[libc] Configure the 'printf' implementation for the GPU

[jhuber6]

Summary

The GPU build is attempt to build with printf support. Currently with
D158246 we can compile snprintf and run it. However, this introduces
problems with the long double table and backtracking. This patch adds
the necessary configs to disable the long double table and support for
%n.

The size of the generated binaries is much smaller with this applied so
it's a good first step. We also will want %n disabled for out handling
of printf as it required backtracking.

[JonChesterfield]

Can we do the long double part in the source?

I'm thinking that lots of targets have double and long double as the same type, so we should be able to branch at compile time to cast long double to double and go through the double code path instead

[jhuber6]

In D158435#4603994, @JonChesterfield wrote:

Can we do the long double part in the source?

I'm thinking that lots of targets have double and long double as the same type, so we should be able to branch at compile time to cast long double to double and go through the double code path instead

I brought that up with @michaelrj previously. The GPU build already sets a macro in libc called LONG_DOUBLE_IS_DOUBLE which should also disable the massive long double table. He said he'd work on it, but in general we should probably configure the other stuff anyway.

[sivachandra]

Can you use https://reviews.llvm.org/D159067?

[jhuber6]

In D158435#4623931, @sivachandra wrote:

Can you use https://reviews.llvm.org/D159067?

Yeah that would probably be good for sprintf stuff. Thinking disabling float would be viable as well.

libc/src/stdio/printf_core/CMakeLists.txt

+set(printf_config "")
+if(LIBC_TARGET_ARCHITECTURE_IS_GPU)
+  list(APPEND printf_config -DLIBC_COPT_PRINTF_DISABLE_WRITE_INT)
+  list(APPEND printf_config -DLIBC_COPT_PRINTF_DISABLE_INDEX_MODE)
+else()
+  list(APPEND printf_config -DLIBC_COPT_FLOAT_TO_STR_USE_MEGA_LONG_DOUBLE_TABLE)
+endif()
 
 add_header_library(
   core_structs
   HDRS
     core_structs.h
   DEPENDS
     libc.src.__support.CPP.string_view
     libc.src.__support.FPUtil.fp_bits
DEPENDS
     libc.src.__support.arg_list
     libc.src.__support.ctype_utils
     libc.src.__support.str_to_integer
     libc.src.__support.CPP.bit
     libc.src.__support.CPP.optional
     libc.src.__support.CPP.string_view
     libc.src.__support.CPP.type_traits
     libc.src.__support.common
+  COMPILE_OPTIONS
+    ${target_config}
 )
 
 add_object_library(
   mock_parser
   SRCS
     parser.cpp
   HDRS
     parser.h
DEPENDS
     libc.src.__support.FPUtil.rounding_mode
     libc.src.__support.common
     libc.src.__support.libc_assert
     libc.src.__support.uint
     libc.src.__support.uint128
     libc.src.__support.integer_to_string
     libc.src.__support.float_to_string
   COMPILE_OPTIONS
-    -DLIBC_COPT_FLOAT_TO_STR_USE_MEGA_LONG_DOUBLE_TABLE
+    ${target_config}
 )
 
 
 add_object_library(
   printf_main
   SRCS
     printf_main.cpp
   HDRS