fix(parse/html): stop reading {{ as an interpolation in Svelte

[dyc3]

Summary

Fixes the parsing of {{ in svelte files. It's a bit of a larger refactor than I'd like, but it makes the most sense to me.

implemented by opus 5 initially, gpt 5.6 sol fixed it up

Test Plan

snapshots

Docs

---

_{Stack created with GitHub Stacks CLI • Give Feedback 💬}

[changeset-bot]

🦋 Changeset detected

Latest commit: 6395bae

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 13 packages

Name	Type
@biomejs/biome	Patch
@biomejs/cli-win32-x64	Patch
@biomejs/cli-win32-arm64	Patch
@biomejs/cli-darwin-x64	Patch
@biomejs/cli-darwin-arm64	Patch
@biomejs/cli-linux-x64	Patch
@biomejs/cli-linux-arm64	Patch
@biomejs/cli-linux-x64-musl	Patch
@biomejs/cli-linux-arm64-musl	Patch
@biomejs/wasm-web	Patch
@biomejs/wasm-bundler	Patch
@biomejs/wasm-nodejs	Patch
@biomejs/backend-jsonrpc	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

✅ Organic activity

No automation signals detected in the analyzed events.

View full analysis →

_{This is an automated analysis by AgentScan}

[codspeed-hq]

Merging this PR will improve performance by 6.44%

⚡ 3 improved benchmarks
✅ 64 untouched benchmarks
⏩ 219 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	synthetic/vue-expressions.vue[cached]	387.9 µs	361.3 µs	+7.36%
⚡	synthetic/vue-expressions.vue[uncached]	433 µs	406.5 µs	+6.52%
⚡	synthetic/svelte-expressions.svelte[cached]	477.1 µs	452.5 µs	+5.44%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing dyc3/svelte-double-curly (6395bae) with main (6556ca3)}

Footnotes

1. 219 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[ematipico]

Left some comment to address

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 03d7d804-8e8a-48f1-b39a-565a4ab18a76

📥 Commits

Reviewing files that changed from the base of the PR and between 408ad70 and 6395bae.

⛔ Files ignored due to path filters (2)

• crates/biome_html_formatter/tests/specs/html/svelte/double_curly.svelte.snap is excluded by !**/*.snap and included by **
• crates/biome_html_parser/tests/html_specs/ok/svelte/double_curly.svelte.snap is excluded by !**/*.snap and included by **

📒 Files selected for processing (4)

• crates/biome_html_formatter/tests/specs/html/svelte/double_curly.svelte
• crates/biome_html_parser/src/syntax/mod.rs
• crates/biome_html_parser/src/syntax/svelte.rs
• crates/biome_html_parser/tests/html_specs/ok/svelte/double_curly.svelte

🚧 Files skipped from review as they are similar to previous changes (4)

• crates/biome_html_formatter/tests/specs/html/svelte/double_curly.svelte
• crates/biome_html_parser/src/syntax/svelte.rs
• crates/biome_html_parser/src/syntax/mod.rs
• crates/biome_html_parser/tests/html_specs/ok/svelte/double_curly.svelte

---

Walkthrough

The HTML parser now derives HtmlFramework from parser options and carries it through lexer contexts and re-lexing. Double-curly recognition is enabled for non-Svelte frameworks and disabled in Svelte-specific paths. HTML and Svelte parsing now use framework-aware contexts for expressions, literals, delimiters, and blocks. Lexer, parser, formatter, and Svelte fixtures cover the updated behaviour. A patch changeset documents the fix.

Possibly related PRs

• biomejs/biome#10665: Updates framework-aware HtmlLexContext handling in related HTML lexer and parser paths.
• biomejs/biome#10275: Changes Svelte text-expression boundary handling in the lexer and syntax parser.
• biomejs/biome#11056: Updates Svelte handling of {...} expressions through related context-sensitive parsing paths.

Suggested reviewers: ematipico

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly describes the main change: preventing Svelte parsing from treating {{ as interpolation.
Description check	✅ Passed	The description explains the Svelte parsing fix and mentions the parser refactor and snapshot tests.
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.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch dyc3/svelte-double-curly

---

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

[coderabbitai]

Actionable comments posted: 1

🤖 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 `@crates/biome_html_parser/src/token_source.rs`:
- Around line 348-353: Update the token-advancing methods around
bump_with_context and skip_as_trivia so they preserve the active Svelte
framework instead of constructing HtmlLexContext::default(). Reuse the current
framework when creating the context, and ensure parse_html_element fallbacks and
Svelte paths continue lexing {{ as two L_CURLY tokens.

🪄 Autofix

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: CHILL

Plan: Pro

Run ID: 7b7b80cd-9829-449a-abbe-28180b708248

📥 Commits

Reviewing files that changed from the base of the PR and between 6556ca3 and 408ad70.

⛔ Files ignored due to path filters (3)

• crates/biome_html_formatter/tests/specs/html/svelte/double_curly.svelte.snap is excluded by !**/*.snap and included by **
• crates/biome_html_formatter/tests/specs/svelte-plugin/formatting/pre-mustache.svelte.snap is excluded by !**/*.snap and included by **
• crates/biome_html_parser/tests/html_specs/ok/svelte/double_curly.svelte.snap is excluded by !**/*.snap and included by **

📒 Files selected for processing (9)

• .changeset/fix-svelte-double-curly.md
• crates/biome_html_formatter/tests/specs/html/svelte/double_curly.svelte
• crates/biome_html_parser/src/lexer/mod.rs
• crates/biome_html_parser/src/lexer/tests.rs
• crates/biome_html_parser/src/parser.rs
• crates/biome_html_parser/src/syntax/mod.rs
• crates/biome_html_parser/src/syntax/svelte.rs
• crates/biome_html_parser/src/token_source.rs
• crates/biome_html_parser/tests/html_specs/ok/svelte/double_curly.svelte

[coderabbitai]

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
set -euo pipefail

rg -n -C 4 \
  '\bp\.(bump|bump_any|bump_remap|skip_as_trivia)\b' \
  crates/biome_html_parser/src/syntax/mod.rs \
  crates/biome_html_parser/src/syntax/svelte.rs

rg -n -C 5 \
  'fn (bump_remap|bump_any|skip_as_trivia)\b' \
  crates --glob '*.rs'

Repository: biomejs/biome

Length of output: 16409

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

printf '%s\n' '--- HTML token source structure ---'
ast-grep outline crates/biome_html_parser/src/token_source.rs --match 'struct HtmlTokenSource' --view expanded
ast-grep outline crates/biome_html_parser/src/token_source.rs --match 'impl TokenSource for HtmlTokenSource' --view expanded
printf '%s\n' '--- HTML token source implementation ---'
sed -n '1,390p' crates/biome_html_parser/src/token_source.rs

printf '%s\n' '--- parser bump implementation ---'
sed -n '250,350p' crates/biome_parser/src/lib.rs

printf '%s\n' '--- HTML context definitions and calls ---'
rg -n -C 6 'enum HtmlLexContext|struct HtmlLexContext|impl Default for HtmlLexContext|fn bump_with_context|fn re_lex|regular_context|HtmlReLexContext' \
  crates/biome_html_parser crates/biome_parser --glob '*.rs'

Repository: biomejs/biome

Length of output: 50369

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

printf '%s\n' '--- parser do_bump paths ---'
sed -n '340,390p' crates/biome_parser/src/lib.rs

printf '%s\n' '--- lexer context-sensitive text handling ---'
rg -n -C 12 'fn consume_html_text|L_DOUBLE_CURLY|framework != HtmlFramework::Svelte|HtmlFramework::Svelte' \
  crates/biome_html_parser/src/lexer/mod.rs

printf '%s\n' '--- all HTML parser default bump calls ---'
rg -n -C 3 '\bp\.(bump|bump_any|bump_remap|skip_as_trivia)\b' \
  crates/biome_html_parser/src --glob '*.rs'

printf '%s\n' '--- Svelte parser context transitions ---'
sed -n '285,370p' crates/biome_html_parser/src/syntax/svelte.rs
sed -n '1160,1220p' crates/biome_html_parser/src/syntax/mod.rs

Repository: biomejs/biome

Length of output: 16362

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

printf '%s\n' '--- framework selection ---'
sed -n '150,225p' crates/biome_html_parser/src/syntax/mod.rs
rg -n -C 5 'HtmlFileSource|HtmlVariant|HtmlFramework::Svelte|initial_context|from_str\(' \
  crates/biome_html_parser/src/parser.rs crates/biome_html_parser/src --glob '*.rs'

printf '%s\n' '--- default-bump call-site bodies ---'
sed -n '500,575p' crates/biome_html_parser/src/syntax/mod.rs
sed -n '1145,1220p' crates/biome_html_parser/src/syntax/mod.rs
sed -n '470,535p' crates/biome_html_parser/src/syntax/svelte.rs
sed -n '750,805p' crates/biome_html_parser/src/syntax/svelte.rs

printf '%s\n' '--- relevant lexer branches and tests ---'
sed -n '1550,1595p' crates/biome_html_parser/src/lexer/mod.rs
sed -n '395,425p' crates/biome_html_parser/src/lexer/tests.rs
rg -n -C 3 'bump_remap_with_context|bump_any_with_context|skip_as_trivia_with_context|expect_with_context' \
  crates/biome_html_parser/src --glob '*.rs'

Repository: biomejs/biome

Length of output: 50369

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

printf '%s\n' '--- Svelte declaration/expression fallback ---'
rg -n -C 8 'fn parse_svelte_declaration_or_expression|parse_svelte_declaration_or_expression' \
  crates/biome_html_parser/src/syntax --glob '*.rs'

printf '%s\n' '--- exact default calls in Svelte-specific code ---'
rg -n -C 2 '\bp\.(bump|bump_any|bump_remap|skip_as_trivia)\b' \
  crates/biome_html_parser/src/syntax/svelte.rs

printf '%s\n' '--- exact default calls in shared HTML parsing ---'
rg -n -C 2 '\bp\.(bump|bump_any|bump_remap|skip_as_trivia)\b' \
  crates/biome_html_parser/src/syntax/mod.rs

printf '%s\n' '--- read-only source invariant probe ---'
python3 - <<'PY'
from pathlib import Path

token_source = Path("crates/biome_html_parser/src/token_source.rs").read_text()
lexer = Path("crates/biome_html_parser/src/lexer/mod.rs").read_text()
svelte = Path("crates/biome_html_parser/src/syntax/svelte.rs").read_text()
syntax = Path("crates/biome_html_parser/src/syntax/mod.rs").read_text()

assert "fn bump(&mut self) {\n        self.bump_with_context(HtmlLexContext::default())" in token_source
assert "fn skip_as_trivia(&mut self) {\n        self.skip_as_trivia_with_context(HtmlLexContext::default())" in token_source
assert "self.consume_token(current, framework != HtmlFramework::Svelte)" in lexer
assert "HtmlLexContext::Regular { framework: HtmlFramework::Svelte }" in Path(
    "crates/biome_html_parser/src/lexer/tests.rs"
).read_text()

for name, source in [("svelte.rs", svelte), ("mod.rs", syntax)]:
    calls = []
    for method in ("bump", "bump_any", "bump_remap", "skip_as_trivia"):
        calls.extend((i + 1, method) for i, line in enumerate(source.splitlines())
                     if f"p.{method}(" in line and f"p.{method}_with_context(" not in line)
    print(name, sorted(calls))

print("source invariant: default TokenSource bumps use HtmlLexContext::default()")
print("source invariant: HtmlLexContext::default() is Regular/Plain")
print("source invariant: Regular/Svelte disables double-curly lexing")
print("source invariant: Svelte-specific code contains default bump calls")
PY

Repository: biomejs/biome

Length of output: 7057

---

Preserve the Svelte framework when advancing tokens.

Default token advances use Regular { framework: Plain }. The Svelte paths and parse_html_element fallbacks therefore can lex a following {{ as L_DOUBLE_CURLY instead of two L_CURLY tokens. Store the active framework or replace these calls with the appropriate *_with_context variant.

🤖 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 `@crates/biome_html_parser/src/token_source.rs` around lines 348 - 353, Update
the token-advancing methods around bump_with_context and skip_as_trivia so they
preserve the active Svelte framework instead of constructing
HtmlLexContext::default(). Reuse the current framework when creating the
context, and ensure parse_html_element fallbacks and Svelte paths continue
lexing {{ as two L_CURLY tokens.

[dyc3]

formatting these inner expressions is not actually implemented yet: https://biomejs.dev/playground/?tab=formatter&pane=Diagnostics&code=ewAgAGMAbwBuAHMAbwBsAGUALgBsAG8AZwAoACAAIAAgACIAIgAgACAAKQB9AA%3D%3D&language=svelte