Sigil

This project publishes no release notes. Between 4.6.0 and 4.6.1 there were 2 commits, 1 of them substantive:

• naturally, introduced a diversion from the spec around tail. injection

This project publishes no release notes. Between 4.5.1 and 4.6.0 there were 11 commits, 9 of them substantive:

• switch to a -beta
• Add overloads to allow pointer loads of unbaked functions
• Don't insert tail calls when returns are mismatched
• rework fix for #34 check on return types in tail. cases
• make a note of a case we're handling implicitly here
• eh, not doing the Issue# style in Sigil right now... maybe change that later
• get ready for 4.5.3
• new failing test for a validation bug
• fix a failing verification case w/ valuetype <-> object

This project publishes no release notes. Between 4.5.0 and 4.5.1 there were 15 commits, 14 of them substantive:

• fix bug where Convert<bool> and ConvertOverflow<bool> threw exceptions
• Add core-clr tooling (project.json, etc); still builds for same platforms - no core-clr changes
• Build for Core-CLR
• Test projects now compile for all framework types
• Down to 6 failures
• Fix everything except the StreamWriter.Flush call; for some reason that breaks the test runner
• Isolated criteria for breaking the dnx xunit test runner - see WTFDNXTEST
• Add SNK; fix disassembler test (ld,dup vs ld,ld - both are valid)
• Mark as a beta (because everything upstream is)
• Update exclusions based on feedback from dotnet/corefx
• Fix the intent of the 3 catch-all tests
• Transition to rc1
• Add support down to netstandard 1.2 (required TypeHelpes.EmptyTypes); separate build for netstandard 1.4 (only code difference is compiled regex); make dependencies less specific (was required previously to yank pre-release bits)
• more TypeHelpers into Impl

This project publishes no release notes. Between 4.4.0 and 4.5.0 there were 14 commits, 11 of them substantive:

• Failing test for instruction size with switch
• Fix InstructionSize.Get for switch
• Add support for creating type initializers
• Fixed the checks on IsBuildingConstructor being the wrong way round
• Fix CreateTypeInitializer calling CreateConstructor instead of the CreateTypeInitializer overload.
• Added syntax highlighting to readme
• fix #29; basically inspecting the ops for real, and encoding a separate set of 'expected ops' in the tests
• fix nongeneric type initializers for #26
• get started on #19; make recursion work for nongeneric dynamic methods
• .NET doesn't let me actually figure out parameters on MethodBuilders, so recursion on Emits with MethodBuilders isn't possible; fixes #19
• adopt suggested NewObject overload; fixes #28

This project publishes no release notes. Between 4.3.1 and 4.4.0 there were 4 commits, 2 of them substantive:

• Make reused local initialization optional
• normalize and extend the initializeReused behavior to all DeclareLocal methods

This project publishes no release notes. Between 4.3.0 and 4.3.1 there were 7 commits, 5 of them substantive:

• Test disassemble Func<float>
• Fix disassemble ldc.r4
• failing test for local reuse to uninitialized value
• fix uninitialized local reuse bug
• fix case when re-defaulting locals where the valuetype is user defined

This project publishes no release notes. Between 4.2.2 and 4.3.0 there were 3 commits, 2 of them substantive:

• LoadIndirect tests for char and bool
• LoadIndirect support for char and bool