fdbg0014
[-] fixed ugly bug of version 0013 procedure when reading unaccessible memory

fdbg0013
[?] fixed odd behaviour of loading DLLs observed in Vista x64 (thank to r22 for producing such a DLL)
    (Vista isn't able to read the whole image of DLL by ReadProcessMemory in one step)

fdbg0012
[+] rounding floating point numbers (e.g. 1.00000000000000000*10^0012 instead of 9.99999999999999999*10^11 when e.g. first 35 digits of floating point number are 9)

fdbg0011
[-] fixed bug in displaying of floating point numbers - thanx to Garthower
[-] fixed bug when changing floating point number registers
[+] started to rewrite procedures without usage of floating point instructions to avoid FPU fnstnev antidebug trick

fdbg0010
[+] more accurate disassemble back routine
[+] added possibility for code 1st to start displaying some instructions back (not from RIP exactly),
    set it in Face -> Various face settings -> Code 1st disassemble back instructions.
    0 is there the default choice, but perhaps somebody wants to have real number there.
    Please note that you can turn on logging single steps and steps over by unchecking
    checkboxes just above this settings.

fdbg000F
[+] status window
[+] you can enter 64-bit register name (rax rbx rcx ... r15 rip) instead of hexa address in address column of code data stack

fdbg000E
[-] fixed bug when disassembling imul reg16,reg16/mem,val as reg32 and missing content on [mem]
[+] more user-friendly displaying of floating point registers

fdbg000D
[+] added support for displaying content at disassembled address like:
    xor qword [rbp+rbx*8+0],9A2FC517 ; [000000010000308C]=0000000078CC9A80=USER32.MessageBoxA
    mov rdx,[0000000100003150] ; []=0000000000000045
    GS mov rax,[00000030] ; [000007FFFFFDE030]=000007FFFFFDE000
    if there is any register in brackets, RIP have to be on the instruction for display something (because executing instruction changes RIP + executing instruction may change register in brackets)
[-] corrected mistake of editing code->hexa when instruction size exceeded 8 bytes
[-] fixed few disassembler mistakes
[+] more accurate 80-bit floatings (due to the setting of floating control word precision bits to 11b = 64 bit precision mantissa)
[+] added Explore->Memory->List of pages
[-] fixed bug in displaying of DLL names in the Log window
[?] final exe was to big, so I decided to compress it, shrinked to less than 40% of its original size
    [?] because I don't know whether it is good [+] or bad [-]
    it should be good to defeat antidebugs searching for e.g. kernel32.WaitForDebugEvent or ntdll.NtWaitForDebugEvent in imports of its parent process
    it could be bad because I could violate some PE32+ rules (any violation isn't known to me yet, but nobody can be sure...)
[+] added support for writing-at-once through borders on pages with different access flags, e.g. writing code (read execute) and data (read write) from one file (or even you can now write the whole executable image = header + code + data + resource + imports + ...  from one file at once !!!)

fdbg000C
[+] added multibyte NOP instruction, AMD64 Pacifica instructions (disallowed in long mode RING3, usefull only for detecting malware)
[-] fixed bugs in proc_UNLOAD_DLL_DEBUG_EVENT, find_first_empty_slot_put_SW_BP_temporary_flag, find_first_empty_slot_put_SW_BP_permanent_flag
[+] added attaching to a process
[+] added debug symbols support (thanx to dbghelp.dll for doing this dirty job...)
[+] doubleclicking on Imports copies the whole line of your interest (all iSubitems of the iItem) into LogEdit
    then you can there comfortable select and copy (Ctrl-C) address and put breakpoint on it (Alt-F2 -> Ctrl-V)
    the return address in the stack after triggering breakpoint is the address of the instruction just after call the API
    I suppose kernel32 GetSystemTime, GetLocalTime, SystemTimeToFileTime, GetSystemTimeAsFileTime, ...
    are the hottest APIs of the interest of some type of guys playing with some type of program not having its sources
[-] I don't know how to fix this: Help doesn't work fast after attaching to test_attach.exe sample
    You have to wait about 10 seconds !!! Help starts to work quickly after exit of second thread of the sample...
[+] Rambo menu from fdbg000A is back !!!

fdbg000B
[+] the source code of fdbg000A completely rewritten from the base
[+] totaly new face, based on MDI, every ListView or Edit is on its own MDI child window
[+] comfortable face settings
[+] SysListView32 items don't loose their colors
[+] new debug loop design
[+] Successfully managed temporary/permanent SW as well HW breakpoints. Please don't set both
    permanent on the same RIP, I'm not able to manage this choice yet.
[+] fixed mistake in disasm engine (some SIB32 indexing opcodes, e.g. disasm of db 43h,80h,7ch,65h,04h,0b8h = cmp byte [r13+r12*2+04],-48)
