mirror of
https://github.com/frida/tinycc
synced 2025-03-12 18:53:00 +03:00

when x86-64 support was added, "for long double, we use x87 FPU". And indeed, tests show that Intel's extended precision is used, not double precision.