That said, I don't think the right solution to avoiding problems like this is a "head in the sand" approach. If there are real performance issues, they should be acknowledged and addressed. I definitely get the sense that a lot of Ruby zealots are trying to pretend that there is no problem, which is something that Java zealots tried to do years ago (and something that the people in control of feature requests still do quite a bit, using the "Nobody can microbenchmark the JVM, it's too smart!" excuse).
I don't know why this stuff gets people so emotional, though. You don't see carpenters have heated arguments about whether a screwdriver is better than a hammer; why do programmers get so emotionally attached to their tools, refusing to acknowledge the limitations and strengths of each particular one?
Required effort causes programmers to know less languages and to value the languages they already know higher as they've had to come up with a rationalization for learning the language. This rationalization hinders acknowledging the downsides of the languages they've learned.
Additionally, languages give us our vocabulary and solutions seen beautiful by a programmer of one language can seem incomprehensible or at least ugly to programmer of another language. This is at least something that I've personally observed in a Java-shop I work in.
Edit: I am specifically thinking about Ruby here. It seems quite neat and worth trying, but... Maybe I'm spoiled with polite communities.
The suggestion to provide patches is a simple displacement mechanism. A correct, reasoned evaluation is not predicated on also including patches, and a reasoned evaluator may have significant cause to not invest further effort in providing improvements to the evaluated software.