I'll have to surprise you then because it actually does work. Just as I had 
posted this I removed the pop instruction and looked at the IL in ildasm,
which actually showed the right IL (ILSpy didn't).
So I tried running the code again and it worked just as it was supposed to. 
So for (static) managed calls, calli with generic arguments seems to be ok.

I've done some rough performance benchmarks which show that this current 
method is actually about twice as slow as a delegate call (even with IL 
that is a bit more
optimized and AggresiveInlining enabled on the method).
This surprised me, as some performance tests on non-generic calli 
implementations actually showed a general increase in speed.

Anyways, this specific issue seems to be solved. So for future reference: 
It does work, I just don't know why it didn't in the first place.

-- 
-- 
--
mono-cecil
--- 
You received this message because you are subscribed to the Google Groups 
"mono-cecil" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to