Revert "core.thread.refactor Part 1"

[wilzbach]

Reverts #2821

[rainers]

I suspect the merge on Azure fails because the PR is from the dlang repo. Should we still force-merge this PR?

[rainers]

Removed auto-merge, doesn't seem to be the only problem. #2825 runs the Azure tests with the revert, but still fails.

[wilzbach]

@rainers the error happens when trying to clone DMD:

Cloning into '../dmd'...
fatal: Remote branch revert-2821-osthread not found in upstream origin
##[error]Cmd.exe exited with code '128'.

The problem is here: https://github.com/dlang/druntime/blob/master/azure-pipelines.yml#L43 :

  DMD_BRANCH:
    Parsing expression: <coalesce(variables['System.PullRequest.TargetBranch'], variables['Build.SourceBranchName'], 'master')>
    Evaluating: coalesce(variables['System.PullRequest.TargetBranch'], variables['Build.SourceBranchName'], 'master')
    Expanded: coalesce(Null, 'revert-2821-osthread', 'master')
    Result: 'revert-2821-osthread'

Maybe the environment variable SYSTEM_PULLREQUEST_TARGETBRANCH (see e.g. https://github.com/YakDriver/azure-pipelines-environment-variables#system_pullrequest_targetbranch) is defined?

[wilzbach]

It's weird that variables['System.PullRequest.TargetBranch'] is Null though:

The branch that is the target of a pull request. For example: refs/heads/master. This variable is initialized only if the build ran because of a Git PR affected by a branch policy.

https://docs.microsoft.com/en-us/azure/devops/pipelines/build/variables?view=azure-devops&tabs=yaml#system-variables

Maybe it's null, because it was re-triggered manually?

[rainers]

I suspect it makes a difference whether the same repo is used for the PR and the target and then some variables behave differently, We should add printing the environment variables to the yml as in the dmd version.

The revert alone didn't help, though, see #2825 . So resolving this merge issue could be in another PR.

[rainers]

So resolving this merge issue could be in another PR.

It seems the branch isn't built as a PR, but a separate branch. I guess we should configure only master and stable as "triggers": https://docs.microsoft.com/en-us/azure/devops/pipelines/yaml-schema?view=azure-devops&tabs=schema#triggers

[codecov]

Codecov Report

Merging #2822 into master will decrease coverage by <.01%.
The diff coverage is 92.51%.

@@            Coverage Diff             @@
##           master    #2822      +/-   ##
==========================================
- Coverage   74.56%   74.55%   -0.01%     
==========================================
  Files         161      160       -1     
  Lines       17147    17147              
==========================================
- Hits        12785    12784       -1     
- Misses       4362     4363       +1

Impacted Files	Coverage Δ
src/core/internal/spinlock.d	77.77% <ø> (ø)	⬆️
src/core/thread.d	82.97% <92.51%> (ø)
src/test_runner.d	69.69% <0%> (-3.04%)	⬇️
src/core/sync/semaphore.d	87.34% <0%> (-1.27%)	⬇️
src/rt/aaA.d	88.01% <0%> (-0.28%)	⬇️
src/gc/impl/conservative/gc.d	79.22% <0%> (-0.07%)	⬇️
src/core/sync/rwmutex.d	91.22% <0%> (+0.58%)	⬆️
src/core/sync/condition.d	93.91% <0%> (+1.73%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0dfe3e9...80aa9aa. Read the comment docs.