fix(get-selector): escape control characters in attribute selectors - #5273
Merged
Conversation
Contributor
There was a problem hiding this comment.
Pull request overview
This PR hardens selector generation (axe.utils.getSelector) so attribute values containing C0 control characters (U+0000–U+001F) or DEL (U+007F) are escaped as CSS numeric escapes, preventing invalid selectors that can throw inside Element.matches and abort an axe.run().
Changes:
- Update
escapeAttributeto escape all C0 control characters + DEL using CSS numeric escapes (\<hex>). - Coerce the attribute value via
String(str)before escaping to ensure consistent string handling. - Add a regression test covering control characters (form feed, vertical tab, U+001F) in attribute selectors.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| lib/core/utils/get-selector.js | Expands attribute escaping to cover all control characters that would otherwise break CSS selector parsing. |
| test/core/utils/get-selector.js | Adds a regression test to ensure generated selectors remain valid and usable with matchesSelector. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
chutchins25
force-pushed
the
chut/5204-escape-attr-control-chars
branch
from
July 31, 2026 13:19
db11901 to
4607470
Compare
chutchins25
marked this pull request as ready for review
July 31, 2026 13:25
straker
reviewed
Aug 3, 2026
Attribute values containing a C0 control character (e.g. form feed U+000C, a CSS newline) produced an invalid selector, so Element.matches threw and the entire axe.run failed. Escape all control characters as CSS numeric escapes. Closes #5204
chutchins25
force-pushed
the
chut/5204-escape-attr-control-chars
branch
from
August 3, 2026 21:03
4607470 to
4b4b568
Compare
straker
approved these changes
Aug 4, 2026
WilcoFiers
added a commit
that referenced
this pull request
Aug 5, 2026
All notable changes to this project will be documented in this file. See [commit-and-tag-version](https://github.com/absolute-version/commit-and-tag-version) for commit guidelines. ## [4.13.0](v4.12.1...v4.13.0) (2026-08-05) ### Features - **aria-actions:** add aria-actions to allowed ARIA attributes ([#5200](#5200)) ([029655d](029655d)), closes [#4584](#4584) [#5199](#5199), references [#5215](#5215) [#5215](#5215) - **aria-allowed-attr:** flag deprecated ARIA attributes as needs-review ([#5246](#5246)) ([518f3cc](518f3cc)), closes [#3341](#3341) - **aria-prohibited-attr:** allow many elements to be named and disallow label and body from being named ([#5259](#5259)) ([d8b1ea5](d8b1ea5)) - **aria-roles:** add sectionheader and sectionfooter roles ([#5238](#5238)) ([c36c109](c36c109)), closes [#4734](#4734), references [#4734](#4734) - **aria/get-aria-value:** new function to get aria values of a node ([#5109](#5109)) ([a7d8f3e](a7d8f3e)), references [#5042](#5042) - **aria/has-attr-value:** new function to check if node has aria value ([#5136](#5136)) ([61f2624](61f2624)), references [#5109](#5109) - **aria:** support role=image as equivalent to role=img ([#5248](#5248)) ([5aa8aaf](5aa8aaf)), closes [#4656](#4656), references [#5272](#5272) - **checks/aria:** support ARIA element internals properties ([#5172](#5172)) ([9b7f754](9b7f754)) - **checks/label:** support ARIA element internals properties ([#5170](#5170)) ([21c5f8b](21c5f8b)) - **checks/navigation:** support ARIA element internals properties ([#5167](#5167)) ([2c3a98f](2c3a98f)) - **commons/aria:** support ARIA element internals properties ([#5171](#5171)) ([31f09e7](31f09e7)) - **commons/dom:** support ARIA element internals properties ([#5163](#5163)) ([f0a12cf](f0a12cf)) - **commons/forms:** support ARIA element internals properties ([#5165](#5165)) ([27a4686](27a4686)) - **commons/matches/fromPrimative:** deprecate in favor of correct spelling ([#5270](#5270)) ([31cfb2e](31cfb2e)) - **commons/text:** support ARIA element internals properties ([#5169](#5169)) ([e841a33](e841a33)) - **commons/text:** support form-associated labels via element internals ([#5182](#5182)) ([57cfe0a](57cfe0a)), closes [#5045](#5045), references [#5170](#5170) [#5039](#5039) [#5151](#5151) [#5039](#5039) - **dom/getResolvedRefs:** new function to get the resolved virtual nodes of idrefs ([#5151](#5151)) ([489cdea](489cdea)), references [#5109](#5109) - **element-internals:** enable ElementInternals by default ([#5284](#5284)) ([2740d42](2740d42)), closes [#5277](#5277) - **i18n:** Add Swedish locale ([#5190](#5190)) ([dcd13f2](dcd13f2)), references [#5189](#5189) - **matches:** add inSectioningContent, hasChild, and isSummaryForDetails matches ([#5262](#5262)) ([c47cdcd](c47cdcd)) - **rules:** support ARIA element internals properties ([#5168](#5168)) ([065baf7](065baf7)) - **standards/ariaAttrs:** add caseInsensitive property for attributes ([#5224](#5224)) ([bcd791c](bcd791c)) ### Bug Fixes - **aria-allowed-role:** allow roles on a non-details summary ([#5242](#5242)) ([3bd9875](3bd9875)), closes [#3911](#3911), references [#3443](#3443) [#3911](#3911) - **aria-allowed-role:** restrict figure roles with child figcaption ([#5240](#5240)) ([178a635](178a635)), closes [#3443](#3443) - **aria-prohibited-attr:** visible aria-labelledby requires review only ([#5285](#5285)) ([fd6fa9f](fd6fa9f)) - **axe.d.ts:** make enabled property of RuleMetadata optional ([#5129](#5129)) ([90fce18](90fce18)) - **color-contrast:** fix various stacking context bugs ([#5214](#5214)) ([d5e5b04](d5e5b04)), references [#8](#8) [#5213](#5213) - **gather-internals:** handle non-HTMLElement nodes ([#5161](#5161)) ([06e84c3](06e84c3)) - **get-selector:** escape control characters in attribute selectors ([#5273](#5273)) ([4b60ac5](4b60ac5)), closes [#5204](#5204) [#5204](#5204) - **image-alt:** allow whitespace alt on presentational images ([#5218](#5218)) ([c5dd0ef](c5dd0ef)), closes [#5216](#5216) - **landmark-unique:** exclude section/form with non-landmark roles from landmark match ([#5085](#5085)) ([c5fd013](c5fd013)), closes [#4722](#4722) [#5064](#5064) - name the image role in role-img-alt and svg-img-alt metadata ([#5279](#5279)) ([995a269](995a269)), closes [#5272](#5272), references [#5248](#5248) [#5248](#5248) - **standards:** update aria-errormessage and aria-details to be idrefs ([#5157](#5157)) ([fb94f8a](fb94f8a)) This PR was opened by a robot 🤖 🎉
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #5204. When an element's attribute value contains a C0 control character (observed: form feed
U+000C),generateSelectoremitted it raw into an attribute selector. In CSS, form feed is a string-terminating newline, soimg[alt=" \f¼"]is invalid,Element.matchesthrows insidegetNthChildString, and the entireaxe.run()fails — one bad attribute value on a page kills the whole audit.Fix
escapeAttribute(inlib/core/utils/get-selector.js) previously escaped only\,", and CR/LF newlines — missing form feed and every other control character. It now escapes all C0 control characters (U+0000–U+001F) and DEL (U+007F) as CSS numeric escapes (\<hex>), which subsumes the existing newline handling (U+000A→\a,U+000C→\c). This matches the issue's minimal proof thatCSS.escape('\f')produces a valid selector.Compatibility
No behavior change for existing cases: newline (
U+000A) still escapes to\a. CR and CRLF now escape to\d/\d \a(previously both collapsed to\a) — more correct round-tripping, and not covered by any existing test. All existingget-selectorescape tests still pass.Test
test/core/utils/get-selector.js— an attribute value containing form feed, vertical tab, andU+001Fnow produces a valid selector that both matches the expected escaped string and passesmatchesSelector(which threw on the pre-fix code).Confirmed by @straker on the issue.
Closes #5204