Ticket #432 (closed bug: fixed)
Running `make fulltest` does not test fully.
Description
Target "fulltest" behaves in a less-than-useful fashion; if any of the groups of tests fail, then none of the later groups will run at all.
For example, if the "doc_tests" group fails, then "manifest_tests" will not run.
Since all the variations of the runcore tests (which rarely *all* pass) are ordered first, testers who rely on the completeness of target "fulltest" may have *never* run the later-ordered meta-quality tests.
Target "fulltest" should ignore errors in its sub-targets. I think that fulltest was *intended* to do just that, but the single dash before the single "make" command is a mis-coding of the concept. A separate "make" for each sub-target is needed, each preceded with its own error-ignoring dash.