Section 1.8.1.2 of Partition 3 talks about the "verification type system." As Peter points out, for verification, the number of types can be drastically reduced. Take a look at this section of the spec for a fairly rigorous explanation.
-----Original Message----- From: Peter Drayton [mailto:[EMAIL PROTECTED]] Sent: Sunday, August 18, 2002 11:32 AM To: [EMAIL PROTECTED] Subject: Re: [DOTNET-ROTOR] sbyte[] == byte[] I think this is by design - there's really little difference between the signed and unsigned versions of the short data types, except how the bits are interpreted. Assigning a value >127 to an int8 simply overflows. See example [1] below. Arrays of type "unsigned int8[]" are assignment-compatible with arrays of type "int8[]" (see example [2] below), which supports this theory. On a quick troll of the ECMA specs I couldn't find the exact spot that asserts this, though there are numerous references to unsigned being merely an alternate form of signed, and interpreted based on modifers (unsigned) and special opcodes (e.g. *.ovf). --Peter http://www.razorsoft.net/weblog http://staff.develop.com/peterd [1] assembly extern mscorlib {} .assembly overflow {} .method static void main() { .entrypoint .locals init (unsigned int8 b, int8 sb) ldc.i4 128 stloc b ldloc b stloc sb ldloc sb box [mscorlib]System.SByte call void [mscorlib]System.Console::WriteLine(object) ldloc b box [mscorlib]System.Byte call void [mscorlib]System.Console::WriteLine(object) ret } [2] .assembly extern mscorlib {} .assembly overflow {} .method static void main() { .entrypoint .locals init (unsigned int8[] rgb, int8[] rgsb) ldc.i4 1 newarr unsigned int8 dup stloc rgsb stloc rgb ret } > -----Original Message----- > From: Discussion of the Rotor Shared Source CLI implementation > [mailto:[EMAIL PROTECTED]] On Behalf Of Jeroen Frijters > Sent: Saturday, August 17, 2002 11:43 AM > To: [EMAIL PROTECTED] > Subject: [DOTNET-ROTOR] sbyte[] == byte[] > > > When I call a method expecting a byte[] with an sbyte[], the verifier > doesn't mind. Is this by design or is it a bug? > > .assembly extern mscorlib {} > .assembly test {} > .module test.exe > .imagebase 0x00400000 > .subsystem 0x00000003 > .file alignment 512 > .corflags 0x00000001 > > .class private auto ansi beforefieldinit Test > extends [mscorlib]System.Object > { > .method public hidebysig static void Main() cil managed > { > .entrypoint > .maxstack 1 > ldc.i4.1 > newarr [mscorlib]System.SByte > call void Test::Func(unsigned int8[]) > ret > } > > .method public hidebysig static void Func(unsigned int8[] b) cil > managed > { > .maxstack 0 > ret > } > > .method public hidebysig specialname rtspecialname > instance void .ctor() cil managed > { > .maxstack 1 > ldarg.0 > call instance void [mscorlib]System.Object::.ctor() > ret > } > } > > Regards, > Jeroen >
