Adjust testURLBasedPluginAPI logging checks

[neonichu]

It seems like there's some flakiness with these log statements on Linux, but since they're not material for the test (which mostly wants to test that the build succeeds), we can remove the assertions for them for now to avoid blocking CI.

[neonichu]

@swift-ci please test

[neonichu]

Hm, interesting, seems like a lot of tests are failing on Linux self-hosted?

Test Case 'BuildToolTests.testBuildStartMessage' started at 2023-12-18 17:42:12.384
**** FAILURE EXECUTING SUBPROCESS ****
output: 
stderr: error: 'dependencyresolution_internal_simple': Invalid manifest (compiled with: ["/home/build-user/latest_toolchain/swift-DEVELOPMENT-SNAPSHOT-2023-12-17-a-ubuntu20.04/usr/bin/swiftc", "-vfsoverlay", "/tmp/TemporaryDirectory.8iHWxo/vfs.yaml", "-L", "/home/build-user/swiftpm/.build/x86_64-unknown-linux-gnu/debug", "-lPackageDescription", "-Xlinker", "-rpath", "-Xlinker", "/home/build-user/swiftpm/.build/x86_64-unknown-linux-gnu/debug", "-swift-version", "4.2", "-I", "/home/build-user/swiftpm/.build/x86_64-unknown-linux-gnu/debug", "-package-description-version", "4.2.0", "-module-cache-path", "/home/build-user/swiftpm/.build/x86_64-unknown-linux-gnu/debug", "/tmp/DependencyResolution_Internal_Simple.jpYwHJ/DependencyResolution_Internal_Simple/Package.swift", "-Xfrontend", "-disable-implicit-concurrency-module-import", "-Xfrontend", "-disable-implicit-string-processing-module-import", "-o", "/tmp/TemporaryDirectory.p2FAVq/dependencyresolution_internal_simple-manifest"])
error: fatalError
/tmp/DependencyResolution_Internal_Simple.jpYwHJ/DependencyResolution_Internal_Simple/Package.swift:3:8: error: no such module 'PackageDescription'
import PackageDescription
       ^

[neonichu]

Confirmed here that the issue is present in baseline and it seems like macOS is passing. Since builds were passing on Friday, most likely culprit seems to be the toolchain?

[neonichu]

Interesting that we're not seeing the issue on Linux smoke test.

[tomerd]

@atrick @neonichu feels like merging this papers over something else that has changed in the compiler / driver and may be a problem

[neonichu]

@tomerd it's a new test, so likely not any change, but I agree that it is interesting to figure out why we have inconsistent logging on Linux