I took over a project that use the following function to read files:
char *fetchFile(char *filename) {
char *buffer;
int len;
FILE *f = fopen(filename, "rb");
if(f) {
if(verbose) {
fprintf(stdout, "Opened file %s successfully\n", filename);
}
fseek(f, 0, SEEK_END);
len = ftell(f);
fseek(f, 0, SEEK_SET);
if(verbose) {
fprintf(stdout, "Allocating memory for buffer for %s\n", filename);
}
buffer = malloc(len + 1);
if(buffer) fread (buffer, 1, len, f);
fclose (f);
buffer[len] = '\0';
} else {
fprintf(stderr, "Error reading file %s\n", filename);
exit(1);
}
return buffer;
}
The rb mode is used because sometimes the file can be a spreadsheet and therefore I want the information as in a text file.
The program runs on a linux machine but the files to read come from linux and windows.
I am not sure of what approach is better to not have windows line ending mess with my code.
I was thinking of using dos2unix at the start of this function.
I also thought of opening in r mode, but I believe that could potentially mess things up when opening non-text files.
I would like to understand better the differences between using:
dos2unix,
r vs rb mode,
or any other solution which would fit
better the problem.
Note: I believe that I understand r vs rb modes, but if you could explain why it is a bad or good solution for this specific situation (I think it wouldn't be good because sometimes it opens spreadsheets but I am not sure of that).
If my understanding is correct the rb mode is used because sometimes the file can be a spreadsheet and therefore the programs just want the information as in a text file.
You seem uncertain, and though perhaps you do understand correctly, your explanation does not give me any confidence in that.
C knows about two distinct kinds of streams: binary streams and text streams. A binary stream is simply an ordered sequence of bytes, written and / or read as-is without any kind of transformation. On the other hand,
A text stream is an ordered sequence of characters composed into
lines, each line consisting of zero or more characters plus a
terminating new-line character. Whether the last line requires a
terminating new-line character is implementation-defined. Characters
may have to be added, altered, or deleted on input and output to
conform to differing conventions for representing text in the host
environment. Thus, there need not be a one- to-one correspondence
between the characters in a stream and those in the external
representation. [...]
(C2011 7.21.2/2)
For some implementations, such as POSIX-compliant ones, this is a distinction without a difference. For other implementations, such as those targeting Windows, the difference matters. In particular, on Windows, text streams convert on the fly between carriage-return / line-feed pairs in the external representation and newlines (only) in the internal representation.
The b in your fopen() mode specifies that the file should be opened as a binary stream -- that is, no translation will be performed on the bytes read from the file. Whether this is the right thing to do depends on your environment and the application's requirements. This is moot on Linux or another Unix, however, as there is no observable difference between text and binary streams on such systems.
dos2unix converts carriage-return / line-feed pairs in the input file to single line-feed (newline) characters. This will convert a Windows-style text file or one with mixed Windows / Unix line terminators to Unix text file convention. It is irreversible if there are both Windows-style and Unix-style line terminators in the file, and it is furthermore likely to corrupt your file if it is not a text file in the first place.
If your inputs are sometimes binary files then opening in binary mode is appropriate, and conversion via dos2unix probably is not. If that's the case and you also need translation for text-file line terminators, then you first and foremost need a way to distinguish which case applies for any particular file -- for example, by command-line argument or by pre-analyzing the file via libmagic. You then must provide different handling for text files; your main options are
Perform the line terminator conversion in your own code.
Provide separate versions of the fetchFile() function for text and binary files.
The code just copies the contents of a file to an allocated buffer. The UNIX way (YMMV) is to just memory map the file instead of reading it. Much faster.
// untested code
void* mapfile(const char *name)
{
int fd;
struct stat st;
if ((fd = open(name, O_RDONLY)) == -1)
return NULL;
if (fstat(fd, &st)) {
close(fd);
return NULL;
}
void *p = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, 0, fd);
close(fd);
if (p == (void *)MAP_FAILED)
p = NULL;
return p;
}
Something along these lines will work. Adjust settings if you want to write to the file as well.
Related
What translation occurs when writing to a file that was opened in text mode that does not occur in binary mode? Specifically in MS Visual C.
unsigned char buffer[256];
for (int i = 0; i < 256; i++) buffer[i]=i;
int size = 1;
int count = 256;
Binary mode:
FILE *fp_binary = fopen(filename, "wb");
fwrite(buffer, size, count, fp_binary);
Versus text mode:
FILE *fp_text = fopen(filename, "wt");
fwrite(buffer, size, count, fp_text);
I believe that most platforms will ignore the "t" option or the "text-mode" option when dealing with streams. On windows, however, this is not the case. If you take a look at the description of the fopen() function at: MSDN, you will see that specifying the "t" option will have the following effect:
line feeds ('\n') will be translated to '\r\n" sequences on output
carriage return/line feed sequences will be translated to line feeds on input.
If the file is opened in append mode, the end of the file will be examined for a ctrl-z character (character 26) and that character removed, if possible. It will also interpret the presence of that character as being the end of file. This is an unfortunate holdover from the days of CPM (something about the sins of the parents being visited upon their children up to the 3rd or 4th generation). Contrary to previously stated opinion, the ctrl-z character will not be appended.
In text mode, a newline "\n" may be converted to a carriage return + newline "\r\n"
Usually you'll want to open in binary mode. Trying to read any binary data in text mode won't work, it will be corrupted. You can read text ok in binary mode though - it just won't do automatic translations of "\n" to "\r\n".
See fopen
Additionally, when you fopen a file with "rt" the input is terminated on a Crtl-Z character.
Another difference is when using fseek
If the stream is open in binary mode, the new position is exactly offset bytes measured from the beginning of the file if origin is SEEK_SET, from the current file position if origin is SEEK_CUR, and from the end of the file if origin is SEEK_END. Some binary streams may not support the SEEK_END.
If the stream is open in text mode, the only supported values for offset are zero (which works with any origin) and a value returned by an earlier call to std::ftell on a stream associated with the same file (which only works with origin of SEEK_SET.
Even though this question was already answered and clearly explained, I think it would be interesting to show the main issue (translation between \n and \r\n) with a simple code example. Note that I'm not addressing the issue of the Crtl-Z character at the end of the file.
#include <stdio.h>
#include <string.h>
int main() {
FILE *f;
char string[] = "A\nB";
int len;
len = strlen(string);
printf("As you'd expect string has %d characters... ", len); /* prints 3*/
f = fopen("test.txt", "w"); /* Text mode */
fwrite(string, 1, len, f); /* On windows "A\r\nB" is writen */
printf ("but %ld bytes were writen to file", ftell(f)); /* prints 4 on Windows, 3 on Linux*/
fclose(f);
return 0;
}
If you execute the program on Windows, you will see the following message printed:
As you'd expect string has 3 characters... but 4 bytes were writen to file
Of course you can also open the file with a text editor like Notepad++ and see yourself the characters:
The inverse conversion is performed on Windows when reading the file in text mode.
We had an interesting problem with opening files in text mode where the files had a mixture of line ending characters:
1\n\r
2\n\r
3\n
4\n\r
5\n\r
Our requirement is that we can store our current position in the file (we used fgetpos), close the file and then later to reopen the file and seek to that position (we used fsetpos).
However, where a file has mixtures of line endings then this process failed to seek to the actual same position. In our case (our tool parses C++), we were re-reading parts of the file we'd already seen.
Go with binary - then you can control exactly what is read and written from the file.
In 'w' mode, the file is opened in write mode and the basic coding is 'utf-8'
in 'wb' mode, the file is opened in write -binary mode and it is resposible for writing other special characters and the encoding may be 'utf-16le' or others
What translation occurs when writing to a file that was opened in text mode that does not occur in binary mode? Specifically in MS Visual C.
unsigned char buffer[256];
for (int i = 0; i < 256; i++) buffer[i]=i;
int size = 1;
int count = 256;
Binary mode:
FILE *fp_binary = fopen(filename, "wb");
fwrite(buffer, size, count, fp_binary);
Versus text mode:
FILE *fp_text = fopen(filename, "wt");
fwrite(buffer, size, count, fp_text);
I believe that most platforms will ignore the "t" option or the "text-mode" option when dealing with streams. On windows, however, this is not the case. If you take a look at the description of the fopen() function at: MSDN, you will see that specifying the "t" option will have the following effect:
line feeds ('\n') will be translated to '\r\n" sequences on output
carriage return/line feed sequences will be translated to line feeds on input.
If the file is opened in append mode, the end of the file will be examined for a ctrl-z character (character 26) and that character removed, if possible. It will also interpret the presence of that character as being the end of file. This is an unfortunate holdover from the days of CPM (something about the sins of the parents being visited upon their children up to the 3rd or 4th generation). Contrary to previously stated opinion, the ctrl-z character will not be appended.
In text mode, a newline "\n" may be converted to a carriage return + newline "\r\n"
Usually you'll want to open in binary mode. Trying to read any binary data in text mode won't work, it will be corrupted. You can read text ok in binary mode though - it just won't do automatic translations of "\n" to "\r\n".
See fopen
Additionally, when you fopen a file with "rt" the input is terminated on a Crtl-Z character.
Another difference is when using fseek
If the stream is open in binary mode, the new position is exactly offset bytes measured from the beginning of the file if origin is SEEK_SET, from the current file position if origin is SEEK_CUR, and from the end of the file if origin is SEEK_END. Some binary streams may not support the SEEK_END.
If the stream is open in text mode, the only supported values for offset are zero (which works with any origin) and a value returned by an earlier call to std::ftell on a stream associated with the same file (which only works with origin of SEEK_SET.
Even though this question was already answered and clearly explained, I think it would be interesting to show the main issue (translation between \n and \r\n) with a simple code example. Note that I'm not addressing the issue of the Crtl-Z character at the end of the file.
#include <stdio.h>
#include <string.h>
int main() {
FILE *f;
char string[] = "A\nB";
int len;
len = strlen(string);
printf("As you'd expect string has %d characters... ", len); /* prints 3*/
f = fopen("test.txt", "w"); /* Text mode */
fwrite(string, 1, len, f); /* On windows "A\r\nB" is writen */
printf ("but %ld bytes were writen to file", ftell(f)); /* prints 4 on Windows, 3 on Linux*/
fclose(f);
return 0;
}
If you execute the program on Windows, you will see the following message printed:
As you'd expect string has 3 characters... but 4 bytes were writen to file
Of course you can also open the file with a text editor like Notepad++ and see yourself the characters:
The inverse conversion is performed on Windows when reading the file in text mode.
We had an interesting problem with opening files in text mode where the files had a mixture of line ending characters:
1\n\r
2\n\r
3\n
4\n\r
5\n\r
Our requirement is that we can store our current position in the file (we used fgetpos), close the file and then later to reopen the file and seek to that position (we used fsetpos).
However, where a file has mixtures of line endings then this process failed to seek to the actual same position. In our case (our tool parses C++), we were re-reading parts of the file we'd already seen.
Go with binary - then you can control exactly what is read and written from the file.
In 'w' mode, the file is opened in write mode and the basic coding is 'utf-8'
in 'wb' mode, the file is opened in write -binary mode and it is resposible for writing other special characters and the encoding may be 'utf-16le' or others
What is the different between these blocks of code. I tried to search for "wb" but don't see it anywhere. The file containing "wb" is from on of my tutors
FILE *f = fopen(DB_FILE_NAME, "wb");
if (f == NULL) {
printf("Write error\n");
} else {
/* write n_students elements of the studentlist array */
fwrite(studentlist, sizeof(student_t), n_students, f);
fclose(f);
}
and
FILE *f = fopen(DB_FILE_NAME, "w");
if (f == NULL) {
printf("Write error\n");
} else {
/* write n_students elements of the studentlist array */
fwrite(studentlist, sizeof(student_t), n_students, f);
fclose(f);
}
Specifying "b" in the access mode prevents (some implementations of) the standard library from translating a few characters when reading/writing to the file.
Most common translation is for end of line: \n is translated to \r\n in Windows.
Absolutely any reference on the fopen() function would have told you this. For instance the manual page which is the common documentation used in Unix-like environments:
The mode string can also include the letter 'b' either as a last character
or as a character between the characters in any of the two-character
strings described above. This is strictly for compatibility with C89 and
has no effect; the 'b' is ignored on all POSIX conforming systems,
including Linux. (Other systems may treat text files and binary files
differently, and adding the 'b' may be a good idea if you do I/O to a
binary file and expect that your program may be ported to non-UNIX
environments.)
So, it stands for binary and is useful to indicate that you intend to treat the contents of the file as not being text.
For your code, binary access seems right. However, directly writing raw struct values is generally a very bad idea, since you don't know the exact internal format used by the compiler and it can change unexpectedly. For files that should be shared and/or accessed "later", this is not the proper way to do it in C. Look into serialization.
In fopen documentation :
With the mode specifiers above the file is open as a text file. In order to open a file as a binary file, a "b" character has to be included in the mode string. This additional "b" character can either be appended at the end of the string (thus making the following compound modes: "rb", "wb", "ab", "r+b", "w+b", "a+b") or be inserted between the letter and the "+" sign for the mixed modes ("rb+", "wb+", "ab+").
In C, we can find the size of file using fseek() function. Like,
if (fseek(fp, 0L, SEEK_END) != 0)
{
// Handle repositioning error
}
So, I have a question, Is it recommended method for computing the size of a file using fseek() and ftell()?
If you're on Linux or some other UNIX like system, what you want is the stat function:
struct stat statbuf;
int rval;
rval = stat(path_to_file, &statbuf);
if (rval == -1) {
perror("stat failed");
} else {
printf("file size = %lld\n", (long long)statbuf.st_size;
}
On Windows under MSVC, you can use _stati64:
struct _stati64 statbuf;
int rval;
rval = _stati64(path_to_file, &statbuf);
if (rval == -1) {
perror("_stati64 failed");
} else {
printf("file size = %lld\n", (long long)statbuf.st_size;
}
Unlike using fseek, this method doesn't involve opening the file or seeking through it. It just reads the file metadata.
The fseek()/ftell() works sometimes.
if (fseek(fp, 0L, SEEK_END) != 0)
printf("Size: %ld\n", ftell(fp));
}
Problems.
If the file size exceeds about LONG_MAX, long int ftell(FILE *stream) response is problematic.
If the file is opened in text mode, the return value from ftell() may not correspond to the file length. "For a text stream, its file position indicator contains unspecified information," C11dr ยง7.21.9.4 2
If the file is opened in binary mode, fseek(fp, 0L, SEEK_END) is not well defined. "Setting the file position indicator to end-of-file, as with fseek(file, 0, SEEK_END), has undefined behavior for a binary stream (because of possible trailing null characters) or for any stream with state-dependent encoding that does not assuredly end in the initial shift state." C11dr footnote 268. #Evert This most often applies to earlier platforms than today, but it is still part of the spec.
If the file is a stream like a serial input or stdin, fseek(file, 0, SEEK_END) makes little sense.
The usual solution to finding file size is a non-portable platform specific one. Example good answer #dbush.
Note: If code attempts to allocate memory based on file size, the memory available can easily be exceeded by the file size.
Due to these issues, I do not recommend this approach.
Typically the problem should be re-worked to not need to find the file size, but to grow the data as more input is processed.
LL disclaimer: Note that C spec footnotes are informative and so not necessarily normative.
The best method in my opinion is fstat(): https://linux.die.net/man/2/fstat
Well, you can estimate the size of a file in several ways:
You can read(2) the file from the beginning to the end, and the number or chars read is the size of the file. This is a tedious way of getting the size of a file, as you have to read the whole file to get the size. But if the operating system doesn't allow to position the file pointer arbitrarily, then this is the only way to get the file size.
Or you can move the pointer at the end of file position. This is the lseek(2) you showed in the question, but be careful that you have to do the system call twice, as the value returned is the actual position before moving the pointer to the desired place.
Or you can use the stat(2) system call, that will tell you all the administrative information of the file, like the owner, group, permissions, size, number of blocks the file occupies in the disk, disk this file belongs to, number of directory entries pointing to it, etc. This allows you to get all this information with only one syscall.
Other methods you point (like the use of the ftell(3) stdio library call) will work also (with the same problem that it results in two system calls to set and retrieve/restore the file pointer) but have the problem of involving libraries that probably you are not using for anything else. It should be complicated to get a FILE * pointer (e.g. fdopen(3)) on a int file descriptor, just to be able to use the ftell(3) function on it (twice), and then fclose(3) it again.
I am designing an image decoder and as a first step I tried to just copy the using c. i.e open the file, and write its contents to a new file. Below is the code that I used.
while((c=getc(fp))!=EOF)
fprintf(fp1,"%c",c);
where fp is the source file and fp1 is the destination file.
The program executes without any error, but the image file(".bmp") is not properly copied. I have observed that the size of the copied file is less and only 20% of the image is visible, all else is black. When I tried with simple text files, the copy was complete.
Do you know what the problem is?
Make sure that the type of the variable c is int, not char. In other words, post more code.
This is because the value of the EOF constant is typically -1, and if you read characters as char-sized values, every byte that is 0xff will look as the EOF constant. With the extra bits of an int; there is room to separate the two.
Did you open the files in binary mode? What are you passing to fopen?
It's one of the most "popular" C gotchas.
You should use freadand fwrite using a block at a time
FILE *fd1 = fopen("source.bmp", "r");
FILE *fd2 = fopen("destination.bmp", "w");
if(!fd1 || !fd2)
// handle open error
size_t l1;
unsigned char buffer[8192];
//Data to be read
while((l1 = fread(buffer, 1, sizeof buffer, fd1)) > 0) {
size_t l2 = fwrite(buffer, 1, l1, fd2);
if(l2 < l1) {
if(ferror(fd2))
// handle error
else
// Handle media full
}
}
fclose(fd1);
fclose(fd2);
It's substantially faster to read in bigger blocks, and fread/fwrite handle only binary data, so no problem with \n which might get transformed to \r\n in the output (on Windows and DOS) or \r (on (old) MACs)