RE: [abc] Tracematch-Question: Symbols both negated and unnegated

Pavel Avgustinov wrote:
> Correct me if I'm wrong, but it seems to me that the regular pattern
> "a
> c* b" would be equivalent to the above, provided that b() and c() are
> disjoint. (If there are events matched by both b() and c(), define
> symbols b'() := b() && !c() and c'() := c() && !b(), and use the
> pattern "a c'* b'".)

Hmmm, I could really have seen this before :-) Looks like a sufficiently
equivalent version. Thanks!

Anyway, what is the general behaviour if the a state on the trace is matched
by "the last symbol on a tracematch" and at the same time a symbol which is
not contained in the tracematch? On the one hand, this should trigger an
advice, because the state is matched by the last symbol of the tm, but on
the other hand the fact that it is also matched by a symbol which is not
contained should prevent the advice from executing.

Eric

-- 
Eric Bodden
Chair I2 for Programming Languages and Program Analysis
RWTH Aachen University