fix: don't mistake `type` identifier expressions for TS `type` declarations in tags by baseballyama · Pull Request #18330 · sveltejs/svelte

baseballyama · 2026-05-30T06:12:09Z

Detect TypeScript type declarations by actually parsing instead of by character-class blacklists, so that expressions like {type === 'all' ? a : b} or {type instanceof Foo} aren't misclassified as malformed declarations.

Closes #18328

Before submitting the PR, please make sure you do the following

It's really useful if your PR references an issue where it is discussed ahead of time. In many cases, features are absent for a reason. For large changes, please create an RFC: https://github.com/sveltejs/rfcs
Prefix your PR title with feat:, fix:, chore:, or docs:.
This message body should clearly illustrate what problems it solves.
Ideally, include a test that fails without this PR but passes with it.
If this PR changes code within packages/svelte/src, add a changeset (npx changeset).

Tests and linting

Run the tests with pnpm test and lint the project with pnpm lint

…ations in tags Detect TypeScript `type` declarations by actually parsing instead of by character-class blacklists, so that expressions like `{type === 'all' ? a : b}` or `{type instanceof Foo}` aren't misclassified as malformed declarations. Closes #18328

changeset-bot · 2026-05-30T06:12:12Z

🦋 Changeset detected

Latest commit: cca40b1

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

This PR includes changesets to release 1 package

Name	Type
svelte	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

baseballyama · 2026-05-30T06:13:41Z

I also submitted a PR to improve the regular expression (#18329), but I realized that handling edge cases such as generics would be difficult with a regex-based approach alone. As a result, I changed the implementation to first perform a lightweight check using a regular expression and then use Acorn for a more accurate determination.

svelte-docs-bot · 2026-05-30T06:15:08Z

Preview: https://svelte-dev-git-preview-svelte-18330-svelte.vercel.app/

github-actions · 2026-05-30T06:15:24Z

Playground

pnpm add https://pkg.pr.new/svelte@18330

dummdidumm

Goddammit I feared something would slip through. Left a few comments but overall solution makes sense

dummdidumm · 2026-05-30T09:37:39Z

+	{type / 2}
+	{type > 1}
+	{type instanceof Foo}
+	{type in foo}


These additions to this test should be enough we don't need all the other tests

If we’re going with an AST-based solution, then the change at parser.root.comments.length = initial_comment_count; (https://github.com/sveltejs/svelte/pull/18330/changes#diff-07a9d6689d63c982b2e3a45ae4bd624240d27824e8c5bdce8a539b4d2a7c36f1R96) is necessary. Without it, the test at {type instanceof /* probe */ Object} (https://github.com/sveltejs/svelte/pull/18330/changes#diff-37911caa74a576888755175e439f6605c76c5c17695b44de73e2f028295ee3a2R1) fails.

But we can delete packages/svelte/tests/runtime-runes/samples/declaration-tag-type-identifier test. 9826361

dummdidumm · 2026-05-30T09:42:12Z

+		// discard probe comments so `read_expression` doesn't re-add them
+		parser.root.comments.length = initial_comment_count;
+		return null;
+	}


This feels overcomplicated. If you add type to the list of unsupported in the regex and then in the "if unsupported" condition check if it's a usage of type as a reference (using your new regex) then don't throw the error it should come down to the same result

Are you saying that you think the approach in #18329 is preferable to using an AST-based solution?

I feel the AST-based approach is more robust. With the approach in #18329, we’d likely keep running into edge cases and need to add support for them individually—for example, patterns involving generics such as {type<string>('generic-call')}.

Coverage is already provided by the validator samples (declaration-tag-invalid-type-2 exercises the fall-through path; declaration-tag-invalid-type-unicode/-generic exercise the parse-confirmed type-alias error path).

baseballyama · 2026-05-30T10:08:09Z

@@ -0,0 +1 @@
+{type instanceof /* probe */ Object}


This is for https://github.com/sveltejs/svelte/pull/18330/changes#diff-07a9d6689d63c982b2e3a45ae4bd624240d27824e8c5bdce8a539b4d2a7c36f1R96

Without this, comments end up being registered twice.

…tests

Copilot

Pull request overview

Fixes #18328: expressions starting with type (e.g. {type === 'all' ? a : b}, {type instanceof Foo}) were being misclassified as malformed TypeScript type declarations because of a regex character-class blacklist that couldn't tell identifiers from declaration keywords. The fix replaces the blacklist with a "maybe" hint and confirms by actually parsing the contents, falling back to expression parsing when the result is an ExpressionStatement.

Changes:

In state/tag.js, split type out of the unsupported-declaration regex into a regex_maybe_type_declaration hint, then disambiguate via parse_statement_at; if the result is an ExpressionStatement, drop the speculative comments and fall through to expression parsing, otherwise (a TS TSTypeAliasDeclaration) raise declaration_tag_invalid_type.
In utils/bracket.js, make find_regex_end only treat a / as the start of a regex literal when a closing / is found before the next newline, and avoid jumping i to Infinity in find_matching_bracket when no closing slash is present (fixes loose recovery for cases like {let x = a / }).
Added validator coverage for many {type ...} expressions plus a TS type alias, a parser-modern fixture asserting comment handling on the speculative parse, and an extended loose-declaration-tag fixture.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
packages/svelte/src/compiler/phases/1-parse/state/tag.js	Speculatively parse `{type ...}`, fall back to expression on `ExpressionStatement`, restore comment count
packages/svelte/src/compiler/phases/1-parse/utils/bracket.js	Only treat `/` as regex when closing slash precedes newline; guard against `Infinity` cursor advance
packages/svelte/tests/validator/samples/declaration-tag-invalid-type-2/input.svelte	Add `<script lang="ts">` and many `{type ...}` expression cases
packages/svelte/tests/validator/samples/declaration-tag-invalid-type-2/errors.json	Update expected error location to the remaining real type-alias case
packages/svelte/tests/parser-modern/samples/tag-type-identifier-probe-comment/{input.svelte,output.json}	Snapshot showing comments aren't duplicated when speculative parse falls through
packages/svelte/tests/parser-modern/samples/loose-declaration-tag/{input.svelte,output.json}	Cover loose recovery for an unterminated division-like `/`
.changeset/fix-type-identifier-in-tag.md	Patch changeset entry

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

dummdidumm

thank you!

This PR was opened by the [Changesets release](https://github.com/changesets/action) GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated. # Releases ## svelte@5.56.1 ### Patch Changes - fix: error at compile time on duplicate snippet/declaration tag definitions ([#18351](#18351)) - fix: parse declaration tag contents more robustly ([#18353](#18353)) - fix: correctly transform references to earlier declarators in a declaration tag (e.g. `{let a = $state(0), b = $derived(a * 2)}`) ([#18348](#18348)) - fix: avoid spurious `state_referenced_locally` warnings for `$derived` declarations in declaration tags ([#18348](#18348)) - fix: tolerate whitespace before `let`/`const` in declaration tags ([#18348](#18348)) - fix: prevent infinite loop when a tag's expression ends with a trailing `/` at the end of the input ([#18350](#18350)) - fix: more robust parsing of declaration tags with regards to `type` ([#18330](#18330)) - fix: preserve newlines in spread input values when the `type` attribute is applied after `value` ([#18345](#18345)) - fix: update `SvelteURLSearchParams` when setting duplicate keys to the same joined value ([#18336](#18336)) - fix: check references for blockers on server, too ([#18352](#18352)) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

This PR contains the following updates: | Package | Change | [Age](https://docs.renovatebot.com/merge-confidence/) | [Confidence](https://docs.renovatebot.com/merge-confidence/) | |---|---|---|---| | [svelte](https://svelte.dev) ([source](https://github.com/sveltejs/svelte/tree/HEAD/packages/svelte)) | [`5.56.0` → `5.56.1`](https://renovatebot.com/diffs/npm/svelte/5.56.0/5.56.1) | ![age](https://developer.mend.io/api/mc/badges/age/npm/svelte/5.56.1?slim=true) | ![confidence](https://developer.mend.io/api/mc/badges/confidence/npm/svelte/5.56.0/5.56.1?slim=true) | --- ### Release Notes <details> <summary>sveltejs/svelte (svelte)</summary> ### [`v5.56.1`](https://github.com/sveltejs/svelte/blob/HEAD/packages/svelte/CHANGELOG.md#5561) [Compare Source](https://github.com/sveltejs/svelte/compare/svelte@5.56.0...svelte@5.56.1) ##### Patch Changes - fix: error at compile time on duplicate snippet/declaration tag definitions ([#18351](sveltejs/svelte#18351)) - fix: parse declaration tag contents more robustly ([#18353](sveltejs/svelte#18353)) - fix: correctly transform references to earlier declarators in a declaration tag (e.g. `{let a = $state(0), b = $derived(a * 2)}`) ([#18348](sveltejs/svelte#18348)) - fix: avoid spurious `state_referenced_locally` warnings for `$derived` declarations in declaration tags ([#18348](sveltejs/svelte#18348)) - fix: tolerate whitespace before `let`/`const` in declaration tags ([#18348](sveltejs/svelte#18348)) - fix: prevent infinite loop when a tag's expression ends with a trailing `/` at the end of the input ([#18350](sveltejs/svelte#18350)) - fix: more robust parsing of declaration tags with regards to `type` ([#18330](sveltejs/svelte#18330)) - fix: preserve newlines in spread input values when the `type` attribute is applied after `value` ([#18345](sveltejs/svelte#18345)) - fix: update `SvelteURLSearchParams` when setting duplicate keys to the same joined value ([#18336](sveltejs/svelte#18336)) - fix: check references for blockers on server, too ([#18352](sveltejs/svelte#18352)) </details> --- ### Configuration 📅 **Schedule**: (UTC) - Branch creation - At any time (no schedule defined) - Automerge - At any time (no schedule defined) 🚦 **Automerge**: Disabled because a matching PR was automerged previously. ♻ **Rebasing**: Whenever PR is behind base branch, or you tick the rebase/retry checkbox. 🔕 **Ignore**: Close this PR and you won't be reminded about this update again. --- - [ ] If you want to rebase/retry this PR, check this box --- This PR has been generated by [Mend Renovate](https://github.com/renovatebot/renovate).  Co-authored-by: huskas-2189 <huskas-2189@noreply.codeberg.org> Reviewed-on: https://codeberg.org/huskas-2189/Bookmark/pulls/82

baseballyama mentioned this pull request May 30, 2026

fix: don't mistake type identifier expressions for TS type declarations in tags (Regexp based) #18329

Closed

6 tasks

dummdidumm reviewed May 30, 2026

View reviewed changes

baseballyama force-pushed the refactor-declaration-tag-parsing branch from 0b69c5d to 9826361 Compare May 30, 2026 10:02

test: remove redundant runtime test for type identifier expressions

b48ad46

Coverage is already provided by the validator samples (declaration-tag-invalid-type-2 exercises the fall-through path; declaration-tag-invalid-type-unicode/-generic exercise the parse-confirmed type-alias error path).

baseballyama force-pushed the refactor-declaration-tag-parsing branch from 9826361 to b48ad46 Compare May 30, 2026 10:04

baseballyama commented May 30, 2026

View reviewed changes

Comment thread packages/svelte/tests/validator/samples/declaration-tag-invalid-type-unicode/input.svelte Outdated

baseballyama commented May 30, 2026

View reviewed changes

baseballyama force-pushed the refactor-declaration-tag-parsing branch from 014cb4b to b48ad46 Compare May 30, 2026 10:13

baseballyama requested a review from dummdidumm May 30, 2026 10:13

dummdidumm and others added 2 commits May 31, 2026 22:20

simplify / fix "find matching bracket" bug around regex, consolidate …

2cea427

…tests

add test

00479b9

Copilot AI review requested due to automatic review settings June 1, 2026 00:09

Copilot started reviewing on behalf of baseballyama June 1, 2026 00:09 View session

Copilot AI reviewed Jun 1, 2026

View reviewed changes

baseballyama mentioned this pull request Jun 1, 2026

hangs (infinite loop) on a {const} / {let} whose expression ends with a bare / at end of input #18349

Closed

dummdidumm approved these changes Jun 1, 2026

View reviewed changes

Comment thread .changeset/fix-type-identifier-in-tag.md Outdated

Update .changeset/fix-type-identifier-in-tag.md

cca40b1

dummdidumm merged commit c74f44f into main Jun 1, 2026
14 of 15 checks passed

dummdidumm deleted the refactor-declaration-tag-parsing branch June 1, 2026 08:03

github-actions Bot mentioned this pull request Jun 1, 2026

Version Packages #18346

Merged

github-actions Bot mentioned this pull request Jun 2, 2026

Upgrade Svelte submodule to 5.56.1 baseballyama/rsvelte#639

Closed

Uh oh!

Uh oh!

Conversation

baseballyama commented May 30, 2026

Before submitting the PR, please make sure you do the following

Tests and linting

Uh oh!

changeset-bot Bot commented May 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🦋 Changeset detected

Uh oh!

baseballyama commented May 30, 2026

Uh oh!

svelte-docs-bot Bot commented May 30, 2026

Uh oh!

github-actions Bot commented May 30, 2026

Uh oh!

dummdidumm left a comment

Choose a reason for hiding this comment

Uh oh!

dummdidumm May 30, 2026

Choose a reason for hiding this comment

Uh oh!

baseballyama May 30, 2026

Choose a reason for hiding this comment

Uh oh!

dummdidumm May 30, 2026

Choose a reason for hiding this comment

Uh oh!

baseballyama May 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

baseballyama May 30, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

dummdidumm left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

changeset-bot Bot commented May 30, 2026 •

edited

Loading

baseballyama May 30, 2026 •

edited

Loading