toStream(location, state) renders to a web-standard ReadableStream of
UTF-8 bytes: pass it to new Response() on web servers or
Readable.fromWeb().pipe(res) on Node. It awaits store prefetch promises
(state.prefetch, the pattern the #653 thread wanted standardized) and
lazy route views before rendering, then flushes template output
progressively through async holes.
lazy(loader, loadingView?) wraps a dynamic import as a route handler:
loads once, caches forever, renders the loading view (or holds the
current tree via a placeholder) while in flight, emits render on arrival.
toString() refuses lazy routes and prefetching stores with pointers to
toStream() — both halves of the async-route story ship together, unlike
the original PR.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
The server tag now builds a parts list (string runs + async holes) instead
of an eager string. Sync templates collapse to one string and toString()
is byte-identical to before. Promises and async iterables in child
position make the result streamable: iterating it yields everything before
a hole immediately, then each hole as it resolves, in document order —
resolved values get full child semantics (templates, arrays, raw,
escaping). toString() on async content throws with streaming guidance;
async values in attribute position throw immediately.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
npm run bench (happy-dom, all three engines) and npm run bench:browser
(Playwright Chromium, the two ESM engines). Honest numbers recorded in
docs/v8.md: we create ~12% faster than µhtml in real Chromium (the
parse-once/clone design), µhtml updates in place ~5x faster than
fresh-tree + nanomorph (choo's architectural cost, mitigated by component
caching), and server strings are on par with nanohtml v1.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
Three end-to-end scenarios against a static+SSR test server: the
zero-build page (import map + native ESM, no bundler), the SSR page
(rendered content in the raw response, live after hydration, no mismatch
warnings), and adoption (an expando on the server-rendered node survives
a real render). Renders are raf-batched so assertions poll.
Real Chromium flushed out three fixes happy-dom couldn't see:
- nanoraf called an extracted requestAnimationFrame bare — Illegal
invocation under strict-mode ESM (sloppy CJS had masked it); wrapped.
- the counter example only routed '/', so serving it from any subpath
threw; it now has a wildcard fallback.
- hydration mismatch detection is now whitespace-insensitive (the parser
reparents whitespace, e.g. text after </body>), and page scripts belong
in <head> when a view owns <body> — same convention bankai v9 used.
CI gets an e2e job with chromium-headless-shell.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
hydrate(oldNode, newNode) normalizes the client tree to parser text-node
granularity, walks both trees to report the first server/client markup
disagreement (node, text, attribute, or child count, with a path), then
morphs — matching nodes are adopted in place, the client render wins.
choo.mount() now hydrates and console.warns on mismatch.
Building this surfaced two real isomorphism divergences, both fixed:
adjacent text nodes from template holes vs the parser's merged runs
(folded via Node.normalize), and the server serializing onclick="" where
the browser sets a property — event handlers now leave no trace in server
markup at all.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
The integration test runs the counter through the real browser path in
happy-dom: start(), store events, raf-batched morph, DOM click handlers,
emit coalescing. examples/counter is the Phase 2 exit criterion: one app
module mounted zero-build in a browser via import map and string-rendered
by node examples/counter/render.js.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
Ported from nanocomponent 6.6.0 and on-load 3.4.1 (MIT) with attribution:
render/update/morph lifecycle, proxy nodes with isSameNode identity, and
the MutationObserver-driven load/unload hooks. This class is the intended
island/hydration boundary for later phases. @choojs/core exposes ./timing
as a subpath for the instrumentation calls.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
Each template literal parses once, keyed by its strings array in a
WeakMap: static parts become a <template> with hole markers, renders
clone and fill. Interpolated values never pass through innerHTML (child
values become text nodes or adopted DOM; attributes go through
setAttribute). Event handlers set as properties so nanomorph copies them.
Document-level roots (<body>/<head>/<html>) parse via DOMParser because
<template> drops them. Replaces the v7 browserify transform entirely —
production speed is now a runtime property.
Tests run in happy-dom (new devDependency), including a check that server
and browser renderers serialize identical markup.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
Replace nanorouter's regex+decodeURI normalization with the WHATWG URL
parser. Path segments are percent-decoded exactly once in the trie, with
a keep-raw fallback instead of the URIError that crashed v7 on a literal
'%'. Route definitions and locations are NFC-normalized so 'café' matches
regardless of composition. Hash-to-slash rewriting now handles every hash,
not just the first, and state.href is decoded for humans (raw on failure).
Behavior changes from v7: params are no longer double-decoded (%2540 →
'%40', not '@'), and unmatchable encodings fall back to raw segments
instead of silently 404ing.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
Same public API as choo 7.1.0 (use/route/start/mount/toString/emit; choo()
callable with or without new, Choo exported). nanobus, nanorouter+wayfarer,
nanolru+component cache, nanoraf, nanohref, nanotiming, document-ready and
scroll-to-anchor are ported into lib/ with per-file attribution; nanoquery
is replaced by URLSearchParams. The v7 node test suite is ported from tape
to node:test and passes unchanged in behavior.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
ESM ports with attribution: nanohtml 1.10.0 server tag (browserify/babel
transform branches removed — pure runtime), raw-server, and nanomorph 5.4.3
consolidated into one module. browser.js is a Phase 2 placeholder that
throws loudly. Covered by test/server.test.js on node:test.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014NgfSjHE11oFpoSnLVKLXd
As described by #563:
> https://polyfill.io provides a fallback for browsers that don't support it, so for most people this will be a strict improvement.
`discify` is a different module than intended. discify is an alternative
to disc that works as a plugin, but we were using it as a separate
command line program. This switches it out for disc. The disc command
line program is called `discify` which is probably where the confusion
came from.
Also swapped out unassertify and uglifyify for tinyify, which adds tree
shaking.
Replaces #528. I was just going off the features listed in the readme. There are a handful more we should test but this is a start.
Note that this includes a readme change for `state.href` that I believe was incorrectly documented. Let me know if I got that wrong.
Also note that we should delete the original `v6-tests` branch once this is merged. Leaving it for now for reference.
