bug-coreutils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: md5sum coreutils 5.2.1-2 buffer_read(fd): Segmentation fault


From: Bob Proulx
Subject: Re: md5sum coreutils 5.2.1-2 buffer_read(fd): Segmentation fault
Date: Fri, 5 Aug 2005 22:39:12 -0600
User-agent: Mutt/1.5.9i

J. Grant wrote:
> I am using Debian GNU/Linux  coreutils  5.2.1-2
> 
> I burned a DVD-R and when I check the md5s.txt I get this crash.  I
> presume the DVD-R did not burn correctly, but md5sum should not seg fault.
> 
> $ md5sum -c md5s.txt
> error processing Dog.Day.Afternoon.(1975).CD1.avi: failed in
> buffer_read(fd): mdfile: Input/output error
> Segmentation fault
> address@hidden dvd]$

That does look like a problem.  Thanks for reporting it.  The dvd was
mounted as a filesystem at this time?  Looks like it.

> I tried to run it through gdb and catch some call stack or something, but it
> got stuck and never completed:

Oh very good.  Thanks for trying that.  Another thing to try is to run
strace on the binary.

  strace -o /tmp/strace.out md5sum -c md5s.txt

The trace output file will show the system calls and the return values
of the system calls.

> (gdb) exec /usr/bin/md5sum -v -c md5s.txt

But /usr/bin/md5sum is not the coreutils version on Debian.  That one
is the dpkg version.  The coreutils version was renamed to
md5sum.textutils when it was in the textutils package to avoid
replacing the one in the dpkg package.

> Also, could -v and --version options be added to md5sum ?  It took me a
> while to find md5sum was part of textutils, and then that had been
> merged into this coreutils.

If you are not seeing a --version then for sure you have the other
md5sum and not the one from coreutils.  The coreutils md5sum supports
--version fine.

  md5sum.textutils --version
  md5sum (coreutils) 5.2.1

Try the coreutils one and see if it works for you or not.

Future versions of md5sum in Debian will very likely be the coreutils
one and not the dpkg one.  However there is some friction between the
package maintainers about how this transition should occur.
Bug#313258 in the Debian BTS has some interesting discussion on this.

Bob




reply via email to

[Prev in Thread] Current Thread [Next in Thread]