Checked the three test files in Firefox and Chrome:
- stepUp/stepDown use the HTML step base (min, else the value attribute)
and snap off-ladder values to the next rung, counting the snap as the
first step as browsers do; clamping lands on the last rung inside
min/max.
- time strings keep a three-digit fraction.
- an empty pattern attribute is a pattern (matches only "").
- tooLong/tooShort only for values last changed by a user edit, so the
text-entry path marks the value and script/attribute values never trip
them; same for textarea.
- showPicker dropped: browsers throw NotAllowedError without a gesture,
a no-op would be a lie.
- month/week assertions skipped where the browser has no such input.
measureText was missing and roundRect, ellipse, setLineDash, gradients,
patterns, isPointInPath and the shadow/filter properties threw, so
Chart.js, D3 and ApexCharts died at init. Widths come from the
deterministic text model in text_measure.zig; nothing is painted.
https://github.com/lightpanda-io/browser/pull/3333 improved markdown rendering
for flexbox items. But screenshot generation didn't benefit from that change.
This commit introduces a RenderTree that extracts and shares functionality
between markdown and screenshots.
react-hook-form and date pickers read valueAsNumber/valueAsDate and got
undefined. The date types convert through the existing validators; the
same conversion now also drives min/max range validation for
date/time/month/week/datetime-local, which only number/range had.
addListener/removeListener were no-ops and nothing fired when
Emulation.setDeviceMetricsOverride crossed a breakpoint. Each
MediaQueryList now registers on its Page; Browser.setViewportOverride
tells pages to re-evaluate and dispatch MediaQueryListEvent('change') on
the ones whose matches flipped. onchange is registered as a listener so
JS-side dispatchEvent reaches it too.
click(), willValidate and the WebDriver click read only the control's own
attribute. They now use Element.isDisabled, which already walks fieldset
ancestors with the first-legend exception for :disabled. The disabled IDL
attribute keeps reflecting the attribute alone, per spec.
Three one-accessor gaps: getSetCookie returns Set-Cookie values
unjoined, getName maps a constructor back to its tag, part is a
DOMTokenList over the part attribute like relList.
Composes transform attributes up to the nearest <svg> (getCTM) or through
every enclosing <svg>, offset by the outermost one's client rect
(getScreenCTM). Null when detached. d3-zoom, d3-brush and svg-pan-zoom
call these at init and threw.
window.stop() is less destructive than other mechanisms we have. For one, it
seems largely isolated to pending or inflight HTTP requests. For anther, it
keeps the page intact.
To achieve this, HttpClient gains an `cancelRequests` which is a gentler version
of `abortOwner`. It cancels inflight/pending HTTP requests, which results in
error callbacks (not shutdown callbacks) firing.
Just like https://github.com/lightpanda-io/browser/pull/3189 I ran into the
problem that I couldn't distinguish between an HTTP request that was canceled
because of user-action (e.g. calling window.stop(), or xhr.abort()) and an HTTP
request that was internally aborted. These now have distinct errors/flows so
that we can present the correct state. Most places that aborted now all
transfer.cancel() which results in a distinct `error.TransferCanceled` (some
places still abort -> `error.Abort`). It should be possible to revisit 3189 now.
The CDP "Page.stopLoading" now hooks into this new behavior. Fixes
https://github.com/lightpanda-io/browser/issues/3351
We currently have 1 note: it prints the server's listening address:port. Note
is a special un-ignorable level. This keeps the "note" level, but logs it under
a new scope: "note", so that it _can_ be silenced with a `--log-filter note`.
Add a new note, on startup, that displays tips. Currently, only displays when
--obey-robots is not enabled:
NOTE note : config tips . . . . . . . . . . . . . . . . . . . [+0ms]
robots = use '--obey-robots' to use a sites robots.txt
meta = use '--log-filter note' to silence this message
A client that disconnects might get treated as a harsher terminate failure (e.g.
watchdog). This doesn't have a huge impact, but it makes the CI flaky and it
produces more logs than is necessary.
In a terminate state, the driver will now check its inbox to see if this is a
client disconnection.
Three small compliance (WPT) changes with respect to error handling.
1. When a unique index creation fails due to a constraint violation, the
delivery of that failure is asynchronous.
2. When txn.abort() is called, any undelivered request should be aborted
asynchronously (see the common theme here?)
3. An failure in an onupgradeneeded handler correctly fires an AbortError AND
rollsback the version
The Fetch.enable call takes `patterns` which can limit the path and type of
request that should be intercepted. This adds the wildcard support.
As before, RI is currently only enabled for Requests, not response. A warning
is printed if RI for responses is requested.
Fixes: https://github.com/lightpanda-io/browser/issues/3349
In order to support Selenium the way people are used to, it looks like we need
to support both WebDriver classic (WebDriver) and WebDriver BiDi (BiDi). Typical
scripts look like a mix of the two, e.g. using WebDriver to control the browser
and using BiDi to receive notifications. This commit:
1 - adds a --protocol (cdp|webdriver) CLI argument to the `serve` command to
enable one or the other protocol (defaulting to CDP)
2 - adds basic WebDriver endpoint to let a Selenium client connect. This
implementation is hackish and sits on top of our simple Handshake handler.
The handshake handler is well past its original design. Serving /json/version
and /metrics from it was one thing. But Driving the entire browser session? This
will get a follow up PR.
1. Abstract "Driver". Non-CDP things that referenced *CDP now reference a Driver
2. Move the NodeRegistry out of CDP. This created an artificial link between
agent / mcp and CDP
3. Add BiDi driver with enough to navigate to a page
This enforces a limit on the captures responses based on the Network.enable's
maxTotalBufferSize and maxResourceBufferSize fields. It also limits the size of
the post body we echo (e.g. in requestWillBeSent) based on the `maxPostDataSize`
field.
Karl Seguin