1.20.11

Features:

• using raw on a sized array prevents memcpy and instead binds a pointer.
• fields can now be declared as sized arrays
• structs with sized arrays can be initialized in structs

1.20.7

Features:

• open(array[CONST]) is a noop
• U8 <: CONST
• raw bindings in match statements

match u {
     UmbraLong { prefix=prefix, raw ptr=ptr } => ...
 }

1.20.5

Features:

• port tuple.lm to tuple.lsts
• add tests for LIB/cmp.lsts
• add tests for LIB/tuple.lsts

1.20.4

Features:

• moved about 100 lines of library code from LM to LSTS

1.20.3

Made several improvements to LSTS for macro usage.

declare-binop( $"!=", raw-type(U8), raw-type(U8), raw-type(U64), ( l"("; x; l"!="; y; l")"; ) );

1.20.2

Features:

print definitions can now be defined as generic with respect to output destination (stdout, stderr, file, etc).
Defined a generic implementation for the common print(x) and eprint(x) that sends data to stdout and stderr respectively.

1.20.1

Features:

• cmp returns an Ord enum instead of an I64.

1.20.0

Features:

• phi types

if condition {
   fclose(fp);
};

will fail because the two code paths can no longer be merged safely.

1.19.40

Global variables no longer need to be ascripted, because they need to be inferred anyways.

let x = 5;

instead of

let x = 5 :: U64;

1.19.38

Features:

• Clean up and port "apply" to LSTS
• permit list fields to be further destructured with list lhs syntax