The submitForm encoding path was the last duplicate of the "limited to
only known values" canonicalization the previous commit consolidated for
the IDL getters. Now it consumes the same Form.normalizeMethod /
Form.normalizeEnctype helpers, so a single function owns the canonical
mapping (`""` / unknown -> spec default, recognized values pass through
unchanged).
Side effect of routing through the helper: the
`log.warn(.not_implemented, "FormData.encoding", ...)` branch falls out.
After commit 4b693db4 added `text/plain`, the only attribute values that
still reach the urlencoded fallback are spec-invalid ones, which per
HTML §4.10.21.5 silently canonicalize to
`application/x-www-form-urlencoded`. The warning was firing for valid
spec behavior — Chrome doesn't log either.
Behavior-preserving on all observable surfaces: full suite 639/639 green;
existing form-submission integration tests (multipart, urlencoded,
text/plain, GET-ignores-enctype) all pass unchanged.
Closing the divergence introduced by the new IDL accessors: `submitter.formEnctype`
(and `form.enctype`) now return "text/plain" for that attribute value per WHATWG
HTML §4.10.21.5, but `Frame.submitForm` previously fell back to urlencoded with
a `.not_implemented` log when it saw the same value on the submission path.
Implement the spec's text/plain encoding algorithm (HTML §4.10.21.8):
- FormData.EncType gains a `.plaintext` variant.
- FormData.plaintextEncode writes "name=value CRLF" per entry, no URL-encoding,
no escaping — the spec accepts that text/plain is a lossy, human-readable
encoding (values containing "=" or CRLF produce an ambiguous wire format
by design).
- Frame.submitForm recognizes "text/plain" before the urlencoded fallback and
sets the Content-Type header to "text/plain; charset=<form-charset>", per
spec step 21.4.
Two new Zig unit tests cover encoding output (`FormData: plaintext write`,
`FormData: plaintext empty body`). Full suite 639/639 green.
This is bundled with the IDL accessor commits because returning "text/plain"
from the IDL while the submission silently re-encodes as urlencoded is a
spec-internal inconsistency the IDL change itself introduces. Reviewers who'd
prefer to land just the read-only accessors first should feel free to ask for
a split — this commit is self-contained and reverts cleanly.
The "limited to only known values" canonicalization (per WHATWG HTML
§2.2.2) was duplicated five times: Form.getMethod + Form.getEnctype +
{Button,Input}.{getFormMethod,getFormEnctype}. Each callsite differed
only in the missing-value default ("" for submitter overrides, "get" /
"application/x-www-form-urlencoded" for the form-side).
Extract into two pub helpers on Form.zig taking the attribute slice +
the missing-value default. The five callers collapse to one-liners.
Behavior-preserving: existing form.html / button.html / input-attrs.html
fixtures all pass unchanged; full suite 637/637 green.
Net: -36 LOC.
Six form-submission IDL accessors were missing from the JsApi blocks of
HTMLFormElement, HTMLButtonElement, and HTMLInputElement, so reads
produced undefined instead of the spec-mandated string/boolean. The
content-attribute path (clicking a submit button honoring formaction /
formmethod / formenctype) was wired up in #2279; this commit adds the
matching IDL-property accessors per WHATWG HTML §4.10.18.6 and §4.10.21.5.
- Form.enctype: limited to known values, missing+invalid both default to
application/x-www-form-urlencoded (mirrors getMethod's shape).
- Button/Input formAction: returns frame.url when missing/empty, else the
resolved URL (mirrors Form.getAction).
- Button/Input formEnctype, formMethod: limited to known values with no
missing-value default ("" when missing, canonical invalid-value default
application/x-www-form-urlencoded / get when invalid).
- Button/Input formTarget: plain reflection, defaults to "".
- Button/Input formNoValidate: boolean reflection of formnovalidate.
Closes#2449
Remove no-longer needed setTimeouts in test now that messages are queued.
Runner also checks ready_queue when determining doneness.
Co-authored-by: Navid EMAD <design.navid@gmail.com>
At some point recently, we started to process scripts that fail to load (e.g.
404). This stops such scripts from [trying] to be evaluated, and executes the
onerror handler in all script loading paths.
Routes the eval tool through callEval instead of callValue. This
ensures JavaScript execution errors are flagged with is_error=true,
allowing the model to self-correct.
Changes `EvalResult` to a union and uses Zig errors for operational
failures, distinguishing them from JavaScript execution errors.
Also removes redundant `stringifyJson` in favor of standard library.
- Derives the `Action` enum in `src/browser/tools.zig` from `tool_defs` to prevent manual maintenance and potential drift.
- Refactors `verifyElementValue` in `src/script/Verifier.zig` to use a `Check` struct for its parameters.
- Reorganizes and cleans up imports in `src/agent/Agent.zig`.
- Replace `unreachable` with explicit error returns in Agent and Executor.
- Add `EvalResult.err` helper in `tools.zig` to simplify error paths.
- Improve error propagation in `SlashCommand` and `Recorder`.
- Rename `Self` to `Agent` for better clarity in function signatures.
Introduces an `ActionTarget` union to handle selectors and backend node
IDs consistently. This simplifies the `formatActionResult` function and
ensures the identification method is preserved within `NodeAndPage`
during element resolution.
- Add `beginTool` and `endTool` to `Terminal` to encapsulate spinner logic.
- Consolidate navigation awaiting and context tagging in browser tools
via a new `finalizeAction` helper.
Client.tick drains self.queue (assigning conns to queued transfers) only
at the start. When perform / processMessages releases a batch of conns
back to the pool, those conns sit idle until the next tick — a queued
transfer that could have run this tick waits one Runner iteration
(~20 ms in the test runner) for no reason. Adds a second drainQueue
call after perform so newly-freed conns get picked up immediately.
In practice this matters whenever httpMaxHostOpen / httpMaxConcurrent
is exceeded — pages with N > limit subresources had each "wave" of
queue overflow paying one extra tick of latency.
cache=true is problematic for a few reasons
1 - The current cache implementation is known to cause timing issues; i.e. it
executes callbacks synchronously.
2 - Unlike something like robots.txt or proxy, cache tests need to be explicitly
tested. The response has to include cache headers and the resource loaded
again.
Moves page URL and title formatting into a separate `appendPageContext`
function and removes the page parameter from `formatActionResult`.
This simplifies tool execution logic and decouples action descriptions
from page metadata.
- Update `substituteEnvVars` to return `OutOfMemory` error instead of
silently returning the original string on failure.
- Simplify `execFill` to always use the original text for display.
- Inline the `writeHealHeader` helper in `src/script.zig`.
