Except for the few that were actually working, which became
normal tests again.

Implement string.split(sep)

TraceStack: when visiting a range preallocate the elements.

Fix wrong if check: causing allocs smaller than 16byte to end up in the 32byte bucket

Improves performance of raytrace.py from about 10secs to 7secs.

Doing it "right" is more complicated than I thought,
so it's a little bit hacky right now.

Specify that lzma needs to be installed (doesn't really, but to keep things simple), and try to support running pyston not from src/

fix several minor issues found during installing and building

gtest does not compile successfully with "-Wundef", so the following
error will be produced while running unittests:

$ make run_gc_unittests

<snip>
 /path-to-deps/gtest-1.7.0/include/gtest/internal/gtest-port.h:288:6: error: 'GTEST_OS_WINDOWS' is not defined,
       evaluates to 0 [-Werror,-Wundef]
 #if !GTEST_OS_WINDOWS
     ^
 /path-to-deps/gtest-1.7.0/include/gtest/internal/gtest-port.h:299:5: error: 'GTEST_OS_LINUX_ANDROID' is not defined,
       evaluates to 0 [-Werror,-Wundef]
 #if GTEST_OS_LINUX_ANDROID
     ^
<snip>

The workaround for this issue is using '-isystem'.

Result of "import time; print time" can be various from different
distributions, for example:

$ python -c "import time; print time"
<module 'time' from '/usr/lib64/python2.7/lib-dynload/timemodule.so'>

which is diff from the result of "<module 'time' (built-in)>".

So test with result of "print type(time)" instead, which should be
more robust.

The following error was produced when -llzma is not specified:

$ make check
cc -O2 -fPIC -Wimplicit -I../include -c ../test/test_extension/test.c -o ../test/test_extension/test.o -g
cc -shared ../test/test_extension/test.o -o ../test/test_extension/test.so -g
pyston: Compiling runtime/inline/link_forcer.o.bc
pyston: Compiling ../tools/publicize.o
pyston: Linking ../tools/publicize
elfxx.c:193: error: undefined reference to 'lzma_stream_footer_decode'
elfxx.c:200: error: undefined reference to 'lzma_index_buffer_decode'
elfxx.c:204: error: undefined reference to 'lzma_index_size'
elfxx.c:209: error: undefined reference to 'lzma_index_end'
elfxx.c:206: error: undefined reference to 'lzma_index_uncompressed_size'
elfxx.c:209: error: undefined reference to 'lzma_index_end'
elfxx.c:277: error: undefined reference to 'lzma_stream_buffer_decode'
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [../tools/publicize] Error 1

Doesn't handle packages or anything other than importing a .py file
(ex zip import, .so, .pyc), but has some basic support for
sys.path and sys.modules.

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