Printable View
Hi: I have two pointers
char* pointer1;
char* pointer2;
in a loop, new data are read from file into buffer pointed by pointer2 and the old data should move from pointer2 to pointer1; pointer1 always pointer to the old data, and pointer2 always points to new data;
how can I do that?
Thank you
I don't know if I'm reading your post correctly, but it sounds like you want something akin to a fifo maybe?
In that case, you need to do a lot of pointer management. This really depends on how you're using it.
One method might be:
// got new data
pointer1 = pointer2; // Now pointer1 is pointing to the old data.
// increment pointer2
// put new data at pointer2
That's sort-of a general outline. If pointer2 becomes the "old data" just as the new stuff somes in all you need to do is assign pointer1 to pointer2 and then put the new data in the buffer.
Post a more in-depth example of what you need if this is totally off the mark. Thanks!
assuming the pointers have been properly allocated,
strcpy(pointer1, pointer2);
Thank you:
I used the following code to obtain what I need:
char* pointer1;
char* pointer2;
after this, I used memcpy(pointer1,pointer2,size);
here size is known;
then read new data to buffer pointed to pointer2.
thank you
If pointer1 and pointer2 point to (null-terminated) strings, then you can do the following, but whichever way you do it, be sure that the area ponited to by pointer1 is large enough.Code:strcpy(pointer1,pointer2);
Just for clarity
Not a bad idea
Bad ideaCode:char* pointer1;
char* pointer2=new char[128];
pointer1=pointer2;
...
...
...
delete [] pointer2;
A memory leak results because the pointer to the memory originally allocated for pointer 2 is lost. Also a runtime error will result since the memory pointer to by pointer2 is the same memory pointed to by pointer1 which will already have been deleted by delete [] pointer1.Code:char* pointer1=new char[128];
char* pointer2=new char[128];
pointer1=pointer2;
...
...
...
delete [] pointer1;
delete [] pointer2;
TDM
If you really want to use strcpy() function as GCDEF and Sam pointed out, I think first of all, it is advisable for you to make sure of the way you handle those two character pointers. During the time you read the data from file into buffer, those two pointers should never be NULLs or it will be one of many famous causes for which you will receive exceptions thrown from your system. Moreover, both of your char pointers, as Sam mentioned in his post, must always point to NULL_terminated strings, which helps you avoid undefined behaviors when you compile and run your program. Therefore, a check to see if both strings aren't NULL and a comparison between their lengths right before they are used in this error-prone strcpy() function are really of necessity.
Personally, I don't use string handling functions from cstring although my first programming languages are old Pascal and C, it is simply because of their inefficiency and error-prone_NESS. For small "exercises" in colleges, they sure never give many worries, but in large projects, the worries they bring us seem to be much more or "larger" than what we can imagine. That is also why I always try to stick myself with String class and STL.
Regards,
-Vu
Note that john100 needs to have two strings or sets of data; a new and an old, at least that is my understaning.Quote:
Originally posted by TDM
A memory leak results because the pointer to the memory originally allocated for pointer 2 is lost. Also a runtime error will result since the memory pointer to by pointer2 is the same memory pointed to by pointer1 which will already have been deleted by delete [] pointer1.
Uh.....yeah.....that is my understanding as well. One of the other posts in this thread suggested setting one pointer equal to the other pointer. Since Memory would have to be allocated for both pointers to hold the new and old data I didn't think that would be a good idea. Therefore, I made an example and gave a reason why I thought that would not be a good idea. I also gave an additional example where setting one pointer equal to another pointer could be a good idea....................
TDM
Sam, please do not get me wrong but I think TDM's suggestions are really great...Quote:
Originally posted by Sam Hobbs
Note that john100 needs to have two strings or sets of data; a new and an old, at least that is my understaning.
There are a lot of different solutions to only one problem though...
Regards,
-Vu
Yes, it depends on requirements. I was just not sure that the code would satisfy the stated requirements; if it does, then my comments can be ignored. However note that in the original question, there is no use of nor any mention of a need to allocate the data dynamically. Assuming it is possible to allocate two areas of storage large enough for all the data to be processed, then there is no need to reallocate. It is elegant to be able to copy just pointers but the reallocation makes the solution less efficient.Quote:
Originally posted by VuQ_Le
Sam, please do not get me wrong but I think TDM's suggestions are really great...
There are a lot of different solutions to only one problem though...
Regards,
-Vu
If it is advantageous to allocate the exact amount of memory for each occurance of data, then the following might be closer to what is required:Code:char* pointer1=NULL;
char* pointer2;
int n;
while (More) {
delete [] pointer1;
pointer1 = pointer2;
// get current n
pointer2 = new char[n];
// other processing
}
delete [] pointer2;
Looks like an access violation to me.Code:char* pointer1=NULL;
char* pointer2;
int n;
while (More) {
delete [] pointer1;
pointer1 = pointer2;
// get current n
pointer2 = new char[n];
// other processing
}
delete [] pointer2;
Okay, I'll bite. Where would the access violation occur?Quote:
Originally posted by TDM
Looks like an access violation to me.
the seond iteration will cause an access violation on 'delete [] pointer1'.Quote:
Originally posted by Sam Hobbs
Okay, I'll bite. Where would the access violation occur?
Bingo
An access violation or any other kind of undefined behaviour... since pointer2 is uninitialized in the beginning, and pointer1 will have a copy of that undefined value when delete [] is called during the second iteration.Quote:
Originally posted by AvDav
the seond iteration will cause an access violation on 'delete [] pointer1'.
I think I would have said something such as that instead of just saying there would be an access violation.Quote:
Originally posted by AvDav
the seond iteration will cause an access violation on 'delete [] pointer1'.
Okay, well then just initialize pointer2 to NULL (zero) also. There are other variations that are likely but we don't know the details of how the data will be processed; we don't know when the pointer1 and pointer2 data will be accessed for other puposes. Based on my experience, pointer2 would be initialized to the first data before the loop begins.
However if no one wants to admit that the solution is closer to something useful than the one I was commenting on, then I hope others see that at least.
It just seemed a little too ironic to me. You commented on my post which I included on this thread to point out a potential mistake that could be encountered when working with pointers. Then you subsequently made exactly the same mistake that I was trying to empahsize.Quote:
Originally quoted by SAM HOBBS
I think I would have said something such as that instead of just saying there would be an access violation.
TDM
The post was updated to remove any ambiguity.Quote:
I don't know who "You" is. If I am the only one that does not then I assume I don't need to know.
TDM
I am sorry to have offended you or whatever. Note that in my initial response to your sample, it was unclear to me whether it accomplished what was being asked. Was that comment of mine unreasonable? I did not say anything about pointers; I did not say there was a mistake. I said very little.Quote:
Originally posted by TDM
It just seemed a little too ironic to me. You commented on my post which I included on this thread to point out a potential mistake that could be encountered when working with pointers. Then you subsequently made exactly the same mistake that I was trying to empahsize.
TDM
Then someone else made comment that I still don't understand.
I admit that my code had a mistake. If you can admit that your code was very general then that is all I meant to say. Do you agree that my code shows maitenance of two copies of data better than yours?
The important point of my sample is the algorithm. People often post code that has not been tested. I certainly try to test the code I post. I did verify that my code compiles at least.
Note that VuQ_Le's comment indicated that I considered your sample to not be good. That seems to have been what caused the problem. I did not intend to say your sample was not good; I just thought it was not a clear example of what was being asked.
Note that I said "needs to have two strings or sets of data". I really think that is reasonable. I really think VuQ_Le could have said whatever needed to be said without making it sound as if I was being so critical of you.
Did I say it was not good? I only meant that the sample was very general. The original question asks how to keep two sets of data, and the samples don't really show that. Right? It does not show data being put into the memory pointed to by pointer1 and pointer2; it only shows a pointer getting set to allocated memory and a pointer being set to the value of another pointer. I did not say there is anything wrong except that it is quite minimal.Quote:
Originally posted by VuQ_Le
Sam, please do not get me wrong but I think TDM's suggestions are really great...
There are a lot of different solutions to only one problem though...
Regards,
-Vu
You assumed that I considered TDM's suggestions to not be useful. Do you really, really, really think that TDM's suggestion is a complete answer to the original question? I think it shows something that was important to be shown but it is not a complete answer. I was just saying that a complete answer to the original question would require a little bit more.