On 2017-02-16 14:56 +0100, Aurelien Buhrig wrote: > The pb is that both -32768 (0x8000) and 32767 should fit into an int > (an int16_t actually), and the newlib <stdlib.h> defines INT16_MIN as > -32768 . According to the Standard, '32768' is a decimal literal, but '-32768' is not. It's an expression using unary '-' to the literal '32768'. Here '32768' has type int32_t. So the expression also has type int32_t despite the result can fit into an int16_t. -- Xi Ruoyao <ryxi@xxxxxxxxxxxxxxxxx> School of Aerospace Science and Technology, Xidian University
