This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
level3/core/comparedocumentposition03, level3/core/comparedocumentposition05, level3/core/comparedocumentposition16, level3/core/comparedocumentposition23, level3/core/comparedocumentposition33: These test an implementation-defined ordering. They correctly require a result including IMPLEMENTATION_SPECIFIC, but also require that the one particular node FOLLOWs the other. This can't be relied on as that's the point of implementation specificity. Only the consistency of the result should be checked for if anything (ie. does the reverse test give the opposite result for FOLLOWING/PRECEDING). There might be more tests with the same problem here, that randomness prevented from appearing whilst I was using the tests.
Created attachment 96 [details] Example nodecomparedocumentposition05 fix This fixes test 05 only - I haven't done the others as the solution here is a really ugly arithmetic hack. Maybe someone more familiar with TSML could find a better way of doing it.
Created attachment 169 [details] Used bitmasking to eliminate over specification