The appearance promote carries the two-level sharing law the paint list
promote established: a group-free appearance moves its parked header
into the persistent region (once, however many lanes share it), a
paint-held group denies the move and clones through the Cow dispatch
that shares persistent interiors pointer for pointer, and stroke
parameter columns on a coverage deny nothing. Four regression tests pin
the move and share behavior, mirroring the paint promote suite.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Cover, Coverage, and Appearance mirror the upstream appearance model on
the record substrate with a 'static interior: the paint column stores
Graphic<'static>, and the appearance marker's value is Option<&Appearance>,
exactly as the Fill paint marker's is today. Stroke parameters ride the
coverage item as attributes elided at their defaults, with typed markers
whose defaults are pinned to Stroke::default() by test. The deep field
glue carries paint-held groups across the owned, resident, and
persistent seams as the paint list glue does.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
One read node per value type, mirroring master's read family. A name
means one value type, so there is no coercion between them: master's
number read silently widened integers, and these refuse instead, which
is the same rule the write side already enforces.
Each takes any record wire through the opaque reading input and serves
the name's own default where the attribute is absent, so the value
carries the declared type either way.
The gradient enums join the value-type rows now that a family exists to
read them at.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
An element generic declared on a reading input, absent from the output
and from every other input, has nowhere to go and nothing to be: the
node reads its declared attributes and never the element. That is an
attr-to-element map, so it accepts any upstream record wire.
The blessing is inferred rather than marked, since the signature already
says it. The two substrate pieces were both in place: the element rides
as the byte-carried token, so the registry takes one generic row instead
of a row per element type, and the reads resolve against the input's own
layout as they always have. Only the passthrough rule stood in the way,
in the three places that enforced it.
A generic the output carries, or one on an input with no reads, is
unchanged and still owes an implementations list.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The read is a number read, not a read of whatever the name holds: our
model has no coercion, so a name written at another value type is a graph
error rather than a conversion. The name says so.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
An absent read of a name the census declares now serves that name's own
default rather than its value type's, which is the rule for a declared
name. The census stages it exactly as it fills any absent field, as the
declared default's bytes, so the read reuses the census's own mechanism
instead of a second one; the compiler takes them when the name folds and
`set_layout` installs them beside the offset.
A name the census does not declare keeps the value type's default, which
is that case's own rule. One name carries one value type, checked when
the name folds, so the row's value type and width agree with the read's.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The fold already holds the name and the read input's finished layout, so
the offset falls out there rather than at construction: `RecordLayout`
carries the resolved numbers and `set_layout` copies them into the read
slots. Constructors are untouched, and census-marker reads keep their
current installation.
A read meets the value type the name was written at, so a disagreement
between a read here and a write upstream is the same graph error as two
writes disagreeing; the one-name-one-type check now spans reads and
writes together. An absent attribute stays absent and the read serves
the forced default rather than reporting it.
`read_attribute` is the catalog's get half, typed and never `Option` at
the kernel boundary, with the name declared exactly as the write side
declares it.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A name input declares where a placeholder's name is wired and nothing
more. The name is spent resolving the layout when the graph compiles, so
the kernel neither declares it nor is passed it; the return type's
`Named<X, V>` is the only tie between the write slot and the folded name.
The wire is untouched: the input keeps its declared position and crosses
as constant text, so a document's input order is unchanged. The
placeholder stays a concrete token, so it adds no generic for the node
to carry and none to go unconstrained.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The record fixtures spell their metas and resolved layouts out in full,
so the name-from-input fields reach them as empty. `named_value!`'s
example moves to `ignore`: its `for T` arm implements a core-types trait
on the value type, which only core-types itself may do.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
`Named<X>` in parameter position declares where a placeholder's name is
wired, and the macro gives that input constant text; `Attr<Named<X, V>>`
in the return writes under it. The kernel is handed the bare placeholder,
since the name is spent resolving the layout and a folded offset is all
the write needs, so the hot path matches a marker node's exactly.
A name-generic write names its value type through the wired generic,
which only an implementations row resolves, so those nodes emit their
layout meta per row rather than sharing one across rows.
`write_attribute` is the catalog's set half, restoring the identifier
master's documents carry with its input positions. Its name input is a
constant, so those documents resolve without migration, and the reset
marker that stood in for the missing node retires with it.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
`compute_layouts` resolves every name-from-input write against the text
its name input carries, so a folded meta is indistinguishable from a
marker node's and the layout holds a `&'static str` from there on. That
leaves nowhere for a name to be computed, which is the whole of the
rule: a name input that is not a constant is refused right here, per
node, with the path the editor pins its diagnostic to.
One name carries one value type, checked across the census and the
names a node folds together, so no graph can declare one field at two
widths. `compute_layouts` returns those refusals rather than panicking.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
An attribute's name and its value type separate: `AttrValue` declares
everything `Attribute` does except the name and its name-specific
default, and `Named<X, V>` pairs one with a name the graph supplies.
The value type stays concrete in the signature while the name varies
per instance, so a computed name has nowhere to come from.
`LayoutMeta` gains the name-from-input variant the fold consumes: a
`FieldWrite` template with every facet minted from the concrete type
and only the name left as a placeholder, plus the input position the
name's constant sits at. `fold_name` resolves one, preferring the
census row's descriptor where the name is declared so a known name
keeps its census default. `RecordLayout` carries the folded names so
`set_layout` can resolve offsets without a marker's `NAME`.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
