We pass -MD -MF args to clang when doing `@cImport`, which
gives us a complete list of files that the C code read from.
Then we add these to the cache. So even when using `@cImport`
Zig's caching system remains perfect. This is a proof of concept
for the mechanism that the self-hosted compiler will use to
watch and rebuild files.
this prevents the situation where we determine the cache
manifest and write it, but then crash or otherwise error out
before putting the artifacts in the proper place.
now the artifacts will be in place because cache_release
happens after that step is done.
we need somewhere to put .o files and leave them while the user
executes their program, so that stack traces on MacOS can find
the .o files and get at the DWARF info.
if we try to clean up old global tmp dir files, first of all that's
a hard and complicated problem, and secondly it's not clear how
that is better than dumping the .o file inside zig-cache locally.
* add almost all the input parameter state to the hash
- missing items are the detected MSVC installation on Windows
and detected libc installation on POSIX
- also missing are C files and .h files that libclang finds
* artifacts are created in global cache directory instead of
zig-cache.
- exception: builtin.zig is still in zig-cache
* zig run uses the new cache correctly
* zig run uses execv on posix systems
