This is a big and messy change, and here's the gist:
- AvaliableSpace is now 2x AvailableSize (width and height)
- Layout algorithms are redesigned around the idea of available space
- When doing layout across nested formatting contexts, the parent
context tells the child context how much space is available for the
child's root box in both axes.
- "Available space" replaces "containing block width" in most places.
- The width and height in a box's UsedValues are considered to be
definite after they're assigned to. Marking something as having
definite size is no longer a separate step,
This probably introduces various regressions, but the big win here is
that our layout system now works with available space, just like the
specs are written. Fixing issues will be much easier going forward,
since you don't need to do nearly as much conversion from "spec logic"
to "LibWeb logic" as you previously did.
The HTMLIFrameElement does not create the nested browsing context on
insertion if the document does not have browsing context, which is not
set unless it's the active document.
Previously, in FrameLoader the document was not set as active until
after parsing, which led to iframes without nested browsing contexts,
and crashes.
Fixes#14207
We were capturing by reference which lead to the variables going
out of scope even when used in the lambda. Due to this the
update_starting_day_range lambda crashes when called.
Now that no one needs a Window just to create prototypes, we can remove
the internal window Object from the main thread VM and get rid of the
HTML::Window include for it.
This finally solves the reference binding to nullptr error in ladybird
that shows up when compiling it with ASAN.
The note in the spec says that we're supposed to make sure this new
document has the same Window as the old about:blank document, but we
forgot to actually assign to the Window pointer.
These classes only needed Window to get at its realm. Pass a realm
directly to construct Crypto, Encoding, HRT, IntersectionObserver,
NavigationTiming, Page, RequestIdleCallback, Selection, Streams, URL,
and XML classes.
These classes only needed Window to get at its realm. Pass a realm
directly to construct DOM and WebIDL classes.
This change importantly removes the guarantee that a Document will
always have a non-null Window object. Only Documents created by a
BrowsingContext will have a non-null Window object. Documents created by
for example, DocumentFragment, will not have a Window (soon).
This incremental commit leaves some workarounds in place to keep other
parts of the code building.
This Intrinsics object hangs off of a new HostDefined struct that takes
the place of EnvironmentSettingsObject as the true [[HostDefined]] slot
on JS::Realm objects created by LibWeb.
This gets the intrinsics off of the GlobalObject, Window, similar to the
previous refactor of LibJS to move the intrinsics into the Realm's
[[Intrinics]] internal slot.
A side effect of this change is that we cannot fully initialize a Window
object until the [[HostDefined]] slot has been installed into the realm,
which happens with the creation of the WindowEnvironmentSettingsObject.
As such, any Window usage that has not been funned through a WindowESO
will not have any cached Web prototyped or constructors, and will not
have Window APIs available to javascript code. Currently this seems
limited to usage of Window in the CSS parser, but a subsequent commit
will clean those up to take Realm as well. However, this commit compiles
so let's cut it off here :^).
If the stored symbol is custom (non-global), it is allocated on the
heap, and may not be visited by anyone else, so we must register it in
order to keep it alive.
This was previously indirectly forcing Heap/Handle.h to include it
instead. This will let us include Handle.h from PropertyKey, which will
allow us to solve a different issue.
Previously Welcome relied on a bogus executable key value to disable
startup. This always printed an error on login and littered the config
file with a useless entry. Adding/removing the group as needed seems
a bit nicer.
Plumbs synchronous calls for adding and removing group entries to
config files. This is useful for services like SystemServer which
default to group names for executable paths, and for removing all
keys at once.
I thought the spec listing out the elements again was an oversight, but
it isn't, as simply inverting "is_actually_disabled" makes :enabled
apply to every element.
Previously we only considered an element disabled if it was an <input>
element with the disabled attribute, but there's way more elements that
apply with more nuanced disabled/enabled rules.
This check happens very often in LibGUI code. 25% of time spent
layouting the emoji input dialog was wasted on RTTI. Adding a simple
fast_is<Widget>() melts almost all of that away.
If a flex item is itself a flex container, we were previously lying when
asked if the item created a BFC. It creates an FFC, so stop lying about
this in FormattingContext::creates_block_formatting_context().
We were using `>=` instead of `>` when checking if a float with a given
width could fit in the available space. If the width was an exact match,
we rejected it! Oops :^)
If the 2D transform in effect is just a simple translation, we don't
need to draw into a temporary bitmap and then transform it. We can
just translate the painter. :^)
I couldn't find anything in the specs about this, but GMail uses
empty generated boxes (`::before` and `::after` with `content: ""`)
inside a flexbox container in order to vertically center things.
The flexbox spec tells us to not generate flex items for empty
*anonymous* boxes, so we continue not doing that, but generated boxes
(any pseudo-element box) now always produce a flex item. This probably
isn't perfect either, and we'll have to revisit it for stuff like
`::first-letter`.
Let's not assume the containing block has a paintable box just because
someone is calling set_needs_display(). It can just be a no-op in that
case, and nobody gets hurt.
Instead of formatting contexts flailing around to figure out from the
"inside" how much space is available on the "outside", we should
provide the amount of available space in both axes as an input to run().
This basically means that when something creates a nested formatting
context, the parent context is responsible for telling the nested context
how much space is available for layout. This information is provided
immediately when invoking run().
Note that this commit doesn't pass accurate values in all cases yet.
This first step just makes it build, and passes available values in some
cases where getting them was trivial.
Previously, changing the time format caused Taskbar to crash.
This commit also simplifies TaskbarWindow::config_string_did_change()
so that it is in line with the changes made to
Calendar::config_string_did_change().
Fixes#15384
New actions in the Layer Menu allows for the creation of a new layer
from the current selection. Layers can be made by copying the
selection or cutting it from the current layer. The new layer will be
sized to the bounding box of the selection. The newly produced layer
will be added to the layer stack.
URL had properly named replacements for protocol(), set_protocol() and
create_with_file_protocol() already. This patch removes these function
and updates all call sites to use the functions named according to the
specification.
See https://url.spec.whatwg.org/#concept-url-scheme
This inserts some CSS and JS to make images in Markdown documents which
are wider than the viewport, become shrink-to-fit. Clicking on these
toggles them between shrink-to-fit and full size.
Anyone who displays Markdown documents using LibWeb gets this
functionality for free! That's Browser, Help, and Welcome's README
display.
The tzname array stores the abbreviated names of the current time zone
when in standard and daylight time. These abbreviations are ambiguous;
a single abbreviation often maps to multiple time zones. For example,
EST is used by America/New_York and America/Detroit, and CST could be
the abbreviation of Central Standard Time or China Standard Time.
Instead, we mimic a subset of how both ICU and Howard Hinnant's "date"
library determines the current time zone. First, we try to parse the TZ
environment variable. If that fails, or isn't set, we try to resolve the
/etc/localtime symbolic link. On most Linux systems and on macOS, this
is a link to the current TZDB file in use.
If all of the above fails, we fall back to UTC.
In addition to changing the uid, the method also changes the gid and
properly sets groups. So this patch will also mitigate the security
issue of `pls`.
This option, often used with only a lonely dash, allows to "simulate a
login". For now, it just changes the current directory to the home of
the new user.
This is needed to eventually share a header list between a Request or
Response object's internal infra request/response and the object's
exposed Header object.