Skip to content

fix: load the URL shown in the address bar when returning to a pushState/replaceState history entry#16307

Closed
Nic-Polumeyv wants to merge 2 commits into
sveltejs:version-3from
Nic-Polumeyv:fix-shallow-history-url
Closed

fix: load the URL shown in the address bar when returning to a pushState/replaceState history entry#16307
Nic-Polumeyv wants to merge 2 commits into
sveltejs:version-3from
Nic-Polumeyv:fix-shallow-history-url

Conversation

@Nic-Polumeyv

Copy link
Copy Markdown
Contributor

closes #11465

pushState and replaceState write one URL to the address bar and store a different one in the history entry:

const opts = {
	[HISTORY_INDEX]: (current_history_index += 1),
	[NAVIGATION_INDEX]: current_navigation_index,
	[PAGE_URL_KEY]: page.url.href, // the page we were on
	[STATES_KEY]: state
};

history.pushState(opts, '', resolve_url(url)); // the URL the user sees

When you later return to that entry, the popstate handler feeds PAGE_URL_KEY to navigate, so load runs with the old URL while the address bar shows the pushed one. That's the report in the issue, url.searchParams in load missing params that are visibly in the URL after going back.

Both functions now store the same resolved URL they push. This follows teemingc's comment on the issue, keep the page store URL at push time but update the URL that subsequent loads run with.

One existing assertion changes. "Replaces state on a new URL" expected that going forward to a replaced entry renders the old page under the new URL, which is the inconsistency being fixed. It now expects the route the address bar shows, with page.state still restored. A new test covers the reported sequence, push, navigate away, go back, and fails on main.

This doesn't touch the invalidate-after-pushState case from #11503. That needs a decision on push-time load semantics, and an existing test ("Invalidates the correct route after pushing state to a new URL") deliberately locks in the current behavior, so I left it alone. Happy to take it as a follow-up if there's an agreed direction.


Please don't delete this checklist! 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
  • This message body should clearly illustrate what problems it solves.
  • Ideally, include a test that fails without this PR but passes with it.

Tests

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

Changesets

  • If your PR makes a change that should be noted in one or more packages' changelogs, generate a changeset by running pnpm changeset and following the prompts. Changesets that add features should be minor and those that fix bugs should be patch. Please prefix changeset messages with feat:, fix:, or chore:.

Edits

  • Please ensure that 'Allow edits from maintainers' is checked. PRs without this option may be closed.

@pkg-svelte-dev

pkg-svelte-dev Bot commented Jul 10, 2026

Copy link
Copy Markdown

Install the latest version of @sveltejs/kit from 83d6bb7:

pnpm add https://pkg.svelte.dev/@sveltejs/kit/c/83d6bb7dc6cb515dccc04d2379469a3edafccb89

Open in pkg.svelte.dev: https://pkg.svelte.dev/repos/kit/pr/16307

Note

This PR is from a fork. A maintainer must approve approve each commit before it can be built and installed.

@changeset-bot

changeset-bot Bot commented Jul 10, 2026

Copy link
Copy Markdown

🦋 Changeset detected

Latest commit: 83d6bb7

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

This PR includes changesets to release 1 package
Name Type
@sveltejs/kit 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

Comment thread packages/kit/test/apps/options-2/test/env.test.js Outdated
Comment thread packages/kit/src/runtime/server/page/render.js
@Nic-Polumeyv
Nic-Polumeyv force-pushed the fix-shallow-history-url branch from fc3a5e5 to 30677e6 Compare July 10, 2026 18:59
@Nic-Polumeyv
Nic-Polumeyv force-pushed the fix-shallow-history-url branch from dc5938f to 0abc5b7 Compare July 11, 2026 01:19
@Nic-Polumeyv
Nic-Polumeyv requested a review from teemingc July 13, 2026 19:43
teemingc added a commit that referenced this pull request Jul 14, 2026
…DE_ENV (#16313)

The NODE_ENV saga continues. Any build where NODE_ENV isn't `production`
currently ships with every origin check compiled out:

```js
// respond.js
if (!DEV) {
	// cross-site remote request check
	// cross-site form submission check (csrf.checkOrigin / trustedOrigins)
}
```

`DEV` comes from esm-env, which kit bundles into apps resolved by
NODE_ENV, so `NODE_ENV=staging vite build` produces a deployable server
with CSRF protection silently disabled. Vite's docs list non-production
builds as a supported workflow, and khromov raised this exact failure
mode in #14335 when the gate was added. At the time `DEV` was assumed to
only mean `vite dev`, which the esm-env consolidation in #14308 no
longer guarantees (first leak #15632, fixed by #15852, second leak
#16008, fixed by #16023).

This applies the rule from the #15852 body ("reverting to the global
constant where we want the code to only run during `vite dev` ... Every
other place that uses DEV is meant to fire warnings ... and that should
still happen if the user did `NODE_ENV=development vite build`") to two
sites it missed:

- the `respond.js` gate, `!DEV` to `!__SVELTEKIT_DEV__`, covering both
the form CSRF check and the remote functions origin check that now
shares it
- the live-serving condition in `app/server/remote/prerender.js`, same
one-token swap (prerendered remote functions were served live in these
builds)

The three validation `DEV` sites in `respond.js` (`validateHeaders`,
`page_nodes.validate()`, `validate_server_exports`) are deliberately
untouched. They are warning sites and should keep following NODE_ENV per
the rule. The check internals are also untouched, so this stays out of
the way of the `sec-fetch-site` idea in #15992.

The new spec in `test/build-errors` builds a minimal fixture with
`NODE_ENV=staging`, imports the generated server, and asserts a normal
GET returns 200 while a forged cross-site form POST returns 403. It
fails on main (the POST returns 200) and passes with the fix. The
existing basics CSRF tests stay green in both modes, and dev behavior is
unchanged.

There are HMR-coupled `DEV` gates in `client.js` and the remote query
modules from the same class. Left for a follow-up, `client.js` currently
conflicts with #16307.

---

### Please don't delete this checklist! Before submitting the PR, please
make sure you do the following:
- [x] 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
- [x] This message body should clearly illustrate what problems it
solves.
- [x] Ideally, include a test that fails without this PR but passes with
it.

### Tests
- [x] Run the tests with `pnpm test` and lint the project with `pnpm
lint` and `pnpm check`

### Changesets
- [x] If your PR makes a change that should be noted in one or more
packages' changelogs, generate a changeset by running `pnpm changeset`
and following the prompts. Changesets that add features should be
`minor` and those that fix bugs should be `patch`. Please prefix
changeset messages with `feat:`, `fix:`, or `chore:`.

### Edits

- [x] Please ensure that 'Allow edits from maintainers' is checked. PRs
without this option may be closed.

---------

Co-authored-by: Tee Ming <chewteeming01@gmail.com>
@Nic-Polumeyv
Nic-Polumeyv changed the base branch from main to version-3 July 15, 2026 01:07
@Nic-Polumeyv
Nic-Polumeyv force-pushed the fix-shallow-history-url branch from 0abc5b7 to 68acb5f Compare July 15, 2026 01:07
@Nic-Polumeyv
Nic-Polumeyv force-pushed the fix-shallow-history-url branch from 0ee794b to 83d6bb7 Compare July 17, 2026 04:49
@dummdidumm

Copy link
Copy Markdown
Member

Thank you, but this isn't right - when you go back you want to preserve the shallow navigation state. Therefore closing.

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.

pushState and replaceState causing inconsistencies with different url

3 participants