CI: just a false positive in fedora-ci.koji-build.rpminspect.static-analysis

CI failed in fedora-ci.koji-build.rpminspect.static-analysis:

1. the "unicode" failure is a routine false positive

2. /usr/lib/clang/plugins/annobin-for-{llvm,clang}.so fail in annocheck - no problem in practice, those 32-bit libraries don't ship

The required fedora-ci.koji-build.tier0.functional testing failed in CI but it was because of some infrastructure problem. I re-run it in a virtual machine and it passed.

"baseos-qe.koji-build.scratch-build.validation" failed when building kernel on s390x. The cause was an infrastructure error (https://kojipkgs.fedoraproject.org//work/tasks/2249/115072249/checkout.log). I rerun the build successfully: https://koji.fedoraproject.org/koji/taskinfo?taskID=115107590.

Completed my side run of baseos-qe.koji-build.scratch-build.validation. Consider CI as passed.

CI results:

baseos-qe.koji-build.scratch-build.validation - some infrastructure problems, I'm running the test aside, still in progress

fedora-ci.koji-build.rpminspect.static-analysis - false positive

CI tests:

baseos-qe.koji-build.scratch-build.validation - some infrastructure problems, I ran it aside and it worked

fedora-ci.koji-build.rpminspect.static-analysis - false positive

CI failures:

• known fedora-ci.koji-build.rpminspect.static-analysis - "unicode" is false positive, "annocheck" has been filed as https://bugzilla.redhat.com/show_bug.cgi?id=2264000

• baseos-qe.koji-build.scratch-build.validation - infrastructure error, i'm running it manually right now, will take a couple of hours, will report back once done

CI failures:

• known fedora-ci.koji-build.rpminspect.static-analysis - false positive

• baseos-qe.koji-build.scratch-build.validation - infrastructure error, i'm running it manually right now, will take a couple of hours, will report back once done

CI failures:

• known fedora-ci.koji-build.rpminspect.static-analysis - false positive

• baseos-qe.koji-build.scratch-build.validation - infrastructure error, i'm running it manually right now, will take a couple of hours, will report back once done

CI failures:

fedora-ci.koji-build.rpminspect.static-analysis - https://bugzilla.redhat.com/show_bug.cgi?id=2174301 continues

fedora-ci.koji-build.tier0.functional - it's the test's error, already fixed in git

baseos-qe.koji-build.scratch-build.validation failed because of an infrastructure error, i ran it manually and it passed

baseos-qe.koji-build.scratch-build.validation failed because of an infrastructure error, i ran it manually and it passed

Comments on the CI failures:

fedora-ci.koji-build.rpminspect.static-analysis: not new, tracked at https://bugzilla.redhat.com/show_bug.cgi?id=2174301

fedora-ci.koji-build.tier0.functional: Broken test. Once time permits, I'll try to fix it and submit a PR. I checked manually what the test seems to achieve:

[root@vm-10-0-184-119 fibonacci]# rpm -qa | grep tbb- | sort
python3-tbb-2021.11.0-4.fc40.x86_64
tbb-2021.11.0-4.fc40.x86_64
tbb-bind-2021.11.0-4.fc40.x86_64
tbb-devel-2021.11.0-4.fc40.x86_64
tbb-doc-2021.11.0-4.fc40.x86_64

[root@vm-10-0-184-119 ~]# koji download-build --arch=src tbb-2021.11.0-4.fc40
[root@vm-10-0-184-119 ~]# dnf -y builddep ./tbb-2021.11.0-4.fc40.src.rpm 
[root@vm-10-0-184-119 ~]# rpm -i ./tbb-2021.11.0-4.fc40.src.rpm 
[root@vm-10-0-184-119 ~]# rpmbuild -bp ~/rpmbuild/SPECS/tbb.spec 
[root@vm-10-0-184-119 ~]# cd rpmbuild/BUILD/oneTBB*/examples/test_all/fibonacci
[root@vm-10-0-184-119 fibonacci]# cmake .
[root@vm-10-0-184-119 fibonacci]# cmake --build .

[root@vm-10-0-184-119 fibonacci]# ./fibonacci
TEST PASSED
[root@vm-10-0-184-119 fibonacci]# ./fibonacci 1000
Fibonacci numbers example. Generating 1000 numbers..
Serial loop - in 2.105444 msec
Serial matrix   - in 8.456915 msec
Serial vector   - in 145.668340 msec
Serial queue    - in 232.516847 msec

Threads number is 1
Shared serial (mutex)       - in 15.073313 msec
Shared serial (spin_mutex)  - in 18.422482 msec
Shared serial (queuing_mutex)   - in 31.774402 msec
Shared serial (Conc.HashTable)  - in 694.475326 msec
Parallel pipe/queue     - in 278.503233 msec
Parallel reduce         - in 29.575960 msec
Parallel scan           - in 13.254988 msec

Threads number is 2
Shared serial (mutex)       - in 42.154892 msec
Shared serial (spin_mutex)  - in 41.053918 msec
Shared serial (queuing_mutex)   - in 110.920034 msec
Shared serial (Conc.HashTable)  - in 690.518567 msec
Parallel pipe/queue     - in 407.470438 msec
Parallel reduce         - in 17.382027 msec
Parallel scan           - in 15.319373 msec

Threads number is 4
Shared serial (mutex)       - in 47.196572 msec
Shared serial (spin_mutex)  - in 39.449660 msec
Shared serial (queuing_mutex)   - in 112.109725 msec
Shared serial (Conc.HashTable)  - in 794.848958 msec
Parallel pipe/queue     - in 496.953495 msec
Parallel reduce         - in 16.953068 msec
Parallel scan           - in 15.066879 msec
Fibonacci number #1000 modulo 2^64 is 817770325994397771

[root@vm-10-0-184-119 fibonacci]# echo $?
0

Looks OK to me, the fedora-ci.koji-build.tier0.functional failure can be ignored.

CI testing failures can be ignored:

fedora-ci.koji-build.rpminspect.static-analysis reports just one known false positive

fedora-ci.koji-build.rpmdeplint.functional looks to be completely messed up (rawhide repos???), I check manual dnf upgrade/downgrade on f39 between 12.35-1.fc39 and 12.36-1.fc39 and they worked just fine

CI testing failures can be ignored:

fedora-ci.koji-build.rpminspect.static-analysis reports just one known false positive

fedora-ci.koji-build.rpmdeplint.functional looks to be completely messed up (rawhide repos???), I check manual dnf upgrade/downgrade on f38 between 12.30-1.fc38 and 12.36-1.fc38 and they worked just fine

Tha failed CI results can be ignored:

fedora-ci.koji-build.rpminspect.static-analysis - /unicode - false positive (CVE-2021-42574 mitigation tests)

fedora-ci.koji-build.rpmdeplint.functional - looks like the test's issue

baseos-qe.koji-build.scratch-build.validation - infrastructure error, I rerun all the scratch builds manually

CI failures can be ignored, IMHO: fedora-ci.koji-build.rpmdeplint.functiona - weird, looks to me like an issue in the test itself fedora-ci.koji-build.rpminspect.static-analysis - /unicode is a false positive (CVE-2021-42574 mitigation tests)

fedora-ci.koji-build.tier0.functional is bound to fail in tests/Sanity/rebuild-wget. The test needs an update to handle the wget -> wget2-wget change; that will be done later. I've just did an equivalent check manually and it passed (https://pastebin.com/5vcgWW3n) so fell free to ignore this particular failure.

no regression found