C fread() magically reading dynamically allocated struct members, how? - c

This is a test program that I have written for a larger project that I am working on. It has to do with writing struct data to disk with fwrite() and then reading that data back with fread(). One member of the struct is dynamically allocated.
First, here is my code
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define STRING_LEN 128
struct Person {
int age;
char *name;
};
int main(int argc, const char *argv[])
{
struct Person *person = calloc(1, sizeof(struct Person));
person->age = 22;
person->name = calloc(STRING_LEN, sizeof(char));
char *name = "Name that is really, really, really, really, really, really, long.";
strncpy(person->name, name, STRING_LEN);
FILE *out_file = fopen("rw.out", "w");
fwrite(person, sizeof(struct Person), 1, out_file);
fclose(out_file);
FILE *in_file = fopen("rw.out", "r");
struct Person *person_read = calloc(1, sizeof(struct Person));
fread(person_read, sizeof(struct Person), 1, in_file);
fclose(in_file);
printf("%d %s\n", person_read->age, person_read->name);
free(person->name);
free(person);
free(person_read);
return 0;
}
And the outpout
22 Name that is really, really, really, really, really, really, long.
My question is, why is this working? Shouldn't fwrite() only write the address that 'name' contains (i.e., the address of the beginning of the string)? That is, I am passing in sizeof(struct Person) to fwrite() and yet it is writing the string the 'name' is pointing to.
Even more confusing to me is the behavior of fread(). Again, if I am passing sizeof(struct Person), how is the actual value of 'name' being read? How is the memory for it being allocated?
My previous understanding of how to use fwrite() + fread() was that I would have to "manually" write the data that 'name' was pointing to, "manually" read that data, and then copy that string after allocating memory for both the structure and the 'name' member. In other words, I would have to traverse any pointers myself, write the data, and then read that data back in the same order.
EDIT: Dan and the others are correct. I have looked at the output file with xxd:
0000000: 1600 0000 0000 0000 30a0 d900 0000 0000 ........0.......
If I print out the address that 'name' contains before writing and after reading it is the same (0xd9a030), which matches the output from xxd.

You are writing the data in the struct, which is an int followed by a pointer to a string. It's just data like anything else, and you know how long it is because the struct is fixed length - an int plus a pointer. You read the same pointer to the same name string as the original. The name itself is neither written nor read.

Both person->name and person_read->name wind up pointing to the same memory location. Since you didn't deallocate person->name before reading the file back in, the pointer value in person_read->name is still valid.
If you had deallocated person->name or read the file from a different program, the pointer value would no longer be valid, and attempting to reference it would invoke undefined behavior - you would either have printed out gibberish or gotten a segfault.

The *name pointer remains valid throughout the fwrite and fread calls, which is seemingly a fluke to you. If you free(person->name) before printf you would get the result or error your were expecting.

Related

How to fix "realloc(): invalid pointer"

I am trying to write a function to convert a text file into a CSV file.
The input file has 3 lines with space-delimited entries. I have to find a way to read a line into a string and transform the three lines from the input file to three columns in a CSV file.
The files look like this :
Jake Ali Maria
24 23 43
Montreal Johannesburg Sydney
And I have to transform it into something like this:
Jake, 24, Montreal
...etc
I figured I could create a char **line variable that would hold three references to three separate char arrays, one for each of the three lines of the input file. I.e., my goal is to have *(line+i) store the i+1'th line of the file.
I wanted to avoid hardcoding char array sizes, such as
char line1 [999];
fgets(line1, 999, file);
so I wrote a while loop to fgets pieces of a line into a small buffer array of predetermined size, and then strcat and realloc memory as necessary to store the line as a string, with *(line+i) as as pointer to the string, where i is 0 for the first line, 1 for the second, etc.
Here is the problematic code:
#include <stdio.h>
#include<stdlib.h>
#include<string.h>
#define CHUNK 10
char** getLines (const char * filename){
FILE *file = fopen(filename, "rt");
char **lines = (char ** ) calloc(3, sizeof(char*));
char buffer[CHUNK];
for(int i = 0; i < 3; i++){
int lineLength = 0;
int bufferLength = 0;
*(lines+i) = NULL;
do{
fgets(buffer, CHUNK, file);
buffLength = strlen(buffer);
lineLength += buffLength;
*(lines+i) = (char*) realloc(*(lines+i), (lineLength +1)*sizeof(char));
strcat(*(lines+i), buffer);
}while(bufferLength ==CHUNK-1);
}
puts(*(lines+0));
puts(*(lines+1));
puts(*(lines+2));
fclose(file);
}
void load_and_convert(const char* filename){
char ** lines = getLines(filename);
}
int main(){
const char* filename = "demo.txt";
load_and_convert(filename);
}
This works as expected only for i=0. However, going through this with GDB, I see that I get a realloc(): invalid pointer error. The buffer loads fine, and it only crashes when I call 'realloc' in the for loop for i=1, when I get to the second line.
I managed to store the strings like I wanted in a small example I did to try to see what was going on, but the inputs were all on the same line. Maybe this has to do with fgets reading from a new line?
I would really appreciate some help with this, I've been stuck all day.
Thanks a lot!
***edit
I tried as suggested to use calloc instead of malloc to initialize the variable **lines, but I still have the same issue.I have added the modifications to the original code I uploaded.
***edit
After deleting the file and recompiling, the above now seems to work. Thank you to everyone for helping me out!
You allocate line (which is a misnomer since it's not a single line), which is a pointer to three char*s. You never initialize the contents of line (that is, you never make any of those three char*s point anywhere). Consequently, when you do realloc(*(line + i), ...), the first argument is uninitialized garbage.
To use realloc to do an initial memory allocation, its first argument must be a null pointer. You should explicitly initialize each element of line to NULL first.
Additionally, *(line+i) = (char *)realloc(*(line+i), ...) is still bad because if realloc fails to allocate memory, it will return a null pointer, clobber *(line + i), and leak the old pointer. You instead should split it into separate steps:
char* p = realloc(line[i], ...);
if (p == null) {
// Handle failure somehow.
exit(1);
}
line[i] = p;
A few more notes:
In C, you should avoid casting the result of malloc/realloc/calloc. It's not necessary since C allows implicit conversion from void* to other pointer types, and the explicit could mask an error where you accidentally omit #include <stdlib.h>.
sizeof(char) is, by definition, 1 byte.
When you're allocating memory, it's safer to get into a habit of using T* p = malloc(n * sizeof *p); instead of T* p = malloc(n * sizeof (T));. That way if the type of p ever changes, you won't silently be allocating the wrong amount of memory if you neglect to update the malloc (or realloc or calloc) call.
Here, you have to zero your array of pointers (for example by using calloc()),
char **line = (char**)malloc(sizeof(char*)*3); //allocate space for three char* pointers
otherwise the reallocs
*(line+i) = (char *)realloc(*(line+i), (inputLength+1)*sizeof(char)); //+1 for the empty character
use an uninitialized pointer, leading to undefined behaviour.
That it works with i=0 is pure coindicence and is a typical pitfall when encountering UB.
Furthermore, when using strcat(), you have to make sure that the first parameter is already a zero-terminated string! This is not the case here, since at the first iteration, realloc(NULL, ...); leaves you with an uninitialized buffer. This can lead to strcpy() writing past the end of your allocated buffer and lead to heap corruption. A possible fix is to use strcpy() instead of strcat() (this should even be more efficient here):
do{
fgets(buffer, CHUNK, file);
buffLength = strlen(buffer);
lines[i] = realloc(lines[i], (lineLength + buffLength + 1));
strcpy(lines[i]+lineLength, buffer);
lineLength += buffLength;
}while(bufferLength ==CHUNK-1);
The check bufferLength == CHUNK-1 will not do what you want if the line (including the newline) is exactly CHUNK-1 bytes long. A better check might be while (buffer[buffLength-1] != '\n').
Btw. line[i] is by far better readable than *(line+i) (which is semantically identical).

too many open files c

I have been trying to create a simple program. However, I encountered an error:
gmon.out:too many open files
I am not clear on why it says I have "too many open files". It does not appear I am using files.
#include<stdio.h>
#include<ctype.h>
#include<math.h>
#include<stdlib.h>
#include<string.h>
struct position
{
int line;
int place;
struct position *next;
};
struct file
{
struct position *info;
struct file *next;
char *name;
};
struct word
{
char *name;
struct word *right;
struct word *left;
struct file *result;
};
int main()
{
int i;
struct word *d,*c;
char *s="brutus";
printf("%s",s);
c=(struct word*)malloc(sizeof(struct word));
strcpy(c->name,s);
c->left=NULL;
c->right=NULL;
for(i=1;i<=10;i++)
{
d=(struct word*)malloc(sizeof(struct word));
if(d==NULL)
exit(0);
scanf("%s",s);
printf("4");
s=d->name;
printf("%s",d->name);
d->left=NULL;
d->right=NULL;
}
system("pause");
exit(0);
}
What should I do about it?Thank you in advnace for your time!
First off:
gmon.out:too many open files
Means that you're compiling with the -p flag (profiling). gmon.out is the default file-name used by gprof. Just ditch-the-switch, and you won't get that problem anymore.
Of course, not profiling code isn't great, but you'd do well to address a coupe of issues first, before setting about actually profiling your code.
Some of these, quite numerous, issues are:
char *s="brutus";
printf("%s",s);
c=(struct word*)malloc(sizeof(struct word));
strcpy(c->name,s);
List of issues:
char *s should be const char *s, because it points to read-only memory.
Next, Do not cast the return of malloc
Check the return value of functions like malloc, they tell you something
struct wordis a struct of which all members are pointers. After allocating the struct, those pointers are invalid: you need to allocate memory for those members, too
strcpy expects the destination (c->name) to be a valid pointer, as I explained above: this is not the case here
What, then, should this code look like:
const char *s = "brutus";
c = malloc(sizeof *c);
if (c == NULL)
{
fprintf(stderr, "Could not allocate memory for struct word\n");
exit( EXIT_FAILURE );
}
//allocate enough memory to store the string
c->name = malloc(
(strlen(s)+1) * sizeof *c->name
);
//OR same, but shorter, works because the type char is guaranteed by the standard to be 1 byte in size
c->name = malloc(strlen(s)+1);
if (c->name == NULL)
exit( EXIT_FAILURE );//could not allocate mem
c->name[0] = '\0';//set to empty string, now we can use safer functions:
strncat(c->name, s, strlen(s));
After you address these issues, seriously re-think your approach, and ask yourself what it is you're actually trying to do here:
for(i=1;i<=10;i++)
{
d=(struct word*)malloc(sizeof(struct word));
if(d==NULL)
exit(0);
scanf("%s",s);
printf("4");
s=d->name;
}
You're allocating a struct 10 times, each time re-assigning it to d. You never free this memory, though. which is bad practice.
Again: don't cast the return of malloc, but that's the least of your worries.
if (d == NULL)
exit(0);
Ok, now you check the return of malloc. Great. But why on earth are you terminating with 0 (indicative of a successful run). There's a macro for this, too. You could've written:
if (d == NULL)
exit( EXIT_SUCCESS);
Clearly, EXIT_SUCCESS is not what you should communicate.
that const char *s is now being used to store user input. That's not going to work, though, as it points to read-only memory, so forget about the unsafe scanf("%s", s); statement. Use a stack variable, and make sure the input buffer is cleared, or use a safe alternative.
But then you go and do something as absurd as this:
s = d->name;
Again, d->name, like in the case with c, is an invalid pointer. Why assign it to s here? there's no point, no reason... only madness.
Bottom line: Kill this code before it hatches, start again, and please use these tips/recommendations and critiques as a guideline.
I have no idea why you're getting a 'too many open files', but this line:
strcpy(c->name,s)
is writing data to random memory, which could cause all kinds of problems.
You need to malloc() that c->name first.
Also that scanf to s looks suspicious, and d->name is never assigned anything either.
The reason that you're getting 'too many open files' is probably because some memory is getting overwritten in such a way that just happens to trigger that particular error. Welcome to the world of undefined behaviour. IE: If you overwrite random memory, basically anything can happen.
The first bug is in the line
strcpy(c->name,s);
At that point, c->name is an uninitialised pointer so the program will crash if you are lucky.
Reading your comment: You fixed the second bug. The first bug is still unfixed. And there's the third bug in the line
s=d->name;
This string copy will run off through memory, starting at whatever c->name points to until it finds a null terminator.
strcpy(c->name,s);
You have allocated space for c but not for the name pointer in c.
c->name = malloc([some length]);
c->name points somewhere, but you don't know where until you malloc it. That's why you're getting a seemingly random error, because your executing a string copy from an unknown location for an unknown number of bytes and you are clobbering whatever s points to for an unknown number of bytes.

Using fgets() with char* type

I have a simple question about using fgets() with char* string.
....
char *temp;
FILE fp=fopen("test.txt", "r");
fgets(temp, 500, fp);
printf("%s", temp);
....
This code didn't work well.
But after I modified char *temp to char temp[100];, the code worked well as I intended.
What is the difference between those two?
When I googled it, some said that memory must be allocated to char * using malloc()...
But I couldn't understand it.
char *temp is only a pointer. At begin it doesn't points to anything, possibly it has a random value.
fgets() reads 500 bytes from fp to the memory addresse, where this temp pointer points! So, it can overwrite things, it can make segmentation faults, and only with a very low chance will be work relativale normally.
But char temp[500] is a 500 bytes long array. That means, that the compiler does the allocation on the beginning of your process (or at the calling of your function). Thus this 500 bytes will be a useable 500 bytes, but it has a price: you can't reallocate, resize, free, etc. this.
What the google wants from you, is this:
char *temp = (char*)malloc(500);
And a
free(temp);
after you don't need this any more.
When we write
char *temp ;
it means temp is an uninitialized pointer to char i.e. currently it does not contain any address in it .
While using fgets you have to pass a string in which the bytes read from file pointer is to be copied . link
since the temp is uninitialized , the fgets looks like this
fgets(<no string> , 500 , fp ) ;
which is invalid .
Hence , we should give initialized string which can be formed as :
1) char *temp = malloc(sizeof(500)) ;
or
2) char temp[500] ;
Hence if we pass initialized string to fgets , it would look like
fgets( < some string > , 500 , fp) ;
char *temp is uninitialized, that is, it isn't pointing to valid memory. Either make it an array (char temp[]) or use malloc to assign memory for it.

Error reading char* type from .DAT file with C

So, for some reason, I need to make a external file (.DAT) to store data by appending the new one to the end of old data.
#include <stdio.h>
#include <stdlib.h>
int main () {
typedef struct {
char *Name;
int Index;
} DataFile;
static FILE *file;
size_t result;
DataFile *DataTable;
file = fopen("database.DAT","ab");
DataTable = (DataFile *) malloc (sizeof(DataFile));
DataTable[0].Name = "somefile.txt";
DataTable[0].Index = 7;
printf("%s %d \n",DataTable[0].Name,DataTable[0].Index);
result = fwrite(DataTable,sizeof(DataFile),1,file);
fclose(file);
free(DataTable);
return 0;
}
After running code above, I then check if the data stored correctly. So, I make this code below.
#include <stdio.h>
#include <stdlib.h>
int main () {
typedef struct {
char *Name;
int Index;
} DataFile;
static FILE *file;
size_t result;
long size;
int i;
DataFile *DataTable;
file = fopen("database.DAT","rb");
if (file == NULL) printf("Error1");
// Determine the size of file
fseek(file,0,SEEK_END);
size = ftell(file);
rewind(file);
DataTable = (DataFile *) malloc ((size/sizeof(DataFile)) * sizeof(DataFile));
if (DataTable == NULL) printf("Error2");
result = fread(DataTable,sizeof(DataFile),size/sizeof(DataFile),file);
fclose(file);
for (i=0; i<result; i++) {
printf("%s %d \n",DataTable[i].Name,DataTable[i].Index);
}
free(DataTable);
return 0;
}
However, it gives output
somefile.txt 7
from the first code block and
Error1 7
from the second code block.
I notice that the problem is not because the failure either when opening .DAT file or when allocating memory for DataTable. Also, it works for int type (Index) but not for char* type (Name) when reading from .DAT file. I have no idea what to do to solve this char*-type-reading problem (and where 'error1' comes from). (not even google gives me answer.)
Your structure DataFile stores one pointer and one integer. When you write it to the file, you write some program specific pointer to a string, and an integer.
When reading from it, you just refill your structure with the pointer and the integer, wich means that DataFile.Name will be a pointer to a probably-not-initialized memory segment. But since you created your file pointing to the first hard-coded string ("filename.txt"), some undefined but understandable behaviour happens, and your pointer in this case points to the first hard-coded string you wrote in you second program (which in your case is Error1)
What you really want to do is write the real string in your file.
A simple solution, if you want to the keep the hole writing structure thing is to create an array instead of a pointer
typedef struct {
char Name[512];
int Index;
} DataFile;
then initialize your data with
strncpy(DataTable[0].Name, "somefile.txt", sizeof(DataTable[0].Name) - 1); // just to make sure you dont overflow your array size
DataTable[0].Name[sizeof(DataTable[0].Name) - 1] = '\0';
and retreview your data the way you did.
A char* is only a pointer, i.e. the address of the character array containing your strings. You don't write the strings themselves to the file. After reading the file, as the same strings aren't in your memory at the same addresses any more, the application will fail.
You'll have to come up with a way to save the strings themselves to file as well. Probably by first writing their length, and then writing their content. Upon reading, you can use the length information to allocate memory dynamically, then read into that memory.
In your writing code you haven't allocated storage for char *Name. When you perform the DataTable[0].Name = "somefile.txt" instruction you're not actually copying the "somefile.txt" into memory pointed by Name, it's actually assigning a Name a value pointing to a constant characters string (moreover, it will become dangling pointer since the string is an rvalue, i.e. doesn't have a memory to be addressed via). Same goes for your file reading code.
You need to:
Allocate storage for your Name.
Copy the string using memcpy or similar into the allocated storage.

fread/fwrite string in C

I have a binary file which contains records. The structure of the file is as such:
Structure (see below)
Name String
Address String
The structure in question:
typedef struct{
char * name;
char * address;
short addressLength, nameLength;
int phoneNumber;
}employeeRecord;
employeeRecord record;
I get the name as such:
char name[50];
printf("\nName:");
fgets(name,50,stdin);
record.nameLength = strlen(name)-1;
record.name = malloc(sizeof(char)*record.nameLength);
strcpy(record.name,name);
I write the structure, the the name, then the address (as mentioned above).
fwrite(&record.name,sizeof(char),record.nameLength,fp);
where fp is a file pointer.
Now i close the file.
However, if i then want to read from the file to get this data back, I believe I need to read in the structure, read the nameLength variable, malloc enough memory for the name to sit in, then fread the name into the variable.
Like so:
char *nameString = malloc(sizeof(char)*record.nameLength);
fread(nameString,sizeof(char),record.nameLength,fp);
printf("\nName: %s",nameString);
However, when i attempt this, i do not get valid data.
Example:
Input name is: Joseph (6 characters)
Output data:
Name length is 6 (correct),
Name is �A � (aka garbage)
So obviously im doing something wrong. Could someone give me some help?
I see two problems with the write, you are setting record.nameLength to be too small, and you are passing the wrong pointer to fwrite for the name. record.name is already a pointer.
Change this
record.nameLength = strlen(name)-1;
...
fwrite(&record.name,sizeof(char),record.nameLength,fp);
to this
record.nameLength = strlen(name);
...
fwrite(record.name,sizeof(char),record.nameLength,fp);
You also have a problem on the read, since you aren't writing the terminating \0 from the strings into your file, when you read back, you need to add that terminator explicitly.
char *nameString = malloc(sizeof(char)* (record.nameLength + 1));
fread(nameString,sizeof(char),record.nameLength,fp);
nameString[record.NameLength] = '\0';
The problem is that you pass the pointer to the char* in your fwrite:
fwrite(&record.name,sizeof(char),record.nameLength,fp);
This means that instead of writing the name, you're writing the memory address of the name. Fwrite expects a pointer to the data to write—in your case, that's the pointer to the char data, not the pointer to the pointer of the char data.
Pass it record.name instead of &record.name and you should be set:
fwrite(record.name, sizeof(char), record.nameLength, fp);

Resources