Enable PVP flow for nuget.client

[NikolaMilosavljevic]

Contributes to dotnet/source-build#3043

Enables PVP flow for nuget.client repo. Includes a patch for the repo-side work: NuGet/NuGet.Client#5207

[mthalman]

I question whether we really gain much by enabling this now (which requires a patch) versus enabling it once the nuget change flows to installer. The friction here is that the existence of this patch will cause the build for the dependency flow with the nuget change to fail.

[NikolaMilosavljevic]

I question whether we really gain much by enabling this now (which requires a patch) versus enabling it once the nuget change flows to installer. The friction here is that the existence of this patch will cause the build for the dependency flow with the nuget change to fail.

That is generally true for any patch. This one was requested today. cc @MichaelSimons

[mthalman]

That is generally true for any patch. This one was requested today. cc @MichaelSimons

Yes, I know. Some patches are more important than others. I'm questioning whether this is that important.

[MichaelSimons]

I question whether we really gain much by enabling this now (which requires a patch) versus enabling it once the nuget change flows to installer. The friction here is that the existence of this patch will cause the build for the dependency flow with the nuget change to fail.

That's a fair question. NuGet is very sporadic on when it flows in. It can take a long time for these changes to flow in. This is the reason I asked for the patch. The patch itself has a low risk for conflict so the only overhead is the dependency flow.

The benefit is that we can complete the pvp work and validate the complete e2e. Validating each repo independently is not enough.

[NikolaMilosavljevic]

9 source-build smoke-tests failed - likely unrelated as my local VMR build was clean, but with an older base sha.

[NikolaMilosavljevic]

9 source-build smoke-tests failed - likely unrelated as my local VMR build was clean, but with an older base sha.

Hmm, this could be related as I don't see any failures in other CI runs. Likely a transitive dependency on a reference package. Investigating...

[NikolaMilosavljevic]

Fixed the source-build issue with 8078c5a

[mthalman]

In order to reduce the chance of patch conflicts, this change should be removed since it provides no functional behavior.

[NikolaMilosavljevic]

Yeah, makes sense. I'll remove it from the patch.

[NikolaMilosavljevic]

Fixed with 4604dec