Use memory more efficiently when reading files, serialising and deserialising - #9

Merged

kristiker merged 16 commits into

full-sourcegenfrom

memory-efficiency

Sep 8, 2026

Commits on Sep 8, 2026

Read array attributes in bulk
Angel-foxxo
committed
fca8e45
Store elements in a dictionary and a list
Angel-foxxo
committed
0c7301c
Generate element IDs on first use
Angel-foxxo
committed
7c7edbd
Skip the type table for class properties
Angel-foxxo
committed
0290370
Store arrays in shared chunks
Angel-foxxo
committed
d56f2d4
Store attributes inline in slots
Angel-foxxo
committed
d76af7f
Measure the live heap of the typed model alone
Angel-foxxo
committed
a335d06
Set class properties through typed bindings
Angel-foxxo
committed
d65a50c
Size the element tables from the file
Angel-foxxo
committed
f4b2a0f
Shrink attribute slots to 40 bytes
Angel-foxxo
committed
408e2fa
Write elements from their slots and arrays in one piece
Angel-foxxo
committed
e63ea7b
Buffer the output and look attribute names up by identity
Angel-foxxo
committed
3c755d4
Size plain elements from the attribute count
Angel-foxxo
committed
069c7bd
Name the attribute types once, for slots and the binary encoding
Angel-foxxo
committed
9933eca
Add Offset alias to InlineValue for deferred attribute position

Copilot
and
kristiker
authored
5dc983d
Bump version
kristiker
authored
18c1e89