I used to attend the Bay Area Functional Programmers meetings (apparently now defunct). At one point, I mentioned that I was having a hard time applying FP to my work, which requires lots of iterative updates to vectors that represent states in physical systems.
The general reaction was not that I simply wasn't committed enough to FP. Rather, it was a very thoughtful version of "Yeah I can see where pure FP probably isn't a good fit".
With enough work, you can accomplish any task in any language. The only thing you can't do is force yourself to use something you don't like. So I commend you for giving it a try, but I certainly don't begrudge you for not liking it or using it for everything.
Programming language proselytization is one of the most obnoxious things about our industry. We love to cloak our feelings in rational language, but at the end of the day, all of the design decisions in a language are choices that have upsides and downsides. Some of these upsides are going to be valuable to a particular user or a particular problem, a lot of them are going to be irrelevant and a few will be detrimental. The factor that determines success is absolutely the individual and not the language.
Believing your language is the best must help motivate people to use it, but it is definitely an irrational belief, and not a technical fact. Advertising language differences as factual improvements when they are, at best, assumptions derived from untested principles must be doing for our industry what snake oil has done for Western medicine.
Adherents of religion X will tell you how it changed their life, and it can change yours too, if you just believe hard enough.
When you investigate, find problems or just say "it's not for me", it's "your fault" for doing it wrong, or with a bad attitude, or you didn't talk to the right people, or you'll "grow up" later and 'see the light'.
"it's not for me" is not a valid excuse, it means you cannot handle the tool. This is your fault. A valid excuse is "it's not for my application domain".
I understand this has been done a number of times.
I think if it had metabolized, that would mean your body had burnt up the weirdness and gotten rid of it.
C++ certainly has it's faults, and some of them are quite painful. However, I find there are things I can do nicer in Haskell than C++, but some jobs call for an XORed linked list, or implementing trailing by knowing my underlying types are C and just calling memcpy, and I can't do that in Haskell.
At http://minisat.se/ , you will find Minisat, an (in my opinion) neat, easy to read(ish) and highly efficient C++ program. I've seen numerous people try to turn it into Haskell and all they have ended up with is something less readable, and much slower.
After spending quite a bit of time learning Haskell, seeing a number of advantages and disadvantages, and being unconvinced and staying with C++, the assumption is I didn't "do enough".
Of course, I also know a number of lovely functional programmers as well!