in reply to Re^3: named pipe and "Text file busy"
in thread named pipe and "Text file busy"

The expected behavior will be for the writer to block (unless open in non-blocking mode) - not to exit with an error. You can test this in the shell with "echo" and "cat". "text file busy" results after the writer tries to open the FIFO - it does not even get to the writing stage. I can confirm that with the C program i wrote that tries to open the FIFO with "open" system call and then returns errno 26 - "text file busy".

Replies are listed 'Best First'.
Re^5: named pipe and "Text file busy"
by tuxz0r (Pilgrim) on Nov 25, 2007 at 21:35 UTC
    I must be missing something here. Can you post some code? I tried tail -fon the FIFO in one terminal, and then wrote a C program which opens the FIFO in "w" write mode. If the open fails, I use strerror(errno) to show the message and I don't get "Text file busy".

    Here's the small C program to write to the FIFO I used:

    #include <stdio.h> #include <errno.h> #include <fcntl.h> #include <string.h> int main(void) { char buf[100]; int fd = open("file.fifo", O_NONBLOCK|O_WRONLY); if (fd < 0) { fprintf(stderr, "error (%d): %s\n", errno, strerror(er +rno)); return 1; } sprintf(buf, "%s\n", "a new line"); int bytes = write(fd, buf, strlen(buf) + 1); printf("wrote %d bytes", bytes); close(fd); return 0; }
    Is this how you are opening the file? When I run this program without the tail -f in the other window, I get ERRNO 6, "Device not configured", not "Text file busy". If I use open without the O_NONBLOCK the program hangs in the open as we both expect.

    ---
    echo S 1 [ Y V U | perl -ane 'print reverse map { $_ = chr(ord($_)-1) } @F;'
    Warning: Any code posted by tuxz0r is untested, unless otherwise stated, and is used at your own risk.

      Yes, that's exactly how i open the file. I should note that "text file buys" only happens after i open the pipe with my perl script and terminate it. After that, this error occurs irrespective of whether later i've got the script (reader) running together with the writer or a writer by itself.
      To sum up, consider scenarios:

      correct behavior:
      1 echo "data" > file.fifo. Echo blocks on write, given that there's no reader running
      2 start the script that opens a file.fifo and waits for data to come in
      2.in a different terminal- echo "data" > file.fifo
      3. Data is correctly received by the script
      4.kill the script with Ctrl-C

      Wrong behavior
      After all of the above, perform the following:
      a. script is not running. echo "data" > file.fifo results "text file busy".
      b. start the script. The scripts open file.fifo for reading successfully.
      c. echo "data" > file.fifo results "text file busy"
      The script itself is quite big as it does some other processing but the actual code which operates the FIFO is very basic and pretty much boils down to (snippet):
      my $alertsthr = threads->new(\&al_thread); $SIG{'INT'} = 'CLEANUP'; sub CLEANUP { print "\nClosing pipe\n"; close PIPE; exit(); } sub al_thread { my $myline; while (1) # keep reading from the pipe until we're killed { print "$config->{FIFO} open\n" if $config->{VERBOSE}; open (PIPE, "$config->{FIFO}") or die "Could not open the input +pipe $config->{FIFO}\n"; while($myline = <PIPE>) { print "read $myline\n" if $config->{VERBOSE}; process_message($myline); } print "$config->{FIFO} EOF\n" if $config->{VERBOSE}; } }
        I'm really at a loss, since I can't re-create this. I've never seen "Text file busy" in relation to a named pipe error. Did you find any better answers in comp.lang.perl.misc, yet (I noticed you had posted there as well)?

        ---
        echo S 1 [ Y V U | perl -ane 'print reverse map { $_ = chr(ord($_)-1) } @F;'
        Warning: Any code posted by tuxz0r is untested, unless otherwise stated, and is used at your own risk.