4 ** New in version 0.4.2
8 - The disassembler is now in a separate package, and is re-exported for
9 compatibility by Harpy. This allows us to give the more liberal BSD licence
14 - The "mov", "jcc" (jump on condition code), and "cmp" instructions support
17 - Henning Thielemann provided some further floating point and SSE
20 - The loop instruction supports labels now.
24 - The disassembly of Harpy's internal code buffers includes all labels
25 now, even when multiple labels are defined for the same location.
27 ** New in version 0.4.1
31 - Many of Harpy's types are now instances of Eq.
35 - Added support for the prefetching instructions PREFETCH0, PREFETCH1,
36 PREFETCH2 and PREFETCHNTA.
40 - Harpy.X86Disassembler.disassembleBloc was too strict and caused
41 stack overflows for large inputs. This was fixed.
43 - Disassembler: The instruction prefix list was not cleared when
44 beginning to parse a new instruction. This caused incorrect
45 disassembly of SSE instructions.
47 - Disassembler: A bug has been fixed in the parsing routine for the
48 addressing mode "scaled index + 32 bit offset" without base
53 - New convenience top-level module "Harpy", which re-exports
54 Harpy.CodeGenMonad, Harpy.Call and Harpy.X86Assembler
56 - It is now possible to override Harpy's automatic code buffer
57 management. The new field 'customCodeBuffer' in the type
58 'CodeGenConfig' can be set to 'Just (buf, size)', where 'buf' is a
59 pointer to a memory region of 'size' bytes. Harpy will then use the
60 supplied code buffer and will not perform any automatic code buffer
61 allocation on overflow. Overflow checking is still performed and
62 will result in an exception in the CodeGen monad.
64 - When using the high-level assembler in X86Assembler, the code buffer
65 is automatically protected from overflow.
67 - Floating point operations added to X86Assembler (only for double
70 - Preliminary support for SSE instructions. Currently, only the
71 packed and scalar floating-point arithmetic operations are supported
72 (both in the low-level module Harpy.X86CodeGen and as methods in
75 - Code buffer default size has been increased from 128 to 4096 bytes.
77 - The CodeGenMonad fails when a label is defined twice.
79 - It is now possible to associate names with labels, using the new
80 operation newNamedLabel. The given names will show up in the
81 disassembly, which makes debugging of generated code much easier.
83 - The doc directory contains a second, slightly larger tutorial now.
85 - The examples/evaluator directory contains a small example
86 interpreter for arithmetic expressions, which translates expressions
87 entered at the keayboard to machine code on the fly. This is the
88 demo program we presented at the Haskell Workshop 2007.
92 - Everything is new! This is the first released version.