From: GSP <palmergs_no_spam@...>
Date: 2001-07-13T04:13:48+09:00
Subject: [ruby-talk:17743] Re: Funky file.each_byte

What you're running into, I think, is the fact that newlines in ms-dos are
actually two characters (ascii 10 and 13, I think).  When you're reading
your file, the file system read's 10 and 13 and gives you one byte "\n".

I tried a quick test using your code

temp.txt  = "
1234567890
1234567890
1234567890
1234567890
1234567890
1234567890
"

with 6 newlines.  The output:

c:/temp.txt had 66 bytes via .each_byte,
while file.stat.size reports 72 bytes.

= 6 bytes difference.  As Park Hee-Sob mentioned, reading in binary
(f.binmode) format ignores this and gives you the exact number of bytes in
the file.

GSP

"Tyler" <nlper@junglemate.com> wrote in message
news:65d29696.0107111312.4a1e488d@posting.google.com...
> Howdy!  I was filtering some files and ran into an anomaly.
> I'm running Ruby 1.6.4 on an old Windows machine. Everything
> was fine when I was manipulating bytes with getc/putc, but that
> seemed too C-ish, so I tried a tweak to a Ruby idiom. Imagine
> my surprise and delight to discover that a file with some bytes
> deleted was larger than the original file!  Ditto for reading.
>
>   count = 0
>   f = File.new( ARGV[0], "r")
>   f.each_byte do
>     count += 1
>   end
>   f.close
>
>   printf( "%s had %d bytes via .each_byte,\n", ARGV[0], count)
>   printf( "while file.stat.size reports %d bytes.\n",
File.stat(ARGV[0]).size)
>
> generates the following results:
>
>   H:\ruby >ruby fsize.rb 1MEG.1
>   1MEG.1 had 1048553 bytes via .each_byte,
>   while file.stat.size reports 1048576 bytes.
>
> Any ideas? Is this a Windows-specific oddity or showing up elsewhere?
>
> TIA, Tyler
>