Summary
----------
When the same script file gets executed in multiple Runspaces, the RuntimeType generated from the powershell class defined in the file will be shared among those Runspaces. For different Runspaces, different SessionState should be used to run the static methods.
Fix
---
We use the `SessionStateKeeper` to query for the correct `SessionState` to use for a static method call.
We already use `SessionStateKeeper` to track the `Runspace/SessionState` where a PowerShell class is defined, and thus we can leverage this for the static method invocation.
* Update release documentation to reflect the desired release steps and order.
* Update in-doc links
* Update .spelling
* Address comments
* Additional changes to build/package example
* Update .spelling
* Remove another internal URL
* Name and link updates could be automated
* Move 'Note' to the top of 'Release Steps' section
Remove the code that spins up our own assembly load context. Keep the code that registers our `Resolve` method to the default loader's `Resolving` event, so that we can continue to do special assembly resolution as needed (such as the GAC probing logic needed for consuming FullCLR PS modules).
Essentially, the assembly `Microsoft.PowerShell.CoreCLR.AssemblyLoadContext.dll` is not needed anymore, the remaining code should be moved to S.M.A.dll. However, that will break DSC and other native hosts that are hosting powershell. So this assembly is kept for now.
Use the method NativeCommandParameterBinder.NeedQuotes, which is used by powershell native command processor, to check if quotes are needed. If yes, add quotes in the same way as our native command processor.
Also, make 'Invoke-Item' on Linux and OSX able to invoke an executable properly.
* First iteration
* Second iteration
* Third iteration
* Fourth iteration
* Fifth iteration
* Revert the temporary workaround
* beta.1 release only support Mac OSX 10.12+
Use 2.0 CLI/SDK of version `2.0.0-preview1-005952` to keep in sync with the .NET Core 2.0 //Build dogfood instructions. The version of `Microsoft.NetCore.App` in use is `2.0.0-preview1-002106-00`.
This change moves powershell to .NET Core 2.0. Major changes are:
1. PowerShell assemblies are now targeting `netcoreapp2.0`. We are using `microsoft.netcore.app-2.0.0-preview1-001913-00`, which is from dotnet-core build 4/4/17. We cannot target `netstandard2.0` because the packages `System.Reflection.Emit` and `System.Reflection.Emit.Lightweight`, which are needed for powershell class, cannot be referenced when targeting `netstandard2.0`.
2. Refactor code to remove most CLR stub types and extension types.
3. Update build scripts to enable CI builds. The `-cache` section is specified to depend on `appveyor.yml`, so the cache will be invalidated if `appveyor.yml` is changed.
4. Ship `netcoreapp` reference assemblies with powershell to fix the issues in `Add-Type` (#2764). By default `Add-Type` will reference all those reference assemblies when compiling C# code. If `-ReferenceAssembly` is specified, then we search reference assemblies first, then the framework runtime assemblies, and lastly the loaded assemblies (possibly a third-party one that was already loaded).
5. `dotnet publish` generates executable on Unix platforms, but doesn't set "x" permission and thus it cannot execute. Currently, the "x" permission is set in the build script, `dotnet/cli` issue [#6286](https://github.com/dotnet/cli/issues/6286) is tracking this.
6. Replace the use of some APIs with the ones that take `SecureString`.
7. osx.10.12 is required to update to `netcoreapp2.0` because `dotnet-cli` 2.0.0-preview only works on osx.10.12.
8. Add dependency to `System.ValueTuple` to work around a ambiguous type identity issue in coreclr. The issue is tracked by `dotnet/corefx` [#17797](https://github.com/dotnet/corefx/issues/17797). When moving to newer version of `netcoreapp2.0`, we need to verify if this dependency is still needed.
The package 'libpsrpclient' was replaced by 'psrp' in #3271. However, this was missed when moving from project.json to MSBuild. This change makes this update.
- FullCLR build is disabled in this change.
- FullCLR build related functionalities in `build.psm1` and `AppVeyor.psm1` are disabled. They are not cleaned up from `build.psm1` and `AppVeyor.psm1` yet. We need to adopt .NET Core 2.0 to verify the portable module concept, and if that works well, we will remove the Windows PowerShell source code and clean up our scripts.
- `dnxcore50` and `portable-net5+win8` target framework monikers are removed.
- Dependency on `Microsoft.NETCore.Portable.Compatibility` is removed. It's not necessary, but it may come back when we work on supporting the `portable module`. Its necessity can be reviewed at that time.
- I didn't spend the time to try building powershell in Visual Studio 2017. We should have a separate issue for that. It's tracked by #3400
The `TypeCatalogParser` project is replaced by a MSBuild target to gather the dependency information.
Due to .NET Core SDK issue [#1021](https://github.com/dotnet/sdk/issues/1021), our meta-package project `Microsoft.PowerShell.SDK` starts to generate an empty assembly during the build and that results in an empty assembly `Microsoft.PowerShell.SDK.dll` appear in `publish` folder and in `.deps.json` file. We cannot simply remove the assembly because it's now part of the TPA, and removing it will cause powershell to crash at startup. We have to live with this empty assembly until that .NET Core SDK issue is fixed. It's tracked by #3401.
Our assembly cache contains assemblies that are explicitly loaded by powershell egine, such as via module loading or the assembly entries from InitialSessionState. We should search it before searching all loaded assemblies to give preference to resolve a type against the assemblies contained in the cache, so that in case there is a conflict, we might have a preferred assembly to use for a type resolution.
Changes:
- Search from context.AssemblyCache.Values before search from all loaded assemblies.
- Skip assemblies that we already searched and found no matching type.
- Skip checking PS types kept in the scope and type accelerators when it's not necessary.
Get-Content /etc/os-release returns an array of strings, which get feeded to ConvertFrom-StringData. Therefore, $LinuxInfo is actually an array of Hashtable instances before this change. Property access on $LinuxInfo like $LinuxInfo.ID happens to work because powershell supports accessing a property or calling a method on a collection of object, but it's inefficient.
After this fix, $LinuxInfo will be a Hashtable, and it's efficient to access its members.
* Refactor 'FindMatchingHandler' to get rid of recursion
Search state can be saved and used throughout the searching process without the recursion.
* Rank the exception types based on how specific they are
* Refactor code to find handler based on the rank of given handlers
When a PS class is defined in a module and the module gets reloaded, the class would still use the SessionState from the old module for execution, and thus it doesn't reflect changes to the module state during the reload. This fix is to make sure we always use the current EngineSessionState for PS class execution.
* Remove crossgen'ed IL assemblies and then rename NI assemblies to the same as the corresponding IL images.
This is to make sure:
1. The dependency existence check by powershell.exe will pass for all TPAs.
2. CoreCLR by default will load the NI images even though it's using names of the IL images.
* Update the comment to make it easy to understand
* Improve Start-PSPackage to support zip package, downlevel packages for win-plat. Fix the version issue and package name for creating msi and appx
* Address comments about New-PSOptions
* Update Start-PSPackage to not show the warning from New-PSOptions
Dongbo Wang