Revert "URI-5.22"

[oalders]

This reverts commit fc3628d.

The following warnings were introduced:

Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 38.
Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 53.
Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 118.
Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 172.

@jackdeguest FYI ^^^

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (08b00a0) 87.38% compared to head (ec70d7f) 87.38%.
Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #134   +/-   ##
=======================================
  Coverage   87.38%   87.38%           
=======================================
  Files          50       50           
  Lines        1783     1783           
  Branches      462      462           
=======================================
  Hits         1558     1558           
  Misses         77       77           
  Partials      148      148           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jackdeguest]

This reverts commit fc3628d.

The following warnings were introduced:

Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 38.
Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 53.
Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 118.
Use of uninitialized value in string eq at /root/.cpan/build/URI-5.22-0/blib/lib/URI/_query.pm line 172.

@jackdeguest FYI ^^^

Maybe we could add instead if (defined($r) && Scalar::Util::reftype($r) eq "ARRAY") ?
That would resolve this issue, don't you think ?

[oalders]

Yes. That would work. We should also check that the code paths which generated the warnings are covered by tests. Possibly we are not failing tests when warnings occur. Not sure. I'm on my phone right now. 😀