API_CHANGES revision 1.1.1.2
1This file is the short summary of the API changes: 2 329.02.2016 - Non-backward compatible 4 Several types and instructions are renamed to improve 5 readability. In general byte, half, and int are renamed 6 to 8, 16, and 32. Floating point types are also renamed 7 from d and s to f64 and f32. 8 9 [s|u]b -> [s|u]8 (8 bit values) 10 [s|u]h -> [s|u]16 (16 bit values) 11 [s|u]i -> [s|u]32 (32 bit values) 12 d -> f64 (64 bit floating point value) 13 s -> f32 (32 bit floating point value) 14 1518.05.2015 - Non-backward compatible 16 SLJIT_[I|]L[U|S]DIV is renamed to SLJIT_[I|][U|S]DIVMOD 17 1829.09.2014 - Non-backward compatible 19 The sljit_create_compiler, sljit_allocate_stack, and 20 sljit_free_stack functions have an allocator_data 21 argument now. 22 2319.09.2014 - Non-backward compatible 24 Using I, D, S prefixes in conditional and floating 25 point operations. And an L prefix to long multiplication 26 and division (op0 opcodes). 27 2811.08.2014 - Non-backward compatible 29 A currently unused options parameter is added to sljit_emit_enter 30 and sljit_set_context. 31 3206.07.2014 - Non-backward compatible 33 SCRATCH registers are renamed to Rx and SAVED registers 34 are renamed to Sx. See the explanation of these registers 35 in sljitLir.h. 36 3731.05.2014 - Non-backward compatible 38 SLJIT_TEMPORARY_EREGx registers were not renamed to 39 SLJIT_SCRATCH_EREGx when the change was done on 08.11.2012 40 4105.03.2014 - Backward compatible 42 The sljit_set_target now supports those jumps, which 43 does not created with SLJIT_REWRITABLE_JUMP flag. 44 Reason: sljit_emit_ijump does not support conditional 45 jumps. 46 4703.03.2014 - Non-backward compatible 48 SLJIT_MOV_UI cannot be combined with SLJIT_INT_OP. 49 Reason: SLJIT_INT_OP flag is not recommended to use 50 directly, and SLJIT_IMOV has no sign bit. 51 5229.01.2014 - Backward compatible 53 Bits assigned to SLJIT_MEM and SLJIT_IMM flags are changed. 54 Reason: the most common cases are fits into one byte now, 55 and more registers can be supported in the future. 56 5708.11.2012 - Non-backward compatible 58 SLJIT_TEMPORARY_REGx registers are renamed to SLJIT_SCRATCH_REGx. 59 6007.11.2012 - Non-backward compatible 61 sljit_emit_cond_value is renamed to sljit_emit_op_flags. An 62 extra source argument is added which will be used in the future. 63 6405.11.2012 - Backward compatible 65 sljit_emit_cond_value now supports SLJIT_AND and SLJIT_INT_OP 66 flags, which makes this function complete. 67 6801.11.2012 - Non-backward compatible 69 SLJIT_F* opcodes are renamed to SLJIT_*D to show that 70 they are double precision operators. Furthermore 71 SLJIT_*S single precision opcodes are added. 72 7301.11.2012 - Non-backward compatible 74 Register arguments of operations with SLJIT_INT_OP flag 75 must be computed by another operation with SLJIT_INT_OP flag. 76 The same way as SLJIT_SINGLE_OP flag works with floating point 77 numbers. See the description of SLJIT_INT_OP. 78 7901.11.2012 - Backward compatible 80 All operations whose support the SLJIT_INT_OP flag, have an 81 alternate name now, which includes the SLJIT_INT_OP. These 82 names starting with I. 83 8431.10.2012 - Non-backward compatible 85 Renaming sljit_w to sljit_sw, sljit_i to sljit_si, sljit_h 86 to sljit_sh, and sljit_b to sljit_sb. Reason: their sign 87 bit is part of the type now. 88 8920.10.2012 - Non-backward compatible 90 Renaming SLJIT_C_FLOAT_NAN to SLJIT_C_FLOAT_UNORDERED. 91 Reason: all architectures call these unordered comparions. 92