|
I'm reading Feng Yuan's, "Windows Graphics Programming", and while it is an incredible investigative journey, I am really confused about one particular feature. In his technical description of the the internal data structures of the Windows 2000 graphics engine and GDI APIs, he lists numerous undocumented structures which the operating system maintains in kernel address space. This is fascinating stuff, but, even a nitwit like me notices pretty quickly that most of these structs are populated with void pointers to other undocumented structs. I can't imagine why Microsoft would use so many void pointers. The only possibility that seems reasonable is that the void pointer conceals the actual data structure from nitwits like me in user address space.
Anybody have any hard intel on this stuff?
|
|
|
|
|
no but i want to know so if u discover anything please post it back here
"No matter where you go, there your are..." - Buckaoo Banzi
-pete
|
|
|
|
|
void pointers are generic and can point to anything. It allows you to write generic functions and classes. so you could pass a string pointer or a integere pointer or a struct pointer and they'll all be accepted by the coid pointer variable.
Nish
Author of the romantic comedy
Summer Love and Some more Cricket [New Win]
Review by Shog9
Click here for review[NW]
|
|
|
|
|
"undocumented structure" means the structure can't be defined, the reason that it can't be defined is those void pointers, they are not cool at all, they are something can not be defined.
i.e. for dialog template structure ("undocumented structure"), u see these stories:
WORD id;
void* pv;
what does pv mean?
if id=0; pv is null;
if id=1; pv to a DWORD;
if id=2; pv to a char string;
(i can't remember exactly values, here is just idea)
these kind of "structures" can't be defined, so called "undocumented structure" - do body can document (define) it.
hope u understand that void* is not cool, it is "no words to say".
includeh10
|
|
|
|
|
You know, I kind of posted this inquiry half in jest, becuase it was my understanding that void pointers are NOT cool (and are, in fact, unsafe). I mean without typecasting, the potential exists for the wrong type of pointer residing in the structure.
And, apparently, I've confused you guys somewhat, because the author represents much of his technical information on "undocumented internals" as being taken from Microsoft's documentation for the Device Driver SDK. The Feng Yuan also invents soem ingenious methods for obtaining information from OS kernel components (which is theoretically, not accessible from a user application). He does this with Windows 2000 (my machine has Windows 98, and some of the function calls he employs are not available on Windows versions earlier than 2000 or XP).
Anyway, it's a long story. And, I'm a novice to Windows systematics. The main reason I posted is that it seemed like a marginal technique to employ for critical kernel DLLs and the systems calls that they provide.
If I find out anything interesting, I'll get back.
|
|
|
|
|
I'm currently using a CMap because i need to associate a key with a specific value.
But i also need to retreive the keys in a specific order.
Is it possible to sort the keys in a map?
Is there any other way to get the map functionality and be able to sort the keys?
|
|
|
|
|
Stefan Dahlin wrote:
I'm currently using a CMap because i need to associate a key with a specific value.
But i also need to retreive the keys in a specific order.
Is it possible to sort the keys in a map?
From memory CMap uses a hash table, in which case the answer is no.
Stefan Dahlin wrote:
Is there any other way to get the map functionality and be able to sort the keys?
Yes move to STL. Look at Map and Multimap. Also this CP article may be of interest: http://www.codeproject.com/vcpp/stl/bimap.asp[^]
Neville Franks, Author of ED for Windows. www.getsoft.com
Make money with our new Affilate program
|
|
|
|
|
Why is it that when the following code is executed, the first dialog comes up and waits 'til the user clicks OK, but as soon as the second dialog comes up, the whole app abruptly shuts down without waiting for any interaction from the user?
BOOL CNuFunkApp::InitInstance()
{
CNuFunkDlg dlg;
m_pMainWnd = &dlg;
int nResponse = dlg.DoModal();
CNuFunkDlg dlg2;
m_pMainWnd = &dlg2;
int nResponse2 = dlg2.DoModal();
return FALSE;
}
|
|
|
|
|
Because you are creating a modal dialog box. What you need is a modeless dialog box. See the C++ FAQ, I think it was mentioned in there.
Rickard Andersson@Suza Computing
C# and C++ programmer from SWEDEN!
UIN: 50302279
E-Mail: nikado@pc.nu
Speciality: I love C#, ASP.NET and C++!
|
|
|
|
|
No, I'm pretty sure I want "Modal" behavior. What I would expect in this simplified example, is that the first dialog would come up and wait in a modal state until the user dismisses it, after which, the second dialog would come up and wait in a modal state until the user dismisses it, at which time the app would then shut down.
|
|
|
|
|
I missunderstood your point, hm...
Well, I would show these two dialogs in OnInitDialog() if I were you... but that's perhaps a bad code design, I don't know. If you do it in OnInitDialog() it WILL work.
Rickard Andersson@Suza Computing
C# and C++ programmer from SWEDEN!
UIN: 50302279
E-Mail: nikado@pc.nu
Speciality: I love C#, ASP.NET and C++!
|
|
|
|
|
Your application actually exits upon return from:
int nResponse = dlg.DoModal();
This is because of the previous line:
m_pMainWnd = &dlg;
Searching in the MSDN documentation reveals:
The Microsoft Foundation Class Library will automatically terminate your thread when the window referred to by m_pMainWnd is closed. If this thread is the primary thread for an application, the application will also be terminated. If this data member is NULL, the main window for the application's CWinApp object will be used to determine when to terminate the thread. m_pMainWnd is a public variable of type CWnd*.
Typically, you set this member variable when you override InitInstance. In a worker thread, the value of this data member is inherited from its parent thread.
-Sean
----
Shag a Lizard
|
|
|
|
|
Thank you very much for solving my mystery.
jer
|
|
|
|
|
I had the same problem once. After a lot of looking around I found out that the WM_QUIT message from the first dialog box gets left in the message queue. When I emptied the message queue between the calls, (or at least the WM_QUIT message) things worked fine.
Write back and let us know what you find in your case.
|
|
|
|
|
I'm working on a project using a MFC dialog-based application, though the problem lies elsewhere. I seem to be having stray data coming into some arrays i use, an array which is populated once and then left alone for reference. Even worse, there seems to be problem with where i declare my arrays in the header file. Declaring it 2 lines before, makes a difference and puts the program into a never ending loop. This seems very strange. I spoke to friend of mine and he said it happened to him but he could'nt figure out why. Although my problem seems to be a lot more severe. Has anybody else had similiar stray array errors before? it's completely negating the task and my deadline is a week away.
I tried adding extra blocks to my array and this helped cutting down some stray data, i.e declaring an array[140] as array[150], and now if i put the array back to [140] it hangs the program as well, so its definately making a difference.
Please Help!
Ayush
|
|
|
|
|
Post an example of the declaration and how you assign the array and access it.
Kuphryn
|
|
|
|
|
Just in the straightforward/primitive way
int used_nodes[142][2];
int instructions[24][140];
etc..
They're all accessed using loop variables, which are local so they are not shared. The array are declared under private: in the header file. If i declare the 'instructions' array before 'used', i get an access violation. If i do it the other way round, some parts of 'instruction',namely 3 parts, seem to get filled with data of the used array. Moving any array about in the header file like so causes an access violation..
|
|
|
|
|
I believe the one or more of the loops try to access and/or modify an index that is out of bound. Recheck all loops.
Do you do some math using on the integer used to access the array, i.e. int i = something * something1 - something2?
Kuphryn
|
|
|
|
|
Weird. The case of order of declaration is weird. As an experiment, declare both arrays as global. Test your loops.
Kuphryn
|
|
|
|
|
I'm guessing that you are going out of bounds on the end of one of the arrays. When strange problems start to have "solutions" like moving variables around, and you are using arrays, ALWAYS make sure you aren't going out of bounds on the array.
From the clues you've given, I'm betting 90% that you are going out of bounds on the used_nodes array. When you go out of bounds on it, if you have it before the instructions array, you are simply stomping on the instructions array. If you swap the order of declaration, you are modifying whatever variable comes next in your class. Or, if your class doesn't have any more members, you are modifying somebody else's data.
Chris Richardson
C/C++ Include Finder[^]
|
|
|
|
|
Thanks guys,
I guess i'll have it to look thru it again, from what i recall, most of my arrays are called from a fixed for loop so i don't know how there could be data going out of bounds. Although, i can see data being changed in an array while debugging, but i dont know where it came from. Is there any possibility of finding out what causes the change of value of a variable from the debugger, i.e. which line of code?
Cheers.
Ayush
|
|
|
|
|
It's possible to set "Data" breakpoints with VC6. Try to look it up in the MSDN (I'm not on my Dev Machine or I would). It's a little tricky, but it will definately do what you want. What it basically allows you to do, is have a breakpoint hit when some code stomps off the end of the array. If you want some more info, I'll post more tomorrow when I'm back on my dev machine.
Chris Richardson
C/C++ Include Finder[^]
|
|
|
|
|
Does anyone know if there are any c++ examples of parsing date/time from rfc822/1123 email message headers on the net?
(I've done much hunting for this but can't seem to find an actual snippet or source and I don't want to re-invent the wheel)
|
|
|
|
|
all u need is a regular expression engine or even a simple tokenizer class will do the job do u have either of those?
"No matter where you go, there your are..." - Buckaoo Banzi
-pete
|
|
|
|
|
No I don't, but I know that's how it's done in the perl world etc.
I was hoping to avoid all that and find a function in straight c or c++; the problem is that there are variants of the way that date is presented according to the docs I've been reading and it's not uncommon for an algorithm to handle that to have some "finagling" built into it, so rather than re-inventing the wheel...
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
