Release 6.0.0-beta.1

[thomaslevesque]

• Address all coding and style violations detected during a Release build (send a regular pull request which must be merged before continuing)
• ensure completeness of draft release notes, including non-owner contributors, if any (move release notes forward from any pre-releases to the current release)
• before releasing 6.0.0, check integration package compatibility
  • AutoFixture.AutoFakeItEasy - @blairconrad
    broken by ReturnValue not set ReturnValue not set #1696, released in 5.4.1. SetReturnValue is no longer found by FakeObjectCall.
  • Machine.Fakes.FakeItEasy
  • Autofac.Extras.FakeItEasy
  • EntityFrameworkTesting.FakeItEasy
• run tools\prepare_release.cmd next x.y.z vNext to:
  • change "vNext" names to appropriate x.y.z version for the next GitHub Release, milestone, and release checklist (this issue)
  • create a new draft GitHub Release, milestone, and release checklist
• tag master with the new version number and push to upstream. This will cause the build server to
  • push the nupkgs to nuget.org,
  • attach all nupkgs from the tagged build to the GitHub Release,
  • add a comment to each issue and pull request in the release, saying it was published in this release
• wait for NuGet package validation
• de-list pre-release NuGet packages if present
• publish the GitHub release
• tweet, mentioning contributors, and post link as comment here for easy retweeting ;-)
• post tweet in Gitter
• add a note in each issue that has a significant contribution from a community member, thanking the contributor
• close this issue
• if this is not a pre-release, close the milestone for this release

[blairconrad]

#1690 is nearly complete, and this was the last issue that we'd listed as a must-have for the 6.0.0 release. Once it's merged, I'd like to issue a beta.1 release, if that's okay.
We could try upgrading the Octokit and NuGet versions before this. Worst case, the release fails.

[thomaslevesque]

Once it's merged, I'd like to issue a beta.1 release, if that's okay.

Works for me.

We could try upgrading the Octokit and NuGet versions before this. Worst case, the release fails.

I'm not very worried about NuGet. Octokit, I don't know, but I guess that if the code still compiles, the behavior probably won't have changed. Is there a benefit to upgrading, or is it just for the sake of upgrading?

[blairconrad]

is it just for the sake of upgrading

Yeah. Which is part of the reason I hadn't pushed it yet. I'm content to leave them as they are until we run into trouble.

[thomaslevesque]

I don't mind either way. If you feel like upgrading, go ahead

[blairconrad]

Nah, it's good. My stress level rises every time we jiggle the deploy tool.

[thomaslevesque]

https://twitter.com/fakeiteasyfx/status/1207622933574823936

[TimLovellSmith]

FYI I tried out the 6.0.0 beta, and found no regressions, tests, continued to run pretty fast, but .... our overall build time went up noticably! My hunch is that this is most likely to be a size-of-files, times number of times we have to copy/load/crossref them thing related to Castle.Core packaging.

[blairconrad]

Glad there were no regressions, @TimLovellSmith. I'm a little surprised about the build time changes. Which TFM are you using?