[jogibear9988]

Member initializers emit constructor arguments inline, leaving earlier arguments on the stack across a later try/finally. Delegate constructor emission to the existing TryEmitNew path, then apply the member bindings.

Class and struct constructors, argument/finally/constructor/binding order and initialized values, compared with System compilation. On .NET Framework the struct EH initializer uses equivalent direct C# as its oracle because the Framework System compiler rejects that shape; class cases use System compilation. Regressions require strict FEC compilation. Native suites cover Standard and LightExpression.

Validation

Local validation uses master baseline 094d067 plus this change. Release solution build: zero warnings/errors.

Runtime	TestX groups	Legacy tests
.NET 10	489 passed	1,756 passed
.NET 8	489 passed	1,755 passed
.NET Framework 4.7.2	468 passed	1,639 passed

Measurements

Master baseline 094d0678 → this change, Release, Windows ARM64, .NET 8.0.28 / 10.0.9. Compilation uses prebuilt trees; runtime measures invocation through the same wrapper. Two fresh process rounds with reversed load order, seven samples per round, about 30 ms calibration, one logical CPU and tiered compilation disabled. Tables use the median of the two per-round medians.

The controls below produce correct results on master for their measured inputs. Newly supported trees have no valid before/after execution comparison because master rejects them, produces incorrect results or emits invalid IL. Timings include machine and JIT/code-placement variation; the unchanged add control helps show that variation. These observations do not prove universal performance neutrality.

net8.0 results (master → change)

API / workload	Compile ns	Invoke ns	Compile B/op	Invoke B/op
Standard / add_control	1148.7 → 1190.3	2.0 → 2.0	872.0 → 872.0	0.0 → 0.0
Standard / member_init	1451.1 → 1544.4	3.8 → 3.7	1064.1 → 1064.1	24.0 → 24.0
Standard / member_init_constructor	1535.0 → 1464.7	3.9 → 4.1	976.1 → 976.1	24.0 → 24.0
Light / add_control	1074.7 → 1061.6	2.1 → 2.0	872.0 → 872.0	0.0 → 0.0
Light / member_init	1457.0 → 1220.9	3.7 → 3.8	1064.0 → 1064.0	24.0 → 24.0
Light / member_init_constructor	1452.7 → 1264.0	3.7 → 3.5	976.0 → 976.0	24.0 → 24.0

net10.0 results (master → change)

API / workload	Compile ns	Invoke ns	Compile B/op	Invoke B/op
Standard / add_control	3289.8 → 3295.8	2.0 → 1.9	872.0 → 872.0	0.0 → 0.0
Standard / member_init	3907.4 → 3918.2	3.9 → 3.9	1064.2 → 1064.2	24.0 → 24.0
Standard / member_init_constructor	4100.7 → 4138.2	4.1 → 4.1	976.2 → 976.2	24.0 → 24.0
Light / add_control	2562.6 → 2517.8	2.0 → 2.0	872.0 → 872.0	0.0 → 0.0
Light / member_init	3249.7 → 3277.8	4.8 → 3.9	1064.0 → 1064.0	24.0 → 24.0
Light / member_init_constructor	2954.2 → 2985.4	3.9 → 3.8	976.0 → 976.0	24.0 → 24.0

Emitted IL differs in these measured workloads: none.
No measured allocation increase above 0.1 B/op in these controls.