |
|---|---|---|
| fail to build | 192 | 141 |
| minified output reads differently in Chromium than the input | 593 |
15 |
| build on `main`, fail here | | 0 |
| right on `main`, wrong here (minified, and at the default targets) | |
0 |
| size of the minified output, no targets | 162.32 MB | 162.47 MB
(+0.093%) |
| size of the output at the default targets | 224.1 MB | 210.2 MB
(-6.2%) |

A syntax error is a build error, with one exception: the IE star hack
`*zoom: 1` is skipped, because tachyons and handsontable ship it. The 35
shipped stylesheets that fail are invalid CSS: 28 have a selector that
Chromium drops (semantic-ui-css, swiper, @ionic/core, Adobe Spectrum,
xp.css), 2 an `@import` after other rules, and one each `@supports not X
and Y`, a nested `@keyframes`, a stray comma, a selector list in
`::slotted()`, `@media screen\0` (govuk-frontend). The 2 that read
differently fold a `color-mix()` inside a gradient (see the known gaps).

The growth of the minified output is rules and declarations that `main`
deleted by mistake: 87 KB of rules, 45 KB of prefixed declarations that
only browser targets can rule out.

#### Popular CSS that `main` breaks today

- normalize.css, modern-normalize, sanitize.css: `font-family:
monospace, monospace` is collapsed to `monospace`, which changes the
font size from 16px to 13px (#46662).
- CKEditor 5, Tabler, Carbon, VitePress, Ionic, PatternFly, Excalidraw:
a rule of custom properties is deleted as a "duplicate" of a later rule
with the same selector and other custom properties. 37 stylesheets lose
1,681 custom properties (#33684).
- Material Components, daisyUI, DocSearch, plyr: a `var()` size loses to
its own fallback (#42101).
- nextra, daisyUI, CKEditor: declarations written after a nested rule
are moved above it (#45462).
- antd 4, purecss, govuk-frontend: browser hacks in selector lists start
to apply everywhere (#42100).
- foundation-sites, flatpickr, 98.css, blueprintjs, scalar,
react-datepicker: the build fails on valid CSS (#46756, #38552, #38534).
tachyons, handsontable: it fails on the IE star hack (#46751).
- nodejs.org, Mantine: a CSS module exports classes that no rule
defines, because the `@apply` or `@mixin` in them is moved out of the
rule (#46759).
- hint.css and 42 more: `z-index: 1000001` and `1000000` become the same
number (#46715).
- Adobe Spectrum: `--version: 5.0.1` prints as `5.1`.
- A quarter of all stylesheets: minified output has a stray space in
`align-items: center` (#41993).

#### Behavior changes to look at first

1. **`light-dark()` is kept as written at the default browser targets**
(#42009). The lowered form `var(--buncss-light, a) var(--buncss-dark,
b)` is invalid unless a compiled `color-scheme` covers the element: 184
of the 215 real stylesheets that get it never define the variable, so
every browser drops the declaration. The docs section on the polyfill is
removed.
2. **`border` is not built from longhands** (#42072):
`border-width:0;border-style:solid;border-color:red` stays three
declarations, because `border` also resets `border-image`. 305 real
stylesheets grow by 30 KB in total; `bootstrap.min.css` by 113 bytes, 5
gzipped. No real stylesheet was found where the old merge shows, so this
is the one to drop if size matters more. 14 existing expectations
change.
3. **A CSS module no longer exports names that only occur inside
`:global(...)`** (#45487). 73 real modules lose 202 export keys, all of
them names that no rule has. `import { g }` or `composes: g` of such a
name becomes a build error.
4. **Container names and `@counter-style` names stay global in a CSS
module** (#45247 by @mikamikasuki). `main` hashes the `@container` name
but not `container-name`, so a named query never matches.
postcss-modules (Vite, css-loader, Next.js) keeps them global;
lightningcss and esbuild hash both. Global keeps a container that is
queried from another file or from JS working.
5. **Syntax errors are build errors, also two that `main` lets
through**: a media query that stops short (`@media screen and {`,
`@media not {`, `@import "a.css" screen and;`) and `@media (width <) foo
{`. No stylesheet of the 10,116 has either. A condition that is valid by
the grammar though no browser knows it, `( ... )` or `fn( ... )`, is
kept as written (#46756): `(min-width:0\0)`, `(not(hover))`.
6. **Declarations after a nested rule stay where they were written**, as
a nested declarations rule (#45462). Changes the output of 199 real
stylesheets; in a fuzz of 2,850 nested programs compared by computed
style, `main` was wrong in 1,056 and the pull request's code, before it
was trimmed here, in none.
7. **`Bun.color`**: `color()` inputs are converted instead of returned
as a string (`Bun.color("color(srgb 1 0 0)", "hex")` was `"color(srgb 1
0 0)"`); `currentColor`, system colors and `light-dark()` give `null` in
the converting formats instead of the keyword as a string; a relative
color or `color-mix()` outside sRGB is clipped, not gamut-mapped
(`hsl(from oklch(100% .399 336.3) h s l)` is `#ff5bff`, was `#fff`).
8. The `:-moz-any()` and `:-webkit-any()` copies of rules with `:is()`
are gone at the default targets (#45945, #42029). The `:-webkit-any()`
copy outranked the original rule: `.title{color:blue}
:is(h1,h2,h3){color:red}` painted red. 5.8% of all default-target output
was the `:-moz-any()` copies.
9. `Property` grows from 216 to 224 bytes: `Url` stores its text as a
slice so that two `url()` values can be compared (#46895). The pull
request packed a pointer and a `u32` behind `unsafe` to avoid that.

#### Not in this PR, needs a decision

- **Raise the default browser targets (#40369) and add a `cssTarget`
option (#40368).** Today: Chrome 80, Edge 80, Firefox 78, Safari 14,
Opera 67, and no public way to change them. At Vite 8's floor the output
of the 9,924 stylesheets is 15.3% smaller (3.9% gzipped). About a dozen
fixes here only matter at today's floor, and code for older targets is
unreachable without the option.
- Vendor prefixes for `user-select`, `appearance`, `hyphens` and five
more (#45814): 1,116 real stylesheets lack a prefix they need at the
default targets. And for `backdrop-filter` (#44648): 173.
- Clip instead of gamut-map in the sRGB fallback of a plain wide-gamut
color: deletes about 90 lines, changes every pinned fallback hex. If the
mapping stays, #44647 by @oddharsh makes it closer to what browsers
paint (colors off by 2/255 or more: 168 to 110 of 3,170 real ones).

#### Changes

**Rules: duplicates, merging, nesting, fallbacks**

- The `:-moz-any()` copy of every rule with `:is()` or `:not(a, b)` is
gone: Firefox 78 has `:is()` (#45945). 1,164 real stylesheets
- Keep declarations written after a nested rule in place with a nested
declarations rule (#45462). 379 real stylesheets
- With nesting compiled away, an unknown at-rule inside a style rule was
moved out of it: `.wrapper{@apply flex}` printed a bare `@apply flex;`
and the rule was gone, so a CSS module exported a class that no rule
defines. It now stays in a block under the compiled selector. 72 real
modules with Tailwind's `@apply`, 25 with Mantine's `@mixin` (#46759).
316 real stylesheets
- Keep !important in the rules generated for logical properties and
color-scheme (#38597). 311 real stylesheets
- With browser targets, remove an earlier duplicate rule only when its
declaration block is equal (#46861). 171 real stylesheets
- Compare custom property names in `PropertyId` equality and importance
in duplicate-rule detection (#33684). 40 real stylesheets
- Keep same-condition @supports rules and empty named @layer blocks
(#46733). 31 real stylesheets
- Keep target-added vendor prefix passes on cloned and re-minified
rules, and never print a bare `:full-screen` (#43339). 12 real
stylesheets
- Keep the selectors on rules added for color fallbacks and logical
properties (#43240). 3 real stylesheets
- A later declaration of a property drops the `@supports` color copies
of the earlier one, which printed after the rule and won (#45560). 2
real stylesheets
- Do not merge style rules in a stylesheet that holds `revert-rule`
(#46953). 1 real stylesheets
- `@import ... supports(-webkit-mask: none)` kept its prefix (#38589)

**Properties and shorthands**

- Print align-items:center without a stray space when no safe/unsafe
keyword is given (#41993). 3,317 real stylesheets
- Emit the physical longhands instead of the inset shorthand for targets
without it (#38583). 1,452 real stylesheets
- Keep the stops of an upward gradient as written, and drop a `50%` hint
only at the midpoint (#46790). 104 real stylesheets
- Emit unparsed border-block-* declarations when compiling logical
borders (#38582). 89 real stylesheets
- Keep repeated font families and dedupe only the inserted system-ui
fallbacks (#46662). 88 real stylesheets
- A `width` / `height` / `min-*` / `max-*` with `var()` or `env()` was
printed before the plain fallback written above it, so the fallback won
(#42101). 26 real stylesheets
- Keep box-orient and box-direction when flex-direction or flex-flow
follows (#45672). 16 real stylesheets
- Write held prefixed background declarations at every flush (#46882). 6
real stylesheets
- Derive equality for Size2D and BorderRadius, and keep every prefix in
the border-radius handler (#45665). 2 real stylesheets
- Print `none` instead of an empty border-image, mask-border or
mask-box-image value (#45143). 2 real stylesheets
- Keep font/transition sub-properties on the side of the shorthand they
were written on (#41990). 2 real stylesheets
- CSS-wide keywords are not names: `font:12px Arial;font-family:inherit`
printed `font:12px inherit`, `color-scheme:inherit` and
`color-scheme:auto` printed `normal`, `transition-property:unset` was
folded into `transition` (#42022). 2 real stylesheets
- Only fold font longhands into `font` when the rule declared a `font`
(#45459)
- Keep a font family name quoted when its identifiers do not read back
as the same name (#46606)
- Stop rewriting valid positions into ones browsers reject (#45197)
- `border` is no longer built from longhands unless the rule also sets
what `border` resets (`border-image`), and an author's `border-image`
reset is no longer dropped. `transition` is likewise only written where
the rule wrote one, because it resets `transition-behavior` (#42072)
- `border; border-image: <gradient>; border-image-slice: 1` lost the
gradient in every default build (#42018)
- A prefixed `linear-gradient` keeps its direction, written or
generated: the prefixed syntax names the starting side and measures
angles differently (#46697)

**At-rules**

- Print @font-face format() keywords as strings (#37979). 439 real
stylesheets
- Print @supports `not` conditions without a leading space (#38559). 88
real stylesheets
- Reset after the number token of a unicode-range (#45835). 33 real
stylesheets
- Keep `<general-enclosed>` media and container conditions as written
(#46756). 24 real stylesheets
- A strict range such as `(width < 768px)` lowered for old targets still
matched at exactly 768px. The step is now `.02px`. The `not all and`
form for other units is not taken (#46806). 18 real stylesheets
- Keep scanning an @supports condition after a nested block (#38552). 9
real stylesheets
- Every `@counter-style name` printed as `@counter-stylename`, which
browsers drop (also #46692). In a CSS module the name now stays global,
like the `list-style` that refers to it (#33679). 9 real stylesheets
- Keep a keyframe whose block has an invalid item (#46288). 2 real
stylesheets
- Read an @Property initial-value on the stylesheet's parser, keep an
unmatched value as written (#45806). 2 real stylesheets
- `@media {}` is `all`. A media query that stops after `and`, `or` or
`not` (`@media screen and {`) is a build error; `main` printed `not all`
or dropped the condition, silently. The `@layer` and type selector parts
are not taken (#46869). 2 real stylesheets
- Accept a name-only `@container` and a `style()` feature with no value
(#46689). 1 real stylesheets
- Parse `<timeline-range-name> <percentage>` keyframe selectors (#45652)
- Allow `or` inside parentheses after a media type (#45088)
- Print the `=` comparison in media and container queries as `=`
(#45071)
- Parse the @font-palette-values font-family descriptor as a list of
family names (#45484)

**Selectors**

- Never wrap a selector list in :is() (#46787). 414 real stylesheets
- Stop emitting :-webkit-any() copies of :is() that outrank the original
rule (#42029). 373 real stylesheets
- With browser targets, a selector list with a vendor-prefixed member
was split, which revives hacks that rely on other engines dropping the
whole list: `.s *::-ms-backdrop,.s{flex:auto}` applied everywhere.
Rewritten; the list is printed as written (#42100). 42 real stylesheets
- Print the vendor prefix on downleveled :any-link, :read-only,
:read-write, :placeholder-shown and :autofill (#42131). 12 real
stylesheets
- Fix nesting lowering of a `&` that does not lead its selector
(#43047). 8 real stylesheets
- `@scope`: `&` and implicit selectors compile to `:where(:scope)`,
relative selectors and bare declarations parse (#43165). 7 real
stylesheets
- Only reject bare :global / :local pseudo-classes in CSS modules
(#38534). 5 real stylesheets
- Stop truncating `@scope to (...)` and stop replacing `&` in the
scope-end selector (#43146). 2 real stylesheets
- `[xl|href$=".pdf"]` lost its namespace prefix (#45159). 2 real
stylesheets
- Match An+B keywords exactly and take only `+` as the leading sign
(#42953). 1 real stylesheets
- Parse ::details-content and ::picker() like ::part() (#45504). 1 real
stylesheets
- Parse :window-inactive as an ordinary pseudo-class, and accept it
after ::selection (#45815)
- Print :nth-col(1) and :nth-last-col(1) in functional form (#41575)
- `li::before::marker` failed the build (#45810)
- `&` inside `:nth-child(... of S)` is resolved when nesting is compiled
away. The second selector serializer (326 lines) is deleted (#43024)

**Colors and `Bun.color`**

- Keep light-dark() as written under the default browser targets
(#42009). 227 real stylesheets
- Number and percentage channels parse in `lab()`, `lch()`, `oklab()`,
`oklch()`, `hsl()` and `hwb()`; the hue of `lch` / `oklch` is no longer
scaled by alpha (#45084). 130 real stylesheets
- `Bun.color` returns `null` for `currentColor`, system colors and
`light-dark()` in the converting formats, instead of the CSS keyword as
a string (#33047). 130 real stylesheets
- Relative colors: channel keywords are on the scale of their function
(`rgb(from #369 calc(r + 10) g b)`), the origin's alpha is kept, a
relative color kept as written gets no `@supports` copies, and a result
outside sRGB is not folded (#45362). 31 real stylesheets
- Keep rotate: 0deg out of none, and keep the origin alpha in relative
colors (#38486). 13 real stylesheets
- Compute the D50 reference white in f32 so greys convert to lab(l 0 0)
like lightningcss (#38535). 7 real stylesheets
- Return a missing hue for a grey from conversions into lch, oklch, hsl
and hwb (#45201). 5 real stylesheets
- Fix color(a98-rgb), srgb-linear relative colors and color-mix(in
xyz-d50) (#38487). 4 real stylesheets
- Match named colors without regard to ASCII case (#45092). 3 real
stylesheets
- Fold min(), max() and clamp() over plain numbers (#39515)
- `calc(l - .08)` in a percentage channel of a relative color became NaN
and printed `#000` (also #45970) (#45971)
- Bun.color: convert color() inputs instead of returning them as a CSS
string (#38488)
- A `color-mix(in srgb, ...)` result outside sRGB prints `color(srgb
...)` instead of a gamut-mapped hex (#39269)

**`calc()` and numbers**

- Rounding to 6 significant digits ate into the integer part:
`z-index:2147483647` became `2147480000`, `max-width:9999999px` became
`10000000px`, `--x:1234567.5` became `1234570`. It now only cuts a
fraction short: a value kept as tokens prints its source text, a typed
value prints every integer digit and at least one digit of the fraction
(#46715). 43 real stylesheets
- A number without a unit is no longer repaired into `px`
(`height:1.5715` is dropped by a standards-mode browser, so it prints as
written), `calc(-1 * 0)` is not a length, and on a property of the range
[0,∞] (sizes, border radii, padding, border and outline widths, gap,
`flex-basis`, `font-size`) a `calc()` that folds below zero prints `0`
instead of an invalid negative literal, and a negative literal no longer
replaces the valid declaration before it (`width:10px;width:-5px`
printed `width:-5px`) (#42104). 24 real stylesheets
- Print numbers from the f32's shortest digits, not its f64 expansion
(#38565). 17 real stylesheets
- Print negative numbers below 1 with a single minus sign (#38516)
- Keep math functions over two different length units as written
(#38489)
- Fold rem() with a truncated quotient, so it keeps the sign of the
dividend (#32286)
- A one-argument `min()` / `max()` around a sum printed without any
function (`width:100% - 20px`) (#46145)
- `round()` breaks ties upward and handles a negative step as browsers
do (#45249)

**CSS modules**

- A CSS module exports its `@keyframes` names (`styles.spin`). Hashing
is unchanged (#46720). 145 real stylesheets
- A name inside `:global(...)` prints plain but was exported as
`name_<hash>`, a class no rule has (also #38573) (#45487). 90 real
stylesheets
- Keep container names global in CSS modules (#45247). 6 real
stylesheets
- Tag the keyframes name when an `animation` value falls back to raw
tokens (#46034). 6 real stylesheets
- In a plain stylesheet `composes: a from "./b.css"` was treated as an
import: the file was inlined and an absolute path printed (#45961). 4
real stylesheets
- Drop a rejected composes instead of failing the file at print time
(#38538). 1 real stylesheets
- Keep every rule's custom properties for the composes conflict check
(#38523)
- Keep composes working after a nested rule in the same block (#38520)
- `IdentOrRef::hash` read 2 of 16 bytes, so in release builds every
class of a module hashed alike (#30784)

**Tokens, `url()`, diagnostics**

- The block skip stack no longer allocates on every block exit (#39533)
- Report the diagnostic's line without its line breaks in
position.lineText (#38342). 192 real stylesheets
- Compare `url()` values by their text (#46895). 26 real stylesheets
- Values kept as tokens keep the blanks they were written with instead
of having them re-invented at print time: `attr(x type(<number>))`
printed `type(< number> )`, which browsers drop. Minified output still
drops the blanks that cannot matter (#45577). 19 real stylesheets
- The IE star hack (`*zoom:1`) failed the whole build. The declaration
is skipped, as every browser does. It is the one syntax error that
builds: tachyons, handsontable and jasmine ship it (#46751). 8 real
stylesheets
- The path of a copied asset spliced into `url()` is always quoted, so a
file name with a space or a parenthesis cannot end the `url()` (#41557).
7 real stylesheets
- Move the skip loop of parse_until_before out of the generic body
(#39520)
- Remove dead code from bun_css: the unused `impl Default for Property`
(#45618)
- Default the at-rule and qualified-rule parser traits, drop the
url_to_css shim (#31999)

**Speed**

Four inputs that took quadratic or exponential time, or produced output
out of proportion, are now linear, with no new limit and no extra pass
(#31639, #42243, #38531, #43421).

**Tests only**

- `css-fuzz.test.ts` never ran in CI and timed out on debug builds; it
is a deterministic table now (#38500)
- `nth-anplusb-ident.test.ts` counts heaps in a child process (#46331)
- `color.test.ts` checks the ansi-16 sweep with one `expect()` (#42824)
- Not CSS, but it failed this PR's CI and fails on `main` too:
`ws-proxy.test.ts` on Windows, "a message sent during a TLS
renegotiation arrives after it". The helper server now renegotiates from
the callback of the frame's write (#46170)

**No pull request covered these**

- `--v: 5.0.1` printed as `5.1`, `--w: 1.0.0` as `10`, `--x: 1+2` as
`12`, `2n+1` as `2n1`: a number token that touches another keeps its
source text.
- `color-scheme: auto` printed as `normal`, and `color-scheme: dark
light` as `light dark`.
- A bare `:global` or `:local` in a CSS module sets the scope of the
names after it, up to the end of the selector, as in css-loader. It was
a build error.
- The scrollbar states (`a:horizontal`,
`::-webkit-scrollbar{&:horizontal{}}`) are valid in a plain compound, as
in WebKit and Blink.
- Without browser targets, `-webkit-box-align`, `-ms-flex-align` and the
like were dropped next to their standard form (flatpickr,
foundation-sites, materialize-css: 28 stylesheets), and
`-webkit-fill-available` before `-moz-available` (beercss).
- `(-o-min-device-pixel-ratio: 2/1)` printed as `2`.
- `sin()`, `cos()` and `tan()` are exact at multiples of 90deg:
`calc(100px * cos(90deg))` printed `-.00000437114px`.
- A `display-p3` color at targets that have P3 but not the other
`color()` spaces (Safari 10.1 to 14) got itself as a fallback.
- `width: rem(-18px, 5px)` folds to a negative literal once `rem()` is
right; it prints `0`.
- Consecutive nested declarations rules, which arise when same-query
`@media` blocks in one rule merge, are folded into one and count against
the existing selector expansion limit as the `&{}` rules they replace
did.

#### Known and not fixed here

- Four longhands are still folded into a `transition` the author did not
write where the targets add a prefixed copy of a property (`mask-image`
at Chrome 80): only the shorthand gives the copy its own times.
- A later `background`, `box-shadow` or `border-*` declaration still
loses to the `@supports` copy of an earlier `var()` value of the same
property with a wide-gamut color. Same on `main`; #45560 is taken for
the properties of the fallback handler and for unparsed values only.
- A merged rule prints the unprefixed name in the passes of the rule it
absorbed: `.d::-webkit-input-placeholder{}.d::-moz-placeholder{}` prints
`::placeholder` for the second (#42106, not taken: it breaks the merge
of prefix-equivalent rules).
- The `border`, `outline` and `font` shorthands, `line-height`,
`scroll-padding`, `background-size` and a few more still print a
`calc()` that folds below zero as a negative literal. No real stylesheet
has one.
- `all` is not a property Bun knows, so `a{margin-top:1px;all:unset}`
prints `all` first. Same on `main`.
- A selector kept as raw tokens loses the blank before `*` when
minified: `--sel: .dark *` prints `.dark*`. Same on `main`.
- `width:calc(0% + 12px)` folds to `12px`, which lays out differently in
a table cell. Same on `main`.
- The `:dir()` / `:lang()` rules of a lowered logical property print
before the nested rules of their rule, so a logical declaration written
after a nested rule of equal specificity still loses to it. Same on
`main`.
- A fraction is still rounded to 6 significant digits (`12345.678`
prints `12345.7`, `33.3333333333%` prints `33.3333%`). Keeping every
digit would add up to 282 KB over the corpus.
- `@keyframes "quoted"` is hashed in a module but not exported.
- In a module, a keyframes name inside a `var()` fallback is not hashed:
`animation: var(--enter, fade 1s)`. Same on `main`.
- A `color-mix()` or relative color folded to a hex inside a gradient
changes the interpolation from Oklab to sRGB (VS Code, clawhub).
Printing `color(srgb ..)` instead needs a fallback for `mask-image`,
`list-style-image` and custom properties at the default targets, which
have none.
- A comma list in `@container` still fails the build.
- After a bare `:global`, names that are hashed at print time
(`::view-transition-old(foo)`) are still hashed; inside `:global(...)`
they are not.

#### Pull requests in scope that do not land here

- Duplicates of a pull request that lands here: #36961 (of #37979),
#38478 (of #46662), #38480 (of #46733), #38482 (of #46715), #38491 (of
#38500), #38508 (of #45201), #38513 (of #45362), #38573 (of #45487),
#38640 (of #38500), #45525 (of #45247), #45589 (of #46756), #45970 (of
#45971)
- Already fixed on `main`: #38674
- Right, but for input nobody writes (0 of 10,116 stylesheets, and not
new syntax): #31916, #38526, #38592, #39349, #41139, #41801, #42042,
#42484, #43078, #45604, #45908, #46086, #46221, #46224, #46273, #46694,
#46707, #46737, #46755, #46816, #46870, #46934
- Right or harmless, but not worth their code, their bigger output or
the behavior they change: #30905, #31924, #33332, #34426, #36225,
#38006, #38512, #38547, #38551, #38561, #38572, #38576, #38607, #38623,
#38667, #38692, #42106, #42912, #42918, #43407, #43434, #44700, #45442,
#45546, #45747, #45759, #45825, #45946, #46061, #46127, #46206
- Need a decision (see above), stay open: #40368, #40369, #44647,
#44648, #45814
- Stays open: #45787

Out of scope: CSS linking and chunking in the bundler, the dev server,
docs-only pull requests, and repository-wide sweeps that touch `src/css`
in passing.

Thanks to @mikamikasuki (#45247), @alii (#39520, #31999) and @oddharsh,
whose measurements on #40369 and #41990 were used.

Fixes #16727.
Fixes #24770.
Fixes #30772.
Fixes #44704.
Fixes #45089.
Fixes #45237.
Fixes #46872.

### How did you verify your code works?

- The 35 CSS test files pass on a debug build, also with
`FORCE_COLOR=1`. 600+ of the tests in this branch fail on a build of
`main`. An earlier head of this branch passed all 179 CI jobs that ran.
- The reproduction of each issue listed above was run on `main` (wrong)
and on this branch (right).
- Differential run over the 10,116 stylesheets, `main` against this
branch, minified without targets and through `Bun.build` at the default
targets: no stylesheet that builds on `main` fails; Chromium 156
compares the output with the input rule by rule (CSSOM, then computed
values on a sized element in a standards-mode document, colors as
painted pixels in sRGB and P3). The comparison has 195 control pairs
that must read equal or different. For stylesheets that read differently
on both builds, mostly because of lowering, the sets of differences were
compared; the three new ones are two minifier fixtures that are wrong on
both and one comparison artifact.
- CSS modules are checked for consistency instead, since hashed names
defeat the comparison: every reference points at a name the output
defines, every export exists.
- `Bun.color`: 59 inputs in 13 formats on both builds; every changed
cell was read.
alii added a commit that referenced this pull request Nov 17, 2026
### What does this PR do?

Each of the 251 CSS properties was written out by hand in 3 enums, a
name map and up to 9 matches, 4,994 lines. Now it is one row:

```rust
"flex-basis": FlexBasis(LengthPercentageOrAuto, VendorPrefix: WEBKIT) with parse_non_negative,
```

- `define_properties!` in `property_table.rs` generates `PropertyIdTag`,
`PropertyId`, `Property` and their matches. Variant order and payloads
are the same.
- The bundler reads `PropertyIdTag::VARIANTS[i]` where it used a
`transmute`. One `unsafe` block goes.
- Deleted: `properties_generated.rs`, `properties_impl.rs` (now inherent
methods), `shape.rs` (unused), `__to_css_helper`, `PREFIX_FLAGS`,
`dc::property`.
- `src`: +692, -5,179.

#41268 also deletes `shape.rs`. #44648, #45814, #47441, #33332, #40438
and #49095 edit the old files and need a rebase.

**Bugs this fixes**

None known.

**Visible changes**

None.

**Kept as is**

- `all` is never parsed, so `a{margin:1px;all:unset}` prints `all` first
(#42042, closed). `eql` of two `All` ignores the keyword.
- `transform-origin`, `perspective-origin` and `mask-position` have the
type of the `position` property, so `left top` stays unparsed.
- `-webkit-mask-composite` never gives its variant, because the prefix
is stripped before the lookup.
- Names that only exist with a prefix, such as `box-orient`, are known
without it too.
- `line-height`, `border`, `font`, `background-size`, `flex-grow` and
some others accept a negative value, which replaces the valid one before
it.
- `PropertyId::Unparsed` is never built. It keeps `tag()` arithmetic.

**Performance**

| instructions, parse + minify + print | before | after | |
| --- | --- | --- | --- |
| bootstrap, tailwind, foundation (Chrome 80, Safari 14) | 128,074,410 |
128,018,694 | -0.04% |
| animate, materialize, pico (Chrome 40, IE 10) | 92,306,972 |
92,251,855 | -0.06% |
| uikit, not minified | 72,133,465 | 72,121,134 | -0.02% |
| `bun build --minify bulma.css`, whole process (noise 0.5%) |
205,648,998 | 205,298,272 | within noise |
| release binary, bytes | 174,344,992 | 174,344,992 | 0 |
| CI size: linux-x64, darwin-aarch64, windows-x64 | | | +0.0, +0.0, +0.5
KB |

A declaration that stays unparsed pays one more compare, for the
`animation` fallback.

### How did you verify your code works?

On Linux x64:

- Compared the expanded macro with the old source, arm by arm. The
enums, the keys and values of the name map and the other matches are the
same. `prefix_feature` has 23 explicit `None` arms that `_` covered. In
`parse`, the fallback of `animation` and `animation-name` moved behind
the match, and the empty `composes` arm became `leave_unparsed`.
- Ran 260 names x 8 prefixes x 236 values x up to 5 target sets through
both release builds: declarations, merges, `transition`, `@supports`,
`@container style()`, style attributes, CSS modules. 147,846 outputs,
all equal.
- Bundled 429 stylesheets (39 tracked, 390 from `node_modules`) 10 ways
each. 4,290 builds, none differ. 108 fail the same way on both.
- Ran 53 test files on debug + ASAN: `test/js/bun/css`,
`test/bundler/css`, the esbuild CSS tests, HTML bundling, `bun pm diff`,
CSS regressions. 51 pass. `test/bake/dev/{css,html}` fail the same tests
on both sides: `happy-dom` was not installed.
- Added 8 cases for prefix lookup, `composes` errors and the order of
conflicts. They pin behaviour, so they pass on main too.
- 140,000 random stylesheets with random targets and 59,000 odd names
and bundler cases give the same output on both release builds.
- CI ran the whole suite on the first commit, except the macOS x64
tests, which were still queued. One `node:http` AOT test failed on macOS
arm64; it has no CSS in it. The sizes are from that build, and 40 of the
local test files predate the last 20 line edit.

This branch has not been deployed

No deployments
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.

2 participants