|
salem wiem,
first of all, you should really read Beej's Guide to Network Programming[^].
i read it once i was at school, and it was really useful to begin with sockets.
for what you're looking for, the functions is gethostbyname() (cf. §4.10)
|
|
|
|
|
Hi all,
Iam trying to understand array of pointers and functions.
I tried doing the following but get warning messages. I have mentioned the warning message above the respective lines.
Can anyone help me fix it.?
Thanks in advance.
char * Names;<br />
int next = 0;<br />
char * list = "Abc";<br />
<br />
Names[ next ] = malloc( sizeof( char )*( strlen( list ) + 1 ) );<br />
memcpy( Names[ next ],list,sizeof( char )*( strlen( list ) + 1 ) );<br />
<br />
list = "cdefg";<br />
Names[ next+1 ] = malloc( sizeof( char )*( strlen( list ) + 1 ) );<br />
memcpy( Names[ next+1 ],list,sizeof( char )*( strlen( list ) + 1 ) );<br />
<br />
printf("vsp list got is %s\n", Names[0] );<br />
printf("vsp list got is %s\n", Names[1] );<br />
|
|
|
|
|
The problem is that Names is not an array of pointers but a simple char array. If you want a pointer array, you need to write this: char** Names; instead.
Also, instead of memcpy, you can simply use strcpy for a char array (if it is a null terminated string but in your case it is).
So, in your code snippet, when you do Names[0] for example, you simply access a character, not a pointer. That's why you get all these warnings.
|
|
|
|
|
malloc() returns a void* pointer. you should cast it into char* in your case.
BTW, are you coding C or C++ ?
if C++, abandon malloc()/free() for new/delete...
for the printf() warning, %s is waiting for a char*, which is exactly what Names is. you should change its declaration into char** Names or char* Names[]
|
|
|
|
|
Hi all,
thanks for ur help.i do in C. I have corrected the code and i get no warnings now. but i would like to know if the following way of coding and passing the address of the first pointer to funcB is threadsafe in multithreaded environment.
is the follwing way of coding correct?Any help?
<br />
funcA{<br />
char * Names[] = {};<br />
int next = 0;<br />
char * list = "Abc";<br />
<br />
<br />
Names[ next ] = malloc( sizeof( char )*( strlen( list ) + 1 ) );<br />
memcpy( Names[ next ],list,sizeof( char )*( strlen( list ) + 1 ) );<br />
<br />
list = "cdefg";<br />
Names[ next+1 ] = malloc( sizeof( char )*( strlen( list ) + 1 ) );<br />
memcpy( Names[ next+1 ],list,sizeof( char )*( strlen( list ) + 1 ) );<br />
<br />
printf("vsp list got is %s\n", Names[0] );<br />
printf("vsp list got is %s\n", Names[1] );<br />
<br />
funcB(....,...,(const char**)&Names);<br />
}<br />
<br />
funcB(...,...,const char ** names)<br />
{<br />
....<br />
...<br />
free( (char*)names[0]);<br />
free( (char*)names[1]);<br />
<br />
}<br />
<br />
<br />
|
|
|
|
|
the way of passing parameters to a function doesn't affect the "thread safety" of the code...
if the resource is potentially alterable by several threads at the same time, then each thread must lock the resource before using it (and of course releasing it then)...
|
|
|
|
|
Say, the pointer of arrays char** Names is used by many threads.
in which case, if i make a local copy of the names[0] and names[1] values inside funcB()
say,
funcB(...,...,char **names)<br />
{<br />
char ** buf;<br />
buf = names;<br />
<br />
<br />
free ((char*)names[0]);<br />
free ((char*)names[1]);<br />
}<br />
|
|
|
|
|
Names[ next ] dereferences the pointer, so it is actually a char. So when you try to assign a pointer to it (malloc returns void*), or assign it to a pointer (memcpy takes void*, %s as format string takes a char *), you try to assign an adress to to a value-type. This is of course not wrong, but it is most likely not what you want, hence the warning. The solution would be to declare char ** Names. Then you wold first have to allocate memory on Names with the size of the char *. Then you could allocate memory for Names[x] and do what you did here. I'm to lazy to write out the code here, also, figuring it out yourself is much better practice 
Hope it helped a little.
|
|
|
|
|
I'm doing a program to solve the problem: find the k-shortest paths in a directed graph from source vertex s to destination vertex t.
Do you have some codes for this?
Please send me.
Thanks!
Nothing Is Impossible !
|
|
|
|
|
Hi all,
I'm trying to read(retrieving) and create xml files using MSXML.
I would like to enable validation before retrieving the xml with a perticular DTD.
The document that I use for retrieving the values does not have any DTD information. I need to specify the external file while loading the XML, and validate it. Please let me know if it is possible and if not possible, how to change the input XML file to contain DTD information.
Please guide me in this regard.
Thanks and regards,
raja pratap
|
|
|
|
|
I posted a similar question a while back without much luck, but it was the weekend 
Take a look at this code sample
for ( int i=0; i<10; i++ )
{
HRGN hRgn = ::CreateRectRgn(0, 0, 10, 10);
::DeleteObject(hRgn);
}
A simple loop that opens a GDI handle then imideatly closes it. If I run the code and watch it in the task managers "GDI Objects" the results are confusing. I enter the loop with 4 handles and leave with 7. This figure will gradualy increase as I create more and more handles, so by the time my initialization is complete I have 400+ GDI objects.
My guess is that some sort of pool is used with a possible garbage collector, ie the handle is only freed as and when it is required. Could anybody please confirm this, and/or run the above example and tell me if you notice the same behaviour.
Obviously I don't have 400+ GDI objects, is there any way I can get the true count and/or force the task manager to update with the true figure?
|
|
|
|
|
I ran your simple loop in a dialog based app except I set it to loop 1000 times. The app started with 20 GDI objects and went up to 26 when I ran the loop once. It then stayed at 26 no matter how many times I ran the loop after that.
In your app, are you sure that you have SelectObject()ed the HRGNs out of the HDCs before you tried to DeleteObject()ed them? Any GDI object that is still selected into a DC will not be deleted by a call to DeleteObject().
You may be right
I may be crazy
-- Billy Joel --
Within you lies the power for good, use it!!!
|
|
|
|
|
So far I have found this behaviour only relates to CreateRgn set of functions. As you probably know already my current project is very region intensive. Nowhere in my code do I directly SelectObject() a region into a DC, I always use the SetClip() set of functions and even those calls are made using global regions which will not be deleted.
It's safe to say that DeleteObject() is working and that there is no leakage, but can you imaging the feedback I would get when users take a look at the task manager?
|
|
|
|
|
Not seeing your code I can not tell for sure, but as you know any GDI object can cause this problem. Just because you use HRGNs extensively does not mean that is where the problem lies. I once had it with an icon in the title bar of a child window. One icon and it did not take long to have a couple hundred GDI objects just floating around. Look very carefully at all your GDI objects.
I do not know of a free tool that will point out which objects are leaking, so good luck in finding it.
You may be right
I may be crazy
-- Billy Joel --
Within you lies the power for good, use it!!!
|
|
|
|
|
PJ Arends wrote: I do not know of a free tool that will point out which objects are leaking, so good luck in finding it.
This is the thing, there are no leaks. That code sample was placed on the very first line of my WinMain(), so no other objects are created and no DC's are present. The object does get deleted, but the count does not decrease to reflect the change. I have confirmed this by checking with "Task Manager" and the GetGuiResources() api call.
To detect memory leaks I am using Deleaker[^] which reports no errors.
|
|
|
|
|
For what it's worth I tried this (in a console app):
for ( int i=0; i<100; i++ )
{
HRGN hRgn = ::CreateRectRgn(0, 0, 10, 10);
::DeleteObject(hRgn);
Sleep(1000);
}
and watched the task manager. GDI Objects climbed to 9 (went from 4 at startup to 11 13) and stayed
there until the end.
I can't find any documentation on Windows caching these internally either 
Good luck!
Mark
*EDIT* This test was on XP Pro Version 2002 SP2
|
|
|
|
|
Hello all,
I would like to solve a mystery:
If you have more that one keyboard layout they are displayed on the taskbar.
Every thread can change its language.
Now if you have a dialog which uses msrdp.ocx the language monitor is hidden.
How can you hide this monitor ??? or better how can you display it.
Can you get the current language for a prticular window that is not your process (this is what the language bar does)?
thanks
Zippo
|
|
|
|
|
Ok, could someone please explain this to me. As it seems a manifest file is stopping my application from loading. And I'd like to make sure my machine isn't going crazy or that I'm not playing the retard today. Anyway, my app in question uses the following manifest...
<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestVersion="1.0">
<assemblyIdentity processorArchitecture="x86" version="5.1.0.0" type="win32" name="Colorz.exe"></assemblyIdentity>
<description>Colorz</description>
<dependency>
<dependentAssembly><assemblyIdentity type="win32" name="Microsoft.Windows.Common-Controls" version="6.0.0.0" publicKeyToken="6595b64144ccf1df" language="*" processorArchitecture="x86"></assemblyIdentity></dependentAssembly>
</dependency>
<dependency>
<dependentAssembly><assemblyIdentity type="win32" name="Microsoft.VC80.CRT" version="8.0.50608.0" processorArchitecture="x86" publicKeyToken="1fc8b3b9a1e18e3b"></assemblyIdentity></dependentAssembly>
</dependency>
</assembly>
The bolded line is cause the problems. Which I added to include xp themes for a native app. Simply removing it allows the app to run, but I don't have themes. I should mention that it's worked fine in the past so I'm hoping it doesn't mean something is messed up on my machine.
The app in question is Colorz[^] if you'd like to try to reproduce this. Could someone please shed some light on this.
TIA
|
|
|
|
|
Let me guess. You took the old VC6 workspace file, opened it in VC8 and let the conversion wizard do it's thing. And now you are trying to retofit XP themes into the app. My experience is that that does not work. It causes more headaches than it is worth. If you really want to upgrade the app then use the VC8 new app creation wizard to create a new app and then copy the source files over, and open the RC files as text and do a copy/paste with the resources. The conversion wizard just leaves way to much crap behind.
|
|
|
|
|
PJ Arends wrote: Let me guess. You took the old VC6 workspace file, opened it in VC8 and let the conversion wizard do it's thing.
 Good guess.
PJ Arends wrote: The conversion wizard just leaves way to much crap behind.
I may very well do this, now and in the future. One thing that gets me is it did work before, using the same projects. It just stopped working and I'm hoping it wasn't due to my system getting messed up. It doesn't seem messed up so I don't know.
|
|
|
|
|
It probably never did work, just that with your old monitor you could not tell the difference.
|
|
|
|
|
PJ Arends wrote: It probably never did work, just that with your old monitor you could not tell the difference
|
|
|
|
|
I went ahead and created a new project, adding the files manually, etc. and it still doesn't work. Out of curiosity, does the themed Colorz run ok on your system?
Thanks for the help btw.
|
|
|
|
|
I downloaded the exe from your article and ran it. Everything does look ok. I took a screen shot so you can compare it to what you are seeing on your machine. http://www3.telus.net/pja/colorz.PNG[^]
You may be right
I may be crazy
-- Billy Joel --
Within you lies the power for good, use it!!!
|
|
|
|
|
Now I'm convinced something is wrong with my machine.  Thanks again.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
