Microsoft.AspNetCore.Mvc.Razor.RuntimeCompilation 3.1.22 has broken dependencies on NuGet

[rjgotten]

The problem

The package Microsoft.AspNetCore.Mvc.Razor.RuntimeCompilation, version 3.1.22 in NuGet refers to unexisting packages:

• Microsoft.AspNetCore.Mvc.Razor.Extensions, version 3.1.22
• Microsoft.CodeAnalysis.Razor, version 3.1.22

To Reproduce

• Inspect the package
• Follow the listed non-existant dependencies
• This takes you to the 6.0.1 versions, as the 3.1.22 versions do not exist.

Note that an attempt to install the Microsoft.AspNetCore.Mvc.Razor.RuntimeCompilation, version 3.1.22 package to any project using .NET Core 3.1 will result in a failing build, because of the bad 6.0.1 CodeAnalysis dependency.

Note

This is a repeat of the same problem with version 3.1.21 reported last month in #38172.
Fix your release process, please. This is shameful.

At the very least incorporate a correct upper bound on the package dependencies, so they cannot automatically resolve versions that go with a later framework. Right now, the dependencies only specify a lower bound:

Under no circumstance whatsoever is it acceptable that when faced with a non-existing version, you allow NuGet to skip into the latest version. Actually; someone should look into that NuGet behavior. It giving you the 'next version up' might be a rational choice - still a stupid choice, if that's a new major version which breaks semantic versioning, but rational. However, what NuGet appears to be doing when the requested dependency doesn't exist is just flat out tossing up its arms and going: "Oh whatever; latest version it is!"

[TanayParikh]

cc/ @dotnet/aspnet-build, seems to be related to the aspnetcore-tooling -> razor-tooling migration (#38172).

[TanayParikh]

Hello @rjgotten, thanks for pointing this out. The packages should now be available, could you please give it a try?

[dougbu]

Sounds like the release process had a hiccup today but everything is now sorted. If you agree @rjgotten, please close this issue.

Hi @rjgotten. We have added the "Needs: Author Feedback" label to this issue, which indicates that we have an open question for you before we can take further action. This issue will be closed automatically in 7 days if we do not hear back from you by then - please feel free to re-open it if you come back to this issue after that time.

[rjgotten]

I've heard back from a co-worker who has verified that builds are succeeding again.
So; guess it's ok again for now.

Hopefully we won't be looking at a repeat again next month.