The 2nd commit reverts the commit 9f3b9c470c05b9a65e4f74ef8aeace792976fb69.
The 3rd commit is the main commit for this PR, setting the environment variable DFLTCC=0. When running test/**/test_*.rb" as a child process from the Ruby code in the Rakefile as a parent process. the environment variable can affect to the tests in the test file. I referred to the way of test task calling the test-internal, seeing the following ruby/openssl's Rakefile.
This PR is a rework for the https://github.com/ruby/zlib/pull/63, to make the workaround simple. Because I was inspired by the comment at https://github.com/ruby/spec/pull/1084#issuecomment-1739243422.
The PR has 3 commit.
The 1st commit is to improve
.travis.yml
printing the used zlib deb package version and other info.The 2nd commit reverts the commit 9f3b9c470c05b9a65e4f74ef8aeace792976fb69.
The 3rd commit is the main commit for this PR, setting the environment variable
DFLTCC=0
. When runningtest/**/test_*.rb"
as a child process from the Ruby code in theRakefile
as a parent process. the environment variable can affect to the tests in the test file. I referred to the way oftest
task calling thetest-internal
, seeing the following ruby/openssl'sRakefile
.https://github.com/ruby/openssl/blob/6b3dd6a372c5eabc88bf35a312937ee3e1a6a105/Rakefile#L21-L35
I confirmed that the Travis CI passing in my forked repository. When running 2 times, the total running times are 28 seconds and 31 seconds. So, it's faster than GitHub Actions around 2 minutes 30 seconds ~ 7 minutes .
@jeremyevans Could you review? The change is essentially only for the 3rd commit, as I was able to revert the 2nd commit without conflict.