Three-part fix for hs-upstream-core/asyncError test 2/2:
1. runtime.sx hs-win-call: when an async call returns a rejected promise,
store the error value in window.__hs_async_error (side-channel) and
raise the sentinel "__hs_async_error__" so the value survives the
raise boundary intact.
2. compiler.sx catch clause: inject `(let ((var (host-hs-normalize-exc var))) ...)`
around the catch body so the sentinel gets swapped for the real error
object before user code runs. Uses let (not set!) so shadowing works
correctly for guard catch variables.
3. tests/hs-run-filtered.js:
- host-promise-state wraps JS Error objects as plain {message:...} dicts
before they cross the WASM boundary (Error.toString() was producing
"Error: boom" strings instead of accessible objects)
- host-hs-normalize-exc native retrieves the side-channel value when
the sentinel arrives in a catch variable
- host-get coercion restricted to El instances — plain JS objects with
a "value" key were being stringified to "[object Object]"
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add `bind` keyword to tokenizer, parse-bind-feat to parser, and
bind-feat no-op case to compiler. Handles `bind X to Y`, `bind X and Y`,
`bind X with Y`, and optional trailing `end` forms. All 43/44 bind tests
pass (1 is an explicit skip).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add (when (not (nil? target)) ...) guards after every hs-null-raise!
call in both the compiler and runtime so execution stops cleanly when
a DOM element is not found, instead of continuing into a JS operation
on null that takes ~34 seconds to propagate.
Compiler: emit-set dot/poss, emit-inc/dec poss case, remove-element,
remove-attr, add-styles all now wrap the action after hs-null-raise!
in a nil guard.
Runtime: hs-toggle-class!, hs-toggle-between!, hs-dispatch!,
hs-set-attr!, hs-toggle-attr!, hs-set-inner-html!, hs-put!,
hs-transition all guarded — hs-settle and hs-measure already were.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
parse-feat's paren-open handler stripped the grouping parens and
returned the inner feature, leaving any trailing ident (like `em`)
as a separate top-level feature. After consuming the closing paren,
now checks if the next token is a non-keyword ident or `%` op and
wraps as (string-postfix inner unit), making `(0 + 1) em` → "1em".
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Two fixes:
(1) compiler.sx: remove `it` from hs-reserved-var?. `it` is the standard
HS loop variable for `repeat in` loops; renaming it to `_hs_lv_it` made
the body reference the outer (nil) `it` rather than the bound element.
Other reserved vars (meta, event, result) still get renamed to prevent
shadowing built-ins in misnamed loops.
(2) runtime.sx: hs-make-object now appends an `_order` list tracking
insertion order, mirroring the pattern used by other dict-building paths.
Without this, `for prop in obj` fell back to `(keys obj)` which gives
non-deterministic key order for objects with string keys.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When parse-feat returns nil but the token stream is not at EOF,
coll-feats now throws a parse error ('Unexpected token X') instead
of silently returning the partial result. Fixes 'extra chars cause
error when evaling': eval-hs("1!") now correctly throws because '!'
is left over after parsing the number expression.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- parse-on-feat: event-vars paren check now restores position and returns empty
list when the first token after '(' is a keyword (command starter). Previously
'(log me)' was consumed as event variable names instead of a parenthesized
command, silently dropping the command body and returning empty innerHTML.
Fixes 'can support parenthesized commands and features'.
- parse-add-cmd: true-fallback now throws instead of returning nil when no 'to'
keyword follows the expression. Makes 'add - to' and similar invalid add forms
throw a parse error, satisfying assert-throws in 'basic parse error messages
work' and '_hyperscript() evaluate API still throws on first error'.
- read-class-name: '(' and ')' now only allowed inside '[...]' bracket groups
(depth > 0). Previously allowing them at top level caused '.innerHTML)' at the
end of a possessive expression to be consumed into the class token, producing
'innerHTML))' as a bogus property name. Tailwind classes like
'group-[:nth-of-type(3)_&]:block' still tokenize correctly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- parser: `toggle $var between v1 and v2 ...` → `(toggle-var-cycle $var (v1 v2 ...))`
- compiler: emit `(hs-toggle-var-cycle! win var-name values)` for new AST node
- runtime: `hs-toggle-var-cycle!` cycles through a list of values on a variable
- parser: `closest .sel to .target` / `closest #id to .target` / `closest sel to .target`
now consumes the `to` keyword and parses the target expr instead of defaulting to beingTold
- tokenizer: `read-class-name` handles backslash escapes and allows `(`, `)`, `&`
chars so Tailwind classes like `group-[:nth-of-type(3)_&]:block` tokenize correctly
- platform.py: `domListen` drives async result via `_driveAsync` after `cekCall`
- test: fixed-time toggle asserts `.foo` IS present after click (toggle started, 10ms window open)
- generate-sx-tests.py: aligned MANUAL_TEST_BODIES for timed toggle with corrected assertion
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- parse-halt-cmd: after consuming 'the event's', check for 'bubbling'
token and return "bubbling" mode instead of "the-event"
- parse-wait-cmd: skip article words (a/an/the) before reading event
name, so 'wait for a customEvent' works correctly
- parse-on-feat: parse optional (vars) paren group before flt and
consume-having!, so 'on intersection(intersecting) having ...' works;
inject event-var refs into body for compiler's event-refs mechanism
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- parser: nil return in parse-cmd for feature keywords (on/init/def/behavior/live)
so "then on click" correctly hands off to outer coll-feats loop
- compiler: cek-try wrap for undefined variable refs in coll-where compilation
so "doesNotExist where it > 1" returns nil instead of throwing
- integration: hs-activate! detects script[type=text/hyperscript-template] and
applies handler to DOM instances via hs-query-all(component attr) not to script el
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- parser.sx: parse-logical now rejects mixed and/or without parens
- parser.sx: parse-arith now rejects mixed +/-/* //%/mod without parens
- generate-sx-tests.py: MANUAL_TEST_BODIES for short-circuit and/or,
typecheck (direct hs-type-assert calls), template string test
- generate-sx-tests.py: Pattern 5 for error("expr") -> assert-throws
- hs-run-filtered.js: redefine try-call to _run-test-thunk after loading
so assert-throws actually catches exceptions (was always {ok true})
- hs-run-filtered.js: clear __hs_deadline immediately after test eval
to prevent cascading timeout fires in result inspection K.eval calls
- hs-run-filtered.js: typecheck suite in _NO_STEP_LIMIT_SUITES and
_SLOW_DEADLINE_SUITES (hs-type-assert JIT is slow on first call)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Capture raised exception in a let-bound variable before the guard
exits, then re-raise after. Avoids the WASM OCaml kernel bug where
(raise e) called from within a guard handler re-invokes the same
handler infinitely.
Affects hs-repeat-forever, hs-repeat-times, hs-repeat-while,
hs-repeat-until, hs-for-each. Repeat suite: 25/30 → 28/29 counted
(1 skipped: 'until event keyword works' requires async event dispatch).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
parse-on-feat now skips 'queue MODE' tokens before parsing the body,
so 'on foo queue first ...' and 'on foo queue last ...' parse correctly.
Compiler ignores queue mode (catch-all drops unknown parts).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Parser: parse-set-cmd now emits (set-el! target value) when target is
a query node (e.g. #id, .class), keeping (set! ...) for all other
targets.
Compiler: add (set-el! ...) handler that calls hs-set-element!; revert
emit-set for query targets back to hs-set-inner-html! so that
put "x" into #target keeps setting innerHTML rather than replacing
the element.
Runtime: hs-set-element! new function — parses value as HTML into a
temp div; if it contains element children, replaces the target element
via replaceChild and boots hyperscript on the new element; otherwise
falls through to hs-set-inner-html!. Removes the spurious
host-to-list wrapper that was causing len() to always return 0.
Result: all 8 assignableElements tests pass (set #id / set .class /
set closest / swap, plus put-into-still-works-as-innerHTML).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- hs-value-of-node: use selectedIndex fallback when SELECT.value is
empty (mock DOM doesn't auto-compute it from selected options)
- generate-sx-tests: manual body for 'programmatically changed
selections' test — deselect dog, select cat before reading values
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- parse-atom: unrecognized keywords (e.g. index) fall back to ref,
fixing 'set index to N' parse failure
- hs-set-inner-html!: join list values as "" so 'put [A,C] into el'
concatenates strings not [object Object]
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add style case to the of-target compiler branch so
'set *color of #el to x' emits dom-set-style correctly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Store target element in meta.owner when hs-on fires;
hs-fetch-impl dispatches beforeFetch on it before the perform.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Track halt mode via __hs-no-stop flag; skip stopPropagation when
handler raised hs-halt-default (from 'halt default'). All other
halt variants (halt, halt the event, halt bubbling) unchanged.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- compiler: wrap catch body in nested guard so (raise e) inside a
catch handler defers the re-raise until after the guard exits,
avoiding the handler-stays-active infinite loop
- generator: MANUAL_TEST_BODIES for rethrown/uncaught exception events,
can-pick-detail/event-property, bootstrap bootstraps; remove from
skip-list; regenerate behavioral spec
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
collect-pairs! in parse-add-cmd now skips the semicolon op token
between CSS properties, so add {color: red; font-family: monospace}
compiles to two dom-set-style calls instead of three malformed ones.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Parser: parse-cmd-list now skips a leading 'then' token so that
'on click from #bar then add .clicked' compiles correctly instead
of producing nil as the body.
Bootstrap tests: fix two broken tests whose assertions were
incomplete or contradictory:
- "cleanup removes event listeners" — deactivate + re-click to
verify listener is gone
- "reinitializes if script attribute changes" — actually change
the _ attribute before re-activating and re-clicking
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
emit-set on call command re-applied so `it`/`the-result` bound after call.
A11 hide now 16/16 via count-filter unlock (was partial +3, now done +4).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
parser.sx: detect bare ident "initial" after "to" in parse-one-transition,
emit string sentinel instead of (ref "initial") which evaluated to nil.
runtime.sx: hs-transition stores pre-first-transition style as
data-hs-init-{prop}; restores it when value=="initial".
Also commits E37 tokenizer and E40 fetch test implementations that
accumulated in the working tree but weren't staged in prior commits.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Implements hs-tokens-of, hs-stream-token, hs-stream-consume,
hs-stream-has-more, hs-token-type, hs-token-value, hs-token-op?,
hs-raw->api-token, hs-eof-sentinel in runtime.sx.
Tokenizer emits whitespace tokens after the first content token;
stream functions skip them for look-ahead and consume. Parser
filters whitespace tokens at hs-parse entry. Dot/hash after close
brackets split into PERIOD/POUND + IDENTIFIER. Template escape \$
produces literal $.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Merge c36fd5b2 stripped the source-info dict unwrapping from hs-to-sx
(the (let ((ast (if (and (dict? ast) (:hs-ast)) ...) wrapper) and also
introduced E37 tokenizer whitespace-token changes that broke the parser.
Reverts tokenizer/runtime to pre-E37 HEAD~1 state, restores hs-to-sx
with AST unwrapping from 61c9697f, and adds back the hs-id= dispatch
clause. Baseline: 178/195.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add hs-raw->api-token, hs-eof-sentinel, hs-api-list, hs-tokens-of,
hs-stream-token, hs-stream-consume, hs-stream-has-more, hs-token-type,
hs-token-value, hs-token-op? to runtime. Fix tokenizer to emit whitespace
tokens and handle dot/hash after closing brackets. Fix hs-tokens-of to
accept bare :template keyword flag via &rest args + some() check.
Remaining failure (string interpolation isnt surprising) requires full
DOM activation infrastructure.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Lost when resolving E37 reformat conflicts — re-added:
- hs-id= function in runtime.sx (JS === for elements, = for scalars)
- hs-id= dispatch in compiler.sx (after = clause)
Parser already uses hs-id= for != operator (unchanged).
- hs-id= uses JS === for DOM elements (hs-ref-eq), = for scalars
- != operator now uses hs-id= for structural correctness
- compound tag[attr=val] selector matching in test runner
- dom-query-all replaces host-call querySelectorAll
- DOM tree structure corrected in 4 generated tests (elements were
appended to wrong parents)
Fix compiler: (block-literal () body) was emitting bare body instead of
(fn () body). Now always wraps in fn regardless of param count.
Generator: MANUAL_TEST_BODIES for all 4 blockLiteral tests using apply
and SX map rather than JS array.map.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Parser: parse-js-block extracts raw JS source by character positions.
Compiler: js-block AST → hs-js-exec call, stores result in it.
Runtime: hs-js-exec creates JS Function, handles promise rejection.
Test runner: host-new-function/host-promise-state natives + promise monkey-patch.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
hs-make-object no longer appends _order to every HS object literal.
Generator emit_eval now uses assert-equal (equal?) for dict-containing
expected values instead of assert= (= reference equality).
Together these fix arrayLiteral "arrays containing objects work".
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
empty .class compiled (empty-target (query ".class")) to
(hs-empty-target! (hs-query-first ".class")) via hs-to-sx — only
emptying the first match. Fix: detect (query ...) target in the
empty-target compiler case and emit (for-each (fn (_el)
(hs-empty-target! _el)) (hs-query-all sel)) instead, mirroring the
add-class pattern. Suite hs-upstream-empty: 12/13 → 13/13.
Smoke 0-195: 175/195 unchanged.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
${}{"val"} pattern in add {prop: ${}{"val"}} uses two consecutive brace
groups: empty ${} followed by {"val"} for the actual expression. The prior
fix called parse-expr when already at the brace-close of the empty group,
returning nil. New fix: detect empty ${} (brace-open then brace-close),
skip the close, then read the actual value from the following {…} block.
Also handles non-empty ${expr} directly as before.
Suite hs-upstream-add: 17/19 → 18/19. Smoke 0-195: 174/195 → 175/195.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
call X then put it into Y was emitting (hs-win-call ...) without
wrapping in emit-set, so it remained nil. Wrap call result in
emit-set(the-result) so it/the-result are updated. Fixes +1 test.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Prevents click events from bubbling into ancestor elements that also
have hs handlers (e.g. parent re-inserting HTML after child click).
Fixes put-reprocessing tests 1147/1149/1150 (+3 tests).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
tell now rebinds beingTold/you/yourself without overwriting me.
Parser implicit targets use beingTold; handler wrapper seeds beingTold=me.
Fixes: attributes refer to the thing being told, does not overwrite me,
your symbol represents the thing being told.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
