Read Unicode into Struct

I have a file with a series of words beginning at offset 1440 and running for a variable number of bytes until the end of file. It does consist of exactly 18 words. Each of the words is in Unicode.

I have written an ifstream statement as follows where
nBytes =1440
l and m are long and I find the size of buffer required by going to end of file and subtracting 1440

Code:

---

ifstream filef((LPCTSTR) dlgFile.GetFileName(), ios::in|ios::binary|ios::ate);
filef.seekg (0, ios::end);
m = filef.tellg();
int catBytes = m-l;
filef.seekg (0, ios::beg);
filef.seekg (nBytes, ios::beg);
catbuffer = new char [catBytes];
filef.read (catbuffer, catBytes);
filef.close();

---

I have a struct defined as follows

Code:

---

struct categories {
                    char category;
                    }catarray[18];

---

Question is how do I get the catbuffer into the catarray. If it helps the format of the unicode is B.r.o.w.n...G.r.e.e.n...O.r.a.n.g.e...
etc

Thanks for any guidance....Andrew
(C++ MFC App using FormView)

Hey Mogul
Thats a lot of code but i would let the computer do the work.
I would read the entire file into a cstring then .replace the seperator ... ?
with a delimiter like ascii 10 then extract the substrings into the array with
::afxExtractSubString( array[x], fullstring, x, 10 ) in a for loop of 18.
Regards

First off, you haven't specified what l is , is it 1440 too ?

Anyways, what you need to do is, have a WCHAR array in your struct instead of simply a char. I assume you are goint to have 18 wide char strings.
Then you need to keep looping through the buffer one by one and copy them to the member data of the structure . When you hit the '\0' terminator, you know you need to advance to the next structure.

Some sample code below:

Code:

---

typedef struct
{
        WCHAR szData[50];
} category;
category g_CatArray[18];

---

Code:

---

int nIndexOfStruct = 0;
int nIndexOfData = 0;
WCHAR* pData = (WCHAR*)catbuffer;

for(int j = 0 ; j < sizeof(catbuffer)/sizeof(WCHAR); j++)
{
        g_CatArray[nIndexOfStruct].szData[nIndexOfData] = (WCHAR)(*pData);
        if((*pData) == '\0')
        {
                nIndexOfStruct++;
                nIndexOfData = 0;
        }
        else
        {
                nIndexOfData++;
        }
        pData++;
}

---

Kirants,

I am not sure I follow, I think I understand that what you are suggesting is that the category in the struct is itself made up of a char array of szData

We then read the buffer and keep extracting letters until we hit a null and then we write the szData array to category and repeat for all 18

However, since pdata will contain the whole string "Green Brown Orange" the comparison if((*pData) == '\0') will never be true until the very last byte and therefore will only write 1 entry rather than 18...

So I am assuming I need compare something that is only looking at the next byte in pData and compare that to Null...

Any ideas....Thanks

You are right. The struct cannot be just a char. You need 18 structs , each one having enough space the copy the string , right ?

Hence the WCHAR szData[50]; 50 assuming is enough for the strings.

Quote:

---

keep extracting letters until we hit a null and then we write the szData array to category and repeat for all 18

---

That is right, however , as we extract , we also copy to the current structure like below:

Code:

---

g_CatArray[nIndexOfStruct].szData[nIndexOfData] = (WCHAR)(*pData);

---

So, that copies all data including the NULL terminator.
After copying, if we have it a NULL terminator, we advance to the next structure.

yes but when I debug this I get

g_CatArray[0]= "Brown Green Orange"

Instead of
g_CatArray[0] = "Brown"
g_CatArray[0]= "Green"
g_CatArray[0]= "Orange"

Because the condition I listed is never met and so

nIndexOfStruct++;
nIndexOfData = 0;

Does not execute until the last Null is read and the whole string is read into one element

Am I being think here?

hehe
Andrew

I would have thought your buffer had NULL terminatorsin between like you showed. Isn't that so ?

Code:

---

B.r.o.w.n...G.r.e.e.n...O.r.a.n.g.e...

---

yes that is correct....After the first pass through the loop pdata is equal to rown green orange....After the second pass through it is equal to own green orange

So when it gets to the first space pData = " green orange" which is not qual to NULL so it keep on going

After one more pass through pdata = reen orange and the array = Brown g

I don't know.. I can't say unless you upload the file that contains the strings. It appears that the .. there is not NULL terminator, but something else.
Open the file with a binary editor ( or open it in visual studio as binary instead of text ) and see what those actual bytes are. If it is in fact a terminator, it should be 00 00.

Also, what does *pData say at that location if you watch the variable ?

Thanks...you nailed it....I looked in hex editor at the ascii to check for ... between each variable but on closer examination of the bytes they were not NULL eg 00 08 00 changing them to 00 fixed the problem....thanks for sticking with me...Regards, Andrew

You are welcome..:wave:

Ok, so I am trying to take the structured array and populate the variables into a comboBox. I am using the following code

Code:

---

int nIndex;
int a;
for (a=0; a < 18; a++)
{
nIndex = m_currtex.AddString((LPCTSTR)&g_CatArray[a]);
m_currtex.SetItemData(nIndex,(DWORD)&g_CatArray[a]);
}

---

But dor some reason it is only picking up the first letter of each word in the array. I feel it has something to do with strings and nulls in C but not sure..

thanks for any feedback...A

Ok I think I am homing in on the problem...

The g_CatArray is a Wide Char array. As a result I think it may be terminating at the second byte. In order to feed the dialog I am assuming I have to convert the wCHAR to CHAR or CSTRING.

This of course will have to be done letter by letter I assume using a for loop...

A

Either that. Or change your project settings to use UNICODE. Is that an option ?

Anyways, see this article:
How to Display UNICODE Characters in a CListCtrl or a CEdit

Its my project I guess I can do anything....hehe

EDIT:

Oops...guess not...Adding Unicode to the project properties C++ section caused a whole bunch of errors...so guess thats out...

A

Note that, if you compile for UNICODE, applications may not run right away on 9x platforms.They would need Microsoft layer for Unicode installed to be able to run it.

If you are targeting NT platforms alone, there is no problem.. In fact on 2k and higher it is good to use UNICODE since the native API on these platforms are UNICODE.