xmldom is a pure JavaScript W3C standard-based (XML DOM Level 2 Core) DOMParser and XMLSerializer module. Prior to @xmldom/xmldom versions 0.8.15 and 0.9.12, and in xmldom versions 0.3.0 through 0.6.0, two independent quadratic paths can cause denial of service. In lib/sax.js, parseElementStartPart repeatedly rescans a malformed tag name to the next > during single-character recovery; in lib/dom.js, normalize() repeatedly removes and appends adjacent text nodes, causing quadratic reindexing and string rebuilding. The first path is reachable through default DOMParser.parseFromString() processing, while the second is also reachable through a direct normalize() call on a programmatically constructed DOM, and endDocument invokes that normalization after parsing. This issue is fixed in @xmldom/xmldom versions 0.8.15 and 0.9.12; no fixed version is available for xmldom.

Affected: npm: @xmldom/xmldom, npm: xmldom

## Summary

`xmldom`'s malformed-input **error-recovery path** has two quadratic-time (O(n²)) behaviors that a single crafted input triggers together, so a tiny, highly compressible document (tens of KB) stalls the Node.js event loop for multiple seconds. It is reachable from `DOMParser.parseFromString` under **default options** — i.e. from unauthenticated, network-delivered XML — making this an unauthenticated denial of service. One of the two behaviors, the `normalize()` adjacent-text merge, is **additionally reachable programmatically** — via a plain `normalize()` call on a DOM built with adjacent text nodes, independent of the parser — so its fix must live in `normalize()`, not only in a parser bound.

## Details

### Finding A — `parseElementStartPart` quadratic re-scan

A `<` character is not a delimiter in any tag-parsing state, so `parseElementStartPart` scans forward character-by-character over any embedded `<` until it reaches the next `>` (or end of input), then validates the accumulated slice as a tag name and throws `invalid tagName:` on failure. The main loop catches this, reports an `error`, sets `end = -1`, and recovers by advancing a single character (`appendText(Math.max(tagStart, start) + 1)`). With a long run of `<` and a distant `>`, each of the O(n) recovery retries performs an O(n) scan plus an O(n) anchored regex validation over the growing candidate ⇒ **O(n²)**.

Code (0.9.x, `bb7a085dc5ba1eea3212388509b97bb4b4af32b9`):

- `parseElementStartPart` cha…

undici: undici: Denial of Service via unrequested WebSocket subprotocol encoding/asn1: golang: Go encoding/asn1: Denial of Service via excessive recursion in Unmarshal ip-address: ip-address: Server-Side Request Forgery via IPv4-mapped/NAT64 IPv6 address misclassification urllib: urllib: Credential leakage via cross-origin redirects net/http: golang: Go net/http: Unencrypted HTTP/2 connections vulnerable to Denial of Service html/template: golang: Go html/template: Cross-Site Scripting via pathological input encoding/xml: golang: Go: Denial of Service via XML decoding recursion depth issue net/url: golang: golang net/url: Denial of Service from quadratic complexity in path resolution crypto/tls: golang: Golang crypto/tls: Denial of Service via indefinite KeyUpdate messages nanoid: nanoid: Denial of Service via negative size input in non-secure module functions axios: axios: Denial of Service via uncontrolled recursion in form data processing pymdown-extensions: Pymdown-extensions: Denial of Service via Regular Expression Vulnerability brace-expansion: DoS via unbounded intermediate arrays, bypassing the CVE-2026-14257 mitigation ip-address: ip-address: Inconsistent IP address parsing leads to Server-Side Request Forgery (SSRF) and trust-boundary bypass nanoid: nanoid: Predictable ID generation due to integer overflow fast-uri: fast-uri: Server-Side Request Forgery via repeated hostname percent-decoding fast-uri: fast-uri: Host confusion via skipped IDN canonicalization fast-…

Affected: Ubuntu:20.04:LTS: node-xmldom, Ubuntu:22.04:LTS: node-xmldom, Ubuntu:24.04:LTS: node-xmldom, Ubuntu:26.04:LTS: node-xmldom

xmldom is a pure JavaScript W3C standard-based (XML DOM Level 2 Core) DOMParser and XMLSerializer module. Prior to @xmldom/xmldom versions 0.8.15 and 0.9.12, and in xmldom versions 0.3.0 through 0.6.0, two independent quadratic paths can cause denial of service. In lib/sax.js, parseElementStartPart repeatedly rescans a malformed tag name to the next > during single-character recovery; in lib/dom.js, normalize() repeatedly removes and appends adjacent text nodes, causing quadratic reindexing and string rebuilding. The first path is reachable through default DOMParser.parseFromString() processing, while the second is also reachable through a direct normalize() call on a programmatically constructed DOM, and endDocument invokes that normalization after parsing. This issue is fixed in @xmldom/xmldom versions 0.8.15 and 0.9.12; no fixed version is available for xmldom.