Index
> Windows > Help combining two codesGoto page Previous 1, 2, 3, 4, 5, 6 Next |
| Author |
|
|
typedef 17 Dec 2011, 05:22
Ctrl + F ----> IDM_NEW
|
|||
17 Dec 2011, 05:22 |
|
|
peet 17 Dec 2011, 13:55
so this is the definition of IDM_NEW in miniPAD.ASM
Code:
cmp eax,IDM_NEW
je .new
these lines i was searching for...! @typedef: What about CTRL-F? |
|||
|
17 Dec 2011, 13:55 |
|
|
peet 17 Dec 2011, 14:05
so i'm again at the point trying to intergrate the fading transparent effect of 'transparent window' to the miniPAD.
could someone teach me how to get those two codes combined the right way please? i do not see some depencies i guess?! |
|||
|
17 Dec 2011, 14:05 |
|
|
typedef 17 Dec 2011, 18:33
Did you understand what revolution said to you though ?
|
|||
|
17 Dec 2011, 18:33 |
|
|
peet 17 Dec 2011, 21:31
typedef wrote: Did you understand what revolution said to you though ? me DID understand - YOU did not understand what i asked, - as i wanted to know where and how IDM_NEW is defined (defined==what it does do) is is defined here Code: cmp eax,IDM_NEW
je .new and further here Code: .new: invoke SendMessage,[edithwnd],WM_SETTEXT,0,0 THIS is what i did ask if you would have understood me. But you made me follow your sight of things, instead of pickin me up where i was standing as a newbe in asm. Vou did hear/read a keyword and told me about it without recognizing my question behind that keyword. But back to learning asm, do i have the point or do still dismiss something about the construct IDM_NEW? |
|||
|
17 Dec 2011, 21:31 |
|
|
JohnFound 17 Dec 2011, 21:42
peet wrote: YOU did not understand what i asked, - as i wanted to know where and how IDM_NEW is defined (defined==what it does do) Hm, peet, maybe we have some language problem here. "defined" is not equal to "what it does do" in my understanding. Revolution, already explained you that the constant IDM_NEW is defined by the following line of code in your source: Code: IDM_NEW = 101 |
|||
|
17 Dec 2011, 21:42 |
|
|
peet 18 Dec 2011, 01:21
yep John, that is the point, but if a professional is talking to newbies he can't expect the newbie to understand his language and sight - otherwise he'd be no newbie. i know this problem from support myself, and we ourself try to keep that thing in mind. because if two people talk together often it is not that made shure they talked about the same thing. and of course if both side are no native speakers thing get eben more complicated sometimes. but i don't want to complain, i want to learn.
so anyway and really thanks for ALL answers, as they all helped me to step further. |
|||
|
18 Dec 2011, 01:21 |
|
|
peet 18 Dec 2011, 01:43
so back to the code, i will reread the combined code and try to find the point where the method is not already known by the code while calling. i'll try.
while with the point of integration i will need some more input. like how to place the scrollbar for transparncy somewhere usable place (like in menubar?). once more here is what i was trying to do: I found the provided miniPAD example and thought it would be very handy tool if the functionallity of the transparent window would be integrated into the miniPAD itself. therefore i tryed to identify the code inside the transparent window source code which would be necessary to add into miniPAD and where. I made this point fix to the similaries between the two window apps. So i compaired the code and took those passages from the transparent window which were not found in exact or similar kind in the miniPAD and adding it in the identical sections at the right point. to make this comprehensible i did mark the imported code with the ----- lines and asked where i did go wrong because of the compiling errors and my blindness to see where the functionallity was coded exactly. how should i find the reason for method (symbol?) not known at calling time when i can't see where it is defined (made?). maybe somebody could teach me, or tell me the right way. like maybe i have first to isolate the functions from transparent window before thinkin about importing it into another peace of code, or maybe you see that i only and simply did mix up some pieces of code because not having understood it really, while you see i mixed apples and pears? |
|||
|
18 Dec 2011, 01:43 |
|
|
AsmGuru62 18 Dec 2011, 04:58
@peet:
See my changes to MiniPAD marked by [AG62]. Use F5/F6 to change more or less transparency and use Ctrl+A to select all text. btw, why the FASM code samples so badly formatted and almost no comments used and no empty lines almost anywhere? Code: ; Simple text editor - fasm example program format PE GUI 4.0 entry start include 'win32a.inc' IDM_NEW = 101 IDM_EXIT = 102 ; ; [AG62]: See, I do not care about the numbers, just ; use a proper NAME for the menu item and make sure ; same number is not used for other menu items. ; IDM_MORE_PERCENTAGE = 52893 IDM_LESS_PERCENTAGE = 8371 IDM_SELECT_ALL = 3333 IDM_ABOUT = 901 FACTOR_STEP = 12 section '.text' code readable executable start: invoke GetModuleHandle,0 mov [wc.hInstance],eax invoke LoadIcon,eax,17 mov [wc.hIcon],eax invoke LoadCursor,0,IDC_ARROW mov [wc.hCursor],eax invoke RegisterClass,wc test eax,eax jz error invoke LoadMenu,[wc.hInstance],37 call AdjustMenu ; <-- [AG62] add some new menu items. invoke CreateWindowEx,WS_EX_LAYERED,\ ; <-- Added WS_EX_LAYERED style [AG62] _class,_title,\ WS_VISIBLE or WS_OVERLAPPEDWINDOW,\ 144,128,256,256,NULL,eax,[wc.hInstance],NULL test eax,eax jz error invoke CreateAcceleratorTable,tblAccel1,3 ; <-- [AG62] mov [hAccelerators], eax msg_loop: mov edi, msg invoke GetMessage,edi,NULL,0,0 cmp eax,1 jb end_loop jne msg_loop ; ; [AG62] ; First check if accelerators were in action ; invoke TranslateAccelerator,[hMainWnd],[hAccelerators],edi test eax, eax jnz msg_loop invoke TranslateMessage,edi invoke DispatchMessage,edi jmp msg_loop error: invoke MessageBox,NULL,_error,NULL,MB_ICONERROR+MB_OK end_loop: invoke DestroyAcceleratorTable,[hAccelerators] ; <-- [AG62] invoke ExitProcess,[msg.wParam] proc WindowProc hwnd,wmsg,wparam,lparam push ebx esi edi cmp [wmsg],WM_CREATE je .wmcreate ; ; To reduce flicker [AG62] ; cmp [wmsg],WM_ERASEBKGND je .wmerasebkgnd cmp [wmsg],WM_SIZE je .wmsize cmp [wmsg],WM_SETFOCUS je .wmsetfocus cmp [wmsg],WM_COMMAND je .wmcommand cmp [wmsg],WM_DESTROY je .wmdestroy .defwndproc: invoke DefWindowProc,[hwnd],[wmsg],[wparam],[lparam] jmp .finish .wmerasebkgnd: ; [AG62] - return TRUE to reduce flicker on resize push 1 pop eax jmp .finish .wmcreate: ; ; Set it to be 50% transparent on startup [AG62] ; mov eax, [hwnd] mov [hMainWnd], eax invoke SetLayeredWindowAttributes,eax,0,[transfactor],LWA_ALPHA invoke GetClientRect,[hwnd],client invoke CreateWindowEx,WS_EX_CLIENTEDGE,_edit,0,WS_VISIBLE+WS_CHILD+WS_HSCROLL+WS_VSCROLL+ES_AUTOHSCROLL+ES_AUTOVSCROLL+ES_MULTILINE,[client.left],[client.top],[client.right],[client.bottom],[hwnd],0,[wc.hInstance],NULL or eax,eax jz .failed mov [edithwnd],eax invoke CreateFont,16,0,0,0,0,FALSE,FALSE,FALSE,ANSI_CHARSET,OUT_RASTER_PRECIS,CLIP_DEFAULT_PRECIS,DEFAULT_QUALITY,FIXED_PITCH+FF_DONTCARE,NULL or eax,eax jz .failed mov [editfont],eax invoke SendMessage,[edithwnd],WM_SETFONT,eax,FALSE xor eax,eax jmp .finish .failed: or eax,-1 jmp .finish .wmsize: invoke GetClientRect,[hwnd],client invoke MoveWindow,[edithwnd],[client.left],[client.top],[client.right],[client.bottom],TRUE xor eax,eax jmp .finish .wmsetfocus: invoke SetFocus,[edithwnd] xor eax,eax jmp .finish .wmcommand: mov eax,[wparam] and eax,0FFFFh cmp eax,IDM_NEW je .new cmp eax,IDM_ABOUT je .about cmp eax,IDM_EXIT je .wmdestroy ; ; [AG62] ; mov ecx, FACTOR_STEP cmp eax,IDM_LESS_PERCENTAGE je .less_perc cmp eax,IDM_MORE_PERCENTAGE je .more_perc cmp eax,IDM_SELECT_ALL je .select_all_text jmp .defwndproc .select_all_text: invoke SendMessage,[edithwnd],EM_SETSEL,0,-1 jmp .finish .less_perc: neg ecx .more_perc: mov ebx, [hwnd] call ApplyTransparency jmp .finish .new: invoke SendMessage,[edithwnd],WM_SETTEXT,0,0 jmp .finish .about: invoke MessageBox,[hwnd],_about_text,_about_title,MB_OK jmp .finish .wmdestroy: invoke DeleteObject,[editfont] invoke PostQuitMessage,0 xor eax,eax .finish: pop edi esi ebx ret endp ApplyTransparency: ; [AG62] ; ; ECX = factor to add to 'transfactor' ; EBX = HWND ; mov eax, [transfactor] add eax, ecx ; ; Validate against range: [4..252] ; cmp eax, 4 jl .done cmp eax, 252 ja .done ; ; We're good - set transparency! ; mov [transfactor], eax invoke SetLayeredWindowAttributes,ebx,0,eax,LWA_ALPHA .done: ret AdjustMenu: ; [AG62] ; ; EAX = HMENU to adjust ; pusha mov ebx, eax ; ; Create empty popup menu ; invoke CreatePopupMenu mov edi, eax ; ; Add two new items to control the transparency ; invoke AppendMenu,edi,0,IDM_LESS_PERCENTAGE,_smenu1 invoke AppendMenu,edi,0,IDM_MORE_PERCENTAGE,_smenu2 ; ; Insert new popup menu into the one which got loaded ; invoke InsertMenu,ebx,1,MF_BYPOSITION or MF_POPUP,edi,_spopmenu popa ret section '.data' data readable writeable _title TCHAR 'MiniPad',0 _about_title TCHAR 'About MiniPad',0 _about_text TCHAR 'This is Win32 example program created with flat assembler.',0 _error TCHAR 'Startup failed.',0 _class TCHAR 'MINIPAD32',0 _edit TCHAR 'EDIT',0 wc WNDCLASS 0,WindowProc,0,0,NULL,NULL,NULL,COLOR_BTNFACE+1,NULL,_class edithwnd dd 0 editfont dd 0 transfactor dd 128 ; [AG62] - 50% transparent on startup hAccelerators dd 0 hMainWnd dd 0 _smenu1 db '&Less Percentage',9,'F5',0 _smenu2 db '&More Percentage',9,'F6',0 _spopmenu db '&Transparency',0 align 4 tblAccel1 db FVIRTKEY,0 dw VK_F5,IDM_LESS_PERCENTAGE tblAccel2 db FVIRTKEY,0 dw VK_F6,IDM_MORE_PERCENTAGE tblAccel3 db FCONTROL or FVIRTKEY,0 dw 41h,IDM_SELECT_ALL msg MSG client RECT section '.idata' import data readable writeable library kernel,'KERNEL32.DLL',\ user,'USER32.DLL',\ gdi,'GDI32.DLL' import kernel,\ GetModuleHandle,'GetModuleHandleA',\ ExitProcess,'ExitProcess' import user,\ RegisterClass,'RegisterClassA',\ CreateWindowEx,'CreateWindowExA',\ DefWindowProc,'DefWindowProcA',\ SetWindowLong,'SetWindowLongA',\ RedrawWindow,'RedrawWindow',\ GetMessage,'GetMessageA',\ TranslateMessage,'TranslateMessage',\ DispatchMessage,'DispatchMessageA',\ SendMessage,'SendMessageA',\ LoadCursor,'LoadCursorA',\ LoadIcon,'LoadIconA',\ LoadMenu,'LoadMenuA',\ GetClientRect,'GetClientRect',\ MoveWindow,'MoveWindow',\ SetFocus,'SetFocus',\ MessageBox,'MessageBoxA',\ PostQuitMessage,'PostQuitMessage',\ CreatePopupMenu,'CreatePopupMenu',\ AppendMenu,'AppendMenuA',\ InsertMenu,'InsertMenuA',\ CreateAcceleratorTable,'CreateAcceleratorTableA',\ DestroyAcceleratorTable,'DestroyAcceleratorTable',\ TranslateAccelerator,'TranslateAcceleratorA',\ SetLayeredWindowAttributes,'SetLayeredWindowAttributes' import gdi,\ CreateFont,'CreateFontA',\ DeleteObject,'DeleteObject' section '.rsrc' resource data readable ; resource directory directory RT_MENU,menus,\ RT_ICON,icons,\ RT_GROUP_ICON,group_icons,\ RT_VERSION,versions ; resource subdirectories resource menus,\ 37,LANG_ENGLISH+SUBLANG_DEFAULT,main_menu resource icons,\ 1,LANG_NEUTRAL,icon_data resource group_icons,\ 17,LANG_NEUTRAL,main_icon resource versions,\ 1,LANG_NEUTRAL,version menu main_menu menuitem '&File',0,MFR_POPUP menuitem '&New',IDM_NEW menuseparator menuitem 'E&xit',IDM_EXIT,MFR_END menuitem '&Help',0,MFR_POPUP + MFR_END menuitem '&About...',IDM_ABOUT,MFR_END icon main_icon,icon_data,'minipad.ico' versioninfo version,VOS__WINDOWS32,VFT_APP,VFT2_UNKNOWN,LANG_ENGLISH+SUBLANG_DEFAULT,0,\ 'FileDescription','MiniPad - example program',\ 'LegalCopyright','No rights reserved.',\ 'FileVersion','1.0',\ 'ProductVersion','1.0',\ 'OriginalFilename','MINIPAD.EXE' |
|||
|
18 Dec 2011, 04:58 |
|
|
peet 18 Dec 2011, 10:47
@AG62:
thanx a lot for your efforts, your comments make it possible to follow the code a lot. maybe the author/someone of miniPAD would like to comment his lines too, that be good for newbies to learn follow your way of thinkin. i recognized the transpareny is limited to almost visibility, or other way round it is not possible this way to switch transparency off completly? if you want to work inside pad for longer without needing to look behind it would be nice for the eyes to get a non fading window. maybe must be done with an extra switch to turn accelaration off? do you know notepad++? i guess shure. it got a nice feature, if you select some string, it marks all the exact matches throuout the whole content (while input is only used on selecion itself). this is nice for working on scripts and would be perfect if you too had the choice of an option (set in menu) to set to use input only on selection or on all marked matches. and if i decompile the resulting exe i see 80% of the lines with empty DB lines, (good for injecting code i know) is this a result of comments in source code taken out by compiler or what is the reason, and would it be possible to eleminate those lines and get the exe even more smaller? |
|||
|
18 Dec 2011, 10:47 |
|
|
revolution 18 Dec 2011, 11:08
peet wrote: and if i decompile the resulting exe i see 80% of the lines with empty DB lines, (good for injecting code i know) is this a result of comments in source code taken out by compiler or what is the reason, and would it be possible to eleminate those lines and get the exe even more smaller? |
|||
|
18 Dec 2011, 11:08 |
|
|
peet 18 Dec 2011, 11:42
shure if you have source code you don't need to inject, you are able to "inject" by coding the normal way - anything other would be stupid in that case.
without source the DB lines gimme possibility to work with jumps and to hook the programm running order the way i want, and this way injecting code without damaging footprint or funtionallity. why i ask for elimiminating those lines is to make it almost impossible to inject code in the resulting exe, this way i'd try to harden the code against hacking. are there other possibilities do harden code, maybe an internal hash compare or by maybe weird source code order? is there a way to cut or limit headers maybe? maybe by special compiling settings? or how to determine how much senceless codeline must be inserted in witch section to fill headers with fakecode, then create hash and depend running code on correct hash? |
|||
|
18 Dec 2011, 11:42 |
|
|
AsmGuru62 18 Dec 2011, 12:03
@peet:
1. To switch off transparency - you need to set that variable (transfactor) to 255 and window will not be transparent, just like a 'normal' window. 2. The standard edit control cannot select multiple places - you need to use some other control or write one by yourself - it will be hard for a person who is new into FASM or new into Windows programming. I could help you with that, but my edit control will be able to load/edit/copy/paste/save ANSI text files, view them and highlight multiple occurences of some text with color - that is all, i.e. - it be much simpler than usual 'edit' control. 3. To make EXE smaller - you can use the section combining or section alignment modification. There was a lot of threads lately on the subject. The EXE file is at 5Kb - is it really worth the effort to reduce it further? Say, you got it down to 2Kb - it will be loaded by Windows faster of course, but by the amount of time a human is not able to perceive, so why do that? I can see an EXE getting down from 5Mb to 2Mb as useful, but not for such a small EXE as this one. 4. Did you understand how the Accelerator Keys are used in the code? There is a set of structures starting at the address 'tblAccel1'. If you need to add an accelerator key you need to add one more structure into that array and increase count when invoking 'CreateAcceleratorTable'. Then you need to add the case into WM_COMMAND processing and write your code there - just look how 'Select All' is coded. |
|||
|
18 Dec 2011, 12:03 |
|
|
revolution 18 Dec 2011, 12:23
peet wrote: why i ask for elimiminating those lines is to make it almost impossible to inject code in the resulting exe, this way i'd try to harden the code against hacking. are there other possibilities do harden code, maybe an internal hash compare or by maybe weird source code order? is there a way to cut or limit headers maybe? maybe by special compiling settings? If you want to "harden" an exe then the way to do it is to sign it. It is not free though, you have to pay MS to get a signing key. However, this does not encrypt an exe. So any hacker can simply strip the signature and work from there. The only real purpose for signing is for driver/kernel code to stop hackers getting altered code into ring0. |
|||
|
18 Dec 2011, 12:23 |
|
|
peet 18 Dec 2011, 12:41
@AG62:
am i right that extending the existing menu would make corrections throughout several source files neccessary? or why do we inject our new menu items? -> 1. i tried to set the validation range up to 255 but did not reach full visibility, still did see shadow of objects underneath? though we do always add 12 we seem not to reach the limit? if i use step 1 and set range to 255 it works, so we would need to modify adding function to allow more than 255 and cutting it off to limt 255 aftwerwards. or use steps which match smallest overall factor of used values. or use 255 - X *12 as start setting for transfactor am getting into lol Last edited by peet on 18 Dec 2011, 12:59; edited 5 times in total |
|||
|
18 Dec 2011, 12:41 |
|
|
peet 18 Dec 2011, 12:45
revolution wrote: You can't "harden" exe's against any form of modification. Any hacker can simply disable any check code and extend the exe for whatever purposes they desire. shure, but if no empty lines are found every modification will leave resulting errors in recompiled exe, now the hacker must know more than just how to inject, he must find a place where his damging is not seen easily (like an almost never just function of the app) and understand where the code does what and how. this was my intention to eliminate unneeded space (@AS62) in my eyes this is hardened app. |
|||
|
18 Dec 2011, 12:45 |
|
|
revolution 18 Dec 2011, 12:59
peet wrote: shure, but if no empty lines are found every modification will leave resulting errors in recompiled exe, now the hacker must know more than just how to inject, he must find a place where his damging is not seen easily (like an almost never just function of the app) and understand where the code does what and how. |
|||
|
18 Dec 2011, 12:59 |
|
|
peet 18 Dec 2011, 13:03
hm, if there is no unneccessary line of code every change would leave a missing or damged functionality in the recompiled exe. if one line was added the resulting exe should be bigger size of file than original - no?
i do not want to stop modification, i want to recognize it by eyeing the filesize. and to be shure all functions are well a autoit script will do easy checking for us, that run in a sandbox helps me getting shure of app before executing on sensitive systems. on those systems even a weird pointer can kill me if set to false memory adress. |
|||
|
18 Dec 2011, 13:03 |
|
|
revolution 18 Dec 2011, 13:13
peet wrote: hm, if there is no unneccessary line of code every change would leave a missing or damged functionality in the recompiled exe. if one line was added the resulting exe should be bigger size of file than original - no? |
|||
|
18 Dec 2011, 13:13 |
|
| Goto page Previous 1, 2, 3, 4, 5, 6 Next < Last Thread | Next Thread > |
Forum Rules:
|