You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the library writes liquid VM code into separate growable buffers (in vm_assembler_t) for each Liquid::C::BlockBody object. However, for block body serialization and deserialization (#77) we will want to share an immutable memory allocation across block bodies, so block bodies will need offsets into that shared memory for execution.
Since block bodies will be connected in this way for execution, we may also want to make parsing more efficient by re-using a vm_assembler_t for a block body once that block body has finished being compiled (e.g. when its parent block tag is finished compiling), by not embedding it in the block_body_t struct. That way we could save on the number of memory allocations for these vm_assembler_t buffers and will have less vm_assembler_t buffers to combine at the end of parsing / compilation. For instance, we could have one vm_assembler_t per-level of nesting in the template.
The text was updated successfully, but these errors were encountered:
Currently, the library writes liquid VM code into separate growable buffers (in vm_assembler_t) for each Liquid::C::BlockBody object. However, for block body serialization and deserialization (#77) we will want to share an immutable memory allocation across block bodies, so block bodies will need offsets into that shared memory for execution.
Since block bodies will be connected in this way for execution, we may also want to make parsing more efficient by re-using a
vm_assembler_t
for a block body once that block body has finished being compiled (e.g. when its parent block tag is finished compiling), by not embedding it in theblock_body_t
struct. That way we could save on the number of memory allocations for these vm_assembler_t buffers and will have lessvm_assembler_t
buffers to combine at the end of parsing / compilation. For instance, we could have onevm_assembler_t
per-level of nesting in the template.The text was updated successfully, but these errors were encountered: