10. File input/output

File as a stream

• teletype
• serial
• tape
• ...
• File IRL: see prev. lection
• Why stream: sequential nature of I/O

C files

• C: FILE * abstraction

  • You may try cc -E program_with_FILE.c, but it's a complex structure

  • Buffer tricks like ungetc() or scanf() next character preview

  • text/binary abstraction

OS files

• Linux: stream file I/O as descriptors
  • Descriptor: integer stream number
  • All other stuff is hidden by OS
  • 0, 1, 2 — stdin, stdout, stderr, already opened

• ssize_t read(int fd, void *buf, size_t count);

  • note ssize_t type

  • note void *buf

  • also ssize_t write(int fd, const void *buf, size_t count);

       1 #include <stdio.h>
       2 #include <unistd.h>
       3 
       4 int main(int argc, char *argv[]) {
       5         char c;
       6         while(read(0,  &c, 1)) {
       7                 c++;
       8                 write(1, &c, 1);
       9         }
      10 
      11         return 0;
      12 }
    

    • Try to use Tab keys

    • Press ^D at the beginning of the line to notify OS that stdin is closed (see stty --all eof parameter)

    • Try to change char to int (and 1 to … what?). Describe the result

• int open(const char *pathname, int flags);()

  • returns first free descriptor, usually 3

  • flags: (many, see man)

• close

• off_t lseek(int fd, off_t offset, int whence);

  • Return value and SEEK_SET/SEEK_CUR/SEEK_END
  • Cant' seek on a stream!

  • Hole in the file

Files and filesystems

Filesystem objects:

• files
• directories
• symlinks
• devices
• pipes (aka FIFOs)

• sockets (aka i've got a hole in me pocket)

  • async!
• …

There's also f- versions of tell/stat/whatever over C file abstraction

== C: using command line arguments =

• int main(int argc, char *argv[]):

  0. argc is command line size in words (always ⩾ 1)

  1. argv[0] always present, it's the program name

  2. argv[1] etc. are commandline words

     • these are four words brcause of quotation: qwe   "ads df d"  sdfsdf   " sdfsd ,kuii  iu"

• char *argv[] is an array of char *, each pointing to corresponded ASCIIZ data block

  • argv[k] when k>argc is undefined

H/W

TODO