[libc] [llvm] [libc] Move printf long double to simple calc (PR #75414)

[Guillaume Chatelet via llvm-commits]

================
@@ -13,18 +13,27 @@
 
 #include "src/__support/CPP/type_traits.h"
 #include "src/__support/FPUtil/FPBits.h"
+#include "src/__support/FPUtil/FloatProperties.h"
 #include "src/__support/FPUtil/dyadic_float.h"
 #include "src/__support/UInt.h"
 #include "src/__support/common.h"
 #include "src/__support/libc_assert.h"
+#include "src/__support/macros/attributes.h"
 
 // This file has 5 compile-time flags to allow the user to configure the float
-// to string behavior. These allow the user to select which 2 of the 3 useful
-// properties they want. The useful properties are:
-//  1) Speed of Evaluation
-//  2) Small Size of Binary
-//  3) Centered Output Value
-// These are explained below with the flags that are missing each one.
+// to string behavior. These were used to explore tradeoffs during the design
+// phase, and can still be used to gain specific properties. Unless you
+// specifically know what you're doing, you should leave all these flags off.
+
+// LIBC_COPT_FLOAT_TO_STR_NO_SPECIALIZE_LD
+//  This flag disables the separate long double conversion implementation. It is
+//  not based on the Ryu algorithm, instead generating the digits by
+//  multiplying/dividing the written-out number by 10^9 to get blocks. It's
+//  significantly faster than INT_CALC, only about 10x slower than MEGA_TABLE,
+//  and is small in binary size. Its downside is that it always calculates all
+//  of the digits above the decimal point, making it ineffecient for %e calls
+//  with large exponents. If this flag is not set, no other flags will change
+//  the long double behavior.
----------------
gchatelet wrote:

Same comments since it's a copy/paste.

https://github.com/llvm/llvm-project/pull/75414