File.size? incorrect for big files?

Hi,

I've written a script to monitor some files stored on our network. The
files range from a few hundred MB to 20-50g. Above a certain point
which I haven't really narrowed down yet, the file sizes start being
reported incorrectly from my script. It's as if the routine that does
the system query is using a 32 bit thing for the size and it's getting
truncated as integers do. Some files actually come back with a negative
size. The data type that ruby is using is perfectly capable of holding
the size of the files. I've tried File.size? and FileUtils.size?, so
I'm somewhat at a loss short of writing an extension. Any ideas?

The script is running on windows, querying the files via UNC paths
through a samba share on linux.

Thanks!

···

________________________________

Carey Nation <mailto:carey.nation@cnn.com>

Lead Software Engineer

CNN BEST

Broadcast Production Systems

Video Solutions Group

(404)827-2935 (wk)

(404)824-0033 (cell)

________________________________

Some versions, maybe all, of the windows libraries have a bug in
File.size.
Make sure you have the latest win32 gems, it seems to help.

The 'gem outdated' and 'gem update' commands can help.

The only bug I've seen is on moderately sized files where the ruby code
closes the file then immediately checks the file size. It came back as
zero. I worked around it several versions ago, so my info is not as of
today.

···

--
Posted via http://www.ruby-forum.com/.

It's a bug in Ruby.

The win32-file library fixes it.

require 'win32/file' # Do this first
File.size(some_big_file)

Regards,

Dan

···

On Mar 13, 12:59 pm, "Nation, Carey" <Carey.Nat...@turner.com> wrote:

Hi,

I've written a script to monitor some files stored on our network. The
files range from a few hundred MB to 20-50g. Above a certain point
which I haven't really narrowed down yet, the file sizes start being
reported incorrectly from my script. It's as if the routine that does
the system query is using a 32 bit thing for the size and it's getting
truncated as integers do. Some files actually come back with a negative
size. The data type that ruby is using is perfectly capable of holding
the size of the files. I've tried File.size? and FileUtils.size?, so
I'm somewhat at a loss short of writing an extension. Any ideas?

The script is running on windows, querying the files via UNC paths
through a samba share on linux.

Hi,

At Fri, 14 Mar 2008 03:59:30 +0900,
Nation, Carey wrote in [ruby-talk:294466]:

The script is running on windows, querying the files via UNC paths
through a samba share on linux.

Microsoft Visual C's runtime doesn't support large files, so
ruby compiled with it also doesn't. Ruby 1.9 replaces some
functions to fix it.

···

--
Nobu Nakada

That did it! Thanks!

···

-----Original Message-----
From: Daniel Berger [mailto:djberg96@gmail.com]
Sent: Thursday, March 13, 2008 3:35 PM
To: ruby-talk ML
Subject: Re: File.size? incorrect for big files?

On Mar 13, 12:59 pm, "Nation, Carey" <Carey.Nat...@turner.com> wrote:

Hi,

I've written a script to monitor some files stored on our network. The
files range from a few hundred MB to 20-50g. Above a certain point
which I haven't really narrowed down yet, the file sizes start being
reported incorrectly from my script. It's as if the routine that does
the system query is using a 32 bit thing for the size and it's getting
truncated as integers do. Some files actually come back with a negative
size. The data type that ruby is using is perfectly capable of holding
the size of the files. I've tried File.size? and FileUtils.size?, so
I'm somewhat at a loss short of writing an extension. Any ideas?

The script is running on windows, querying the files via UNC paths
through a samba share on linux.

It's a bug in Ruby.

The win32-file library fixes it.

require 'win32/file' # Do this first
File.size(some_big_file)

Regards,

Dan