Current failures

[straight-shoota]

In preparation for releasing Crystal 1.6.1 I've tried to fix the failing tests. I managed to get some improvements in #29.

But there are still a couple of failures that are not trivial to solve here.

Most are warnings for positional parameter mismatch (which was introduced in Crystal 1.5.0)

• crystal-db (PR: Fix positional parameter warnings crystal-db#173)
• mint caused by old kemal dependency (PR: Update dependencies mint-lang/mint#551)
• crystal-patterns (PR: Fix warnings about positional parameter mismatch crystal-community/crystal-patterns#20)

These can easily be ignored (the tests succeed, failure is only caused by warnings).

Two failures are a bit more complex:

• lucky has ambiguous sources in the dependencies, a failing postinstall script and a missing constant. The dependency tree seems to be a huge mess. I'm inclined to ignore it until it becomes decently usable. In any case, all the problems seem to be home-made and there's no indication for regressions in the crystal or shards projects.
• crinja specs fail because of some mismatch in test and lib sources. The test file examples/config/config.cr seems to be checked out from master branch, but the shard source is the latest release 0.8.0. Not sure what exactly causes this, but it seems like the global shards override breaks the example test which is supposed to pull in the shard code from a local path (parent directory). I'm pretty sure this is a fault in the test-ecosystem setup. Crinja tests succeed when run manually.