Is this necessarily leading to my compilation error?
It's not supposed to, but might in reality. One of the jobs of configure
scripts is to try and work around missing/broken functionality in the build environment. Sometimes some of the tests that are run in configure
might not even be directly relevant to your build. In a perfect world, the code being configured should be written to work if either a yes
or a no
is returned for any test (or fail the build, if it somehow doesn't have a work around).
In other words, do I have to have all 'yes' for it to properly (i.e. not just by complete luck) compile?
No.