aboutsummaryrefslogtreecommitdiff
path: root/tests/qapi-schema/alternate-empty.exit
AgeCommit message (Collapse)Author
2016-02-19qapi: Forbid empty unions and useless alternatesEric Blake
Empty unions serve no purpose, and while we compile with gcc which permits them, strict C99 forbids them. We happen to inject a dummy 'void *data' member into the C unions that represent QAPI unions and alternates, but we want to get rid of that member (it pollutes the namespace for no good reason), which would leave us with an empty union if the user didn't provide any branches. While empty structs make sense in QAPI, empty unions don't add any expressiveness to the QMP language. So prohibit them at parse time. Update the documentation and testsuite to match. Note that the documentation already mentioned that alternates should have "two or more JSON data types"; so this also fixes the code to enforce that. However, we have existing uses of a union type with only one branch, so the 2-or-more strictness is intentionally limited to alternates. Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1455778109-6278-3-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2015-10-12qapi: Add tests for empty unionsEric Blake
The documentation claims that alternates are useful for allowing two or more types, although nothing enforces this. Meanwhile, it is silent on whether empty unions are allowed. In practice, the generated code will compile, in part because we have a 'void *data' branch; but attempting to visit such a type will cause an abort(). While there's no technical reason that a degenerate union could not be made to work, it's harder to justify the time spent in chasing known (the current abort() during visit) and unknown corner cases, than it would be to just outlaw them. A future patch will probably take the approach of forbidding them; in the meantime, we can at least add testsuite coverage to make it obvious where things stand. In addition to adding tests to expose the problems, we also need to adjust existing tests that are meant to test something else, but which could fail for the wrong reason if we reject degenerate alternates/unions. Note that empty structs are explicitly supported (for example, right now they are the only way to specify that one branch of a flat union adds no additional members), and empty enums are covered by the testsuite as working (even if they do not seem to have much use). Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1443565276-4535-8-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>