I was learning from @jdelong and he said that you should use
double quotes for local includes and angle brackets for library
includes. I asked why our code was the way it was, and he said he wanted
to clean it up. I beat him to it :)
Conflicts:
hphp/runtime/base/server/admin_request_handler.cpp
hphp/runtime/vm/named_entity.h
Post-hphpc, declaring builtins as inline is useless, which obviates the
need for this layer.
This doesn't fully delete the file, because I wanted to keep mindless
parts and use-your-brain parts separate. This is the mindless part. The
remaining functions in noinline.cpp are no-inline wrappers around
(a) a function that isn't defined in an extension, call_user_func_array,
and (b) calling polymorphic is_* functions, which can probably mostly go
away now. But I wanted to exercise more care around those, so they'll
come in a followup diff.
gzinflate tries to do some kind of exponential growth that
hits the max string limit immediately when you try to deflate a string
larger than 128 megs. The logic is a little convoluted with a few
ways to bail, but rather than try to really fix it I just made it
behave close to the same on small strings, but use a smaller starting
factor on long ones and be willing to try to deflate it into a
max-sized string.
Per @mwilliams' suggestion, this is the first stage in a staggered approach to replacing int64 with int64_t. More precisely I inserted "typedef ::int64_t int64;" in util/base.h and dealt with the consequences.
While tracking down gcc 4.7.1 issues I tried a -O1 build,
and got uninitialized variable warnings. None of them is
real, and at -O3 the compiler can figure out that the variables
are never actually used uninitialized.
This change is mostly for FB internal organizational reasons.
Building is not effected beyond the fact that the target now
lands in hphp/hhvm/hhvm rather than src/hhvm/hhvm.
mwilliams