[erlang-questions] file:read_file/2 bug?

[Per Hedeland]

"Christian S" <chsu79@REDACTED> wrote:
>
>I dont understand why everything is done in efile_openfile that is
>done, but my guess is that if one first open() the file then use
>fstat() on the file descriptor, then the proc filesystem would report
>nonzero file length.

Your guess is wrong:

$ cat stdin_size.c

#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <stdio.h>

main()
{
    struct stat buf;

    fstat(0, &buf);
    printf("size=%d\n", buf.st_size);
}

$ ./stdin_size < /proc/self/environ 
size=0

>The linux kernel generates the proc file's content as a snapshot at
>the instance the file is opened. Before open() it doesnt know how long
>the content will be.

AFAIK, the snapshotting happens when you read(2) the file, which can
have a variety of interesting effects (probably not in this particular
case, but in general). Basically, the Linux /proc file system is handy
for the "I just want to have a peek at some system stuff" scenario, but
if you want to use it programmatically and get dependable results, you
need to be *very* careful. I don't think any of the Erlang file(3)
functionality is quite up to the general task.

--Per