Marked a number of failures that seem to be the fault of implementations, instead of Alexandria, as expected failures. Originally, tried to use
rt:*expected-failures*, but that didn't work: expected failures still fail the test suite, meaning that they also fail our GitLab actions tests. So I had to replace the use of
rt:*expected-failures* with reader macros that disable individual tests.
With this merged, Alexandria should build cleanly on GitLab. Hoping that this makes future enhancements and maintenance easier.