Not using fixed path for valgrind in Makefile

Just replace "$(HOME)/pyston_deps" with "$(DEPS_DIR)" to make things
keep going well when the dependencies are placed elsewhere.

Turn on some more warnings, though the ones that don't have many false positives aren't very helpful

I think I generated an invalid pyc file by ctrl-C'ing while it was
writing out a file, and it took me a while to figure out why the
parser kept crashing.
Ideally this should be a full checksum.

This means we can have non-consecutive block indices, which
is a prereq for things like merging blocks.
Not sure how worth it this was

Add an extra marker to libunwind so that pyston can ensure it's running on a patched libunwind version

LLVM had some breaking changes that I didn't notice; on the plus
side this new range-iteration style is much nicer.

Removing hardcoded filename in ast.cpp

Add the ability to link against a debug build of libunwind for debugging unwinding issues, and some basic instructions on how to do so

Merge in the changesets that were missed:
- type recording
- better polymorphic IC performance
- build system improvements
- update to what was llvm trunk

Fixing the subtle "bug" apparently makes pyston fail some tests, so I'm backing out on this. It doesn't make much sense that the reverse of NotEq is NotEq, and the reverse of Eq is Eq... this needs to be rechecked by someone that knows the code better than me.

- Fixed subtle bug in getReverseOpName(): the reverse of AST_TYPE::NotEq and the reverse of AST_TYPE::Eq were wrong (ironically, they were reversed!)

- Added zsh dependency
- Fixed subtle typo in LLVM + clang section

Found some bugs in the name analyzer around listcomps; got the unittests back up and running so I can write some more

It all feel pretty hacked together though... definitely not
confident that I've gotten all of them.

I guess I had gotten those instructions from somewhere,
that said to just clone from trunk.
I ran into an issue installing gold through the instructions
and thought it might have been that, but I think it was the
bison issue I mention now in the instructions.
Probably better to leave it at a real release instead of tracking
trunk and having every run of the Pyston install giving a different
output.

Should support:
- multiple comprehensions
- multiple if conditions
- nested control flow expressions
- OSR'ing from the list comprehension

Though it tends to hit the OSR bug in the previous commit.

Some extra changes that could have been split out:
- use pointers-to-const instead of references-to-const for attribute-name passing,
  to make it harder to bind to a temporary name that will go away.
- add a 'cls_only' flag to getattr / getattrType to not have to special-case clsattrs
  (or simply get it wrong, in the case of getattrType)