Skip to content

Reject namespaced attribute - #461

Merged
JoviDeCroock merged 1 commit into
mainfrom
fix-link
May 17, 2026
Merged

Reject namespaced attribute#461
JoviDeCroock merged 1 commit into
mainfrom
fix-link

Conversation

@JoviDeCroock

Copy link
Copy Markdown
Member

No description provided.

@changeset-bot

changeset-bot Bot commented May 17, 2026

Copy link
Copy Markdown

🦋 Changeset detected

Latest commit: e32a4cd

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

This PR includes changesets to release 1 package
Name Type
preact-render-to-string 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

@JoviDeCroock
JoviDeCroock merged commit 833043b into main May 17, 2026
1 check passed
@JoviDeCroock
JoviDeCroock deleted the fix-link branch May 17, 2026 09:51
@github-actions github-actions Bot mentioned this pull request May 17, 2026
valthon added a commit to valthon/zigapagos that referenced this pull request Aug 8, 2026
… injection fix

Supersedes dependabot PR #124 (happy-dom 20.11.0 -> 20.11.1); the sweep
below was researched against each package's releases rather than taking the
grouped bump as-is.

preact-render-to-string ^6.6.2 -> ^6.7.0 is the change that matters on its
own. 6.7.0 reordered the attribute serializer so UNSAFE_NAME rejection runs
before namespace normalization (preactjs/preact-render-to-string#461):
previously an attribute key that looked namespaced but contained spaces or
`>` was rewritten and emitted rather than dropped — markup injection through
prop KEYS, e.g. `xlinkHref><script>…`. That lands on our only SSR path:
runtime/src/core.ts re-exports renderToString and runtime/sidecar/render.ts
drives it with author-content props from src/islands/props.zig. The lockfile
already resolved 6.7.0, so built sites were safe; the declared ^6.6.2 was
the hole, because any fresh resolution could legally land back on the
vulnerable 6.6.2. Raising the floor closes it.

preact ^10.29.7 -> ^10.29.8: two perf fixes — flushSync now actually
batches (preactjs/preact#5173; we re-export it but never call it) and
retained-subtree traversal is skipped after sCU/memo bailouts (#5182).
Diff-layer change, so validated through bun test (750/0) plus a full
examples/tsx-site build (SSR + hydration bundle + CSP emitters).

happy-dom + @happy-dom/global-registrator ^20.11.0 -> ^20.11.2 (lockstep,
as they release): 20.11.1 stops constructing DOMException on the
query-selector hot path (a straight speedup for this DOM-heavy suite);
20.11.2 fixes MutationObserver callbacks being held only via WeakRef and
silently dying after a GC — latent here, but exactly the class of defect
that presents as an unreproducible flaky test.

typescript stays at ^6.0.3 — and permanently, not pending a 6.1: 6.0.3 is
the final release of the JavaScript codebase (microsoft/TypeScript#62963,
the 6.0 line is in maintenance mode; the next stable on the registry is the
Go-based 7.0.2, which ships no JS compiler API). The dependabot.yml cap
below 7.0.0 and its unblock conditions (7.1's new API, expected ~Oct 2026,
then porting the three ts.createSourceFile call sites) are unchanged.

All three lockfiles are regenerated in this commit (rm bun.lock + install,
per the dependabot.yml note — site/ and examples/tsx-site embed their own
copy of @z/runtime's resolved set and a plain install does not refresh it),
under the newly pinned bun 1.3.14 so the lockfiles are canonical for the
toolchain that CI runs. The two regenerated trees are pinned back to
"configVersion": 0: a from-scratch lockfile is treated as a new project and
stamped configVersion 1, which flips future installs to the isolated
linker — a silent node_modules-layout change this repo doesn't want, since
props_check.zig resolves tsc from the website root's hoisted tree. 0 is the
value runtime/'s lockfile already carries and preserves hoisted semantics;
frozen installs from scratch were verified to honor it (no .bun store,
tsc resolvable at top level).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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