Debian Bug report logs - #476846
genisoimage: On arm only, does not recognise its own Rock Ridge extensions

version graph

Package: genisoimage; Maintainer for genisoimage is Joerg Jaspert <joerg@debian.org>; Source for genisoimage is src:cdrkit.

Reported by: Barry Tennison <barry@ukph.org>

Date: Sat, 19 Apr 2008 15:06:01 UTC

Severity: important

Found in versions cdrkit/9:1.1.2-1, cdrkit/9:1.1.7.1-1

Reply or subscribe to this bug.

Toggle useless messages

View this report as an mbox folder, status mbox, maintainer mbox


Report forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to Barry Tennison <barry@ukph.org>:
New Bug report received and forwarded. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #5 received at submit@bugs.debian.org (full text, mbox):

From: Barry Tennison <barry@ukph.org>
To: submit@bugs.debian.org
Subject: genisoimage: On arm only, does not recognise its own Rock Ridge extensions
Date: Sat, 19 Apr 2008 16:03:33 +0100
Package: genisoimage
Version: 9:1.1.2-1
Severity: important

The following behaviour is in error on arm only, at least on my NSLU2
running etch.  On i386 and amd64, the error does not occur.

Steps to reproduce on an arm machine:

cd <somewhere safe>
mkdir tmp1
touch tmp1/nullfile-with-long-name
genisoimage -R tmp1/ >RR-arm-1.iso
# note: this succeeds, and then RR-arm-1.iso will loop-mount perfectly
# well with Rock Ridge active, so nullfile-with-long-name shows with
# the right name
genisoimage -M RR-arm-1.iso -C 0,0 -R tmp1/ >RR-arm-2.iso

On my arm machine, the last command produces the following:
----------
I: -input-charset not specified, using utf-8 (detected in locale settings)
genisoimage: **BAD RRVERSION (0) for �
NO Rock Ridge present
Disabling Rock Ridge / XA / AA
----------
(BUT note well that, as above, mount handles the RR extensions of
RR-arm-1.iso perfectly.)

On i386 and on amd64, I get success and the following, as expected:
----------
I: -input-charset not specified, using utf-8 (detected in locale settings)
Rock Ridge signatures found
----------

The effect is exactly the same if -r is used in place of -R.  I doubt
very much whether the -C 0,0 is the cause: it works on i386 and amd64
and I've tried different values.

I've discussed this a bit on the debian-arm list, and have asked whether
others see the same behaviour, in particular on the lenny and sid
version of genisoimage (9:1.1.7.1-1).  I'll report if I get a response.

The best guess (due to Tobias Frost) in the discussion has been that it
could be an endian issue; but if so, it seems to be in the initial read
code of genisoimage.

I've categorised this as important because it renders useless, on arm,
the excellent genisoimage/growisofs ability to grow an isofs (eg for
daily backups).

I'm happy to provide more info if I can - but I'm reluctant to upgrade
my slug to sid just to elucidate this bug.  I may investigate qemu as a
potential way to try that (thanks to Tim Small for that suggestion).

Barry

-- System Information:
Debian Release: 4.0
  APT prefers stable
  APT policy: (990, 'stable'), (500, 'testing')
Architecture: arm (armv5tel)
Shell:  /bin/sh linked to /bin/bash
Kernel: Linux 2.6.18-6-ixp4xx
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)

Versions of packages genisoimage depends on:
ii  libc6                  2.3.6.ds1-13etch5 GNU C Library: Shared libraries
ii  libmagic1              4.17-5etch3       File type determination
library us
ii  zlib1g                 1:1.2.3-13        compression library - runtime

genisoimage recommends no packages.

-- no debconf information









Information forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to Barry Tennison <barry@ukph.org>:
Extra info received and forwarded to list. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #10 received at 476846@bugs.debian.org (full text, mbox):

From: Barry Tennison <barry@ukph.org>
To: 476846@bugs.debian.org
Cc: martinwguy@gmail.com, Tim Small <tim@buttersideup.com>, Tobias Frost <tobi@coldtobi.de>, Bill Gatliff <bgat@billgatliff.com>
Subject: Confirming bug for sid version (9:1.1.7.1-1) of genisoimage (arm only)
Date: Sun, 20 Apr 2008 10:23:00 +0100
Package: genisoimage
Version: 9:1.1.7.1-1
Severity: important

Thanks to Martin Guy, I can confirm this buggy behaviour for genisoimage 
version 1.1.7.1-1 (sid current) on arm (and indeed on a different 
machine from my own NSLU2, whew).

The "steps to reproduce" in the original report again produce the 
unexpected output:
-----------
genisoimage: **BAD RRVERSION (0) for �
NO Rock Ridge present
Disabling Rock Ridge / XA / AA
Warning: Neither Rock Ridge (-R) nor TRANS.TBL (-T)
name translations were found on previous session.
ISO-9660 file names have been used instead.
-----------
and genisoimage fails to produce an updated iso with Rock Ridge signatures.

System details:
uname -a
Linux n2100 2.6.24-1-iop32x #1 Fri Feb 22 13:15:18 UTC 2008 armv5tel 
GNU/Linux
Package versions:
ii  genisoimage              9:1.1.7.1-1
ii  libbz2-1.0               1.0.4-3
ii  libc6                    2.7-8
ii  libmagic1                4.23-2
ii  zlib1g                   1:1.2.3.3.dfsg-11


Barry







Information forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to Barry Tennison <barry@ukph.org>:
Extra info received and forwarded to list. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #15 received at 476846@bugs.debian.org (full text, mbox):

From: Barry Tennison <barry@ukph.org>
To: 476846@bugs.debian.org
Cc: martinwguy@gmail.com, Tim Small <tim@buttersideup.com>, Tobias Frost <tobi@coldtobi.de>, Bill Gatliff <bgat@billgatliff.com>
Subject: Bug does NOT happen for sid version (9:1.1.7.1-1) of genisoimage on armel
Date: Sun, 20 Apr 2008 10:37:29 +0100
Package: genisoimage
Version: 9:1.1.7.1-1
Severity: important

Thanks again to Martin Guy, I can confirm that the buggy behaviour does 
NOT happen for genisoimage version 1.1.7.1-1 (sid current) on armel. 
Instead, the behaviour is just as on i386 and amd64: the RR signatures 
are detected and the iso successfully updated.

Maybe this is support for the explanation as an endian problem?

System details:
uname -a
Linux n2100 2.6.24-1-iop32x #1 Fri Feb 22 13:15:18 UTC 2008 armv5tel 
GNU/Linux
Package versions:
ii  genisoimage              9:1.1.7.1-1
ii  libbz2-1.0               1.0.5-0.1
ii  libc6                    2.7-10
ii  libmagic1                4.23-2
ii  zlib1g                   1:1.2.3.3.dfsg-11

Barry








Information forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to Peter Samuelson <peter@p12n.org>:
Extra info received and forwarded to list. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #20 received at 476846@bugs.debian.org (full text, mbox):

From: Peter Samuelson <peter@p12n.org>
To: Barry Tennison <barry@ukph.org>, 476846@bugs.debian.org
Cc: martinwguy@gmail.com, Tim Small <tim@buttersideup.com>, Tobias Frost <tobi@coldtobi.de>, Bill Gatliff <bgat@billgatliff.com>
Subject: Re: Bug#476846: Bug does NOT happen for sid version (9:1.1.7.1-1) of genisoimage on armel
Date: Sun, 20 Apr 2008 10:24:41 -0500
[Message part 1 (text/plain, inline)]
[Barry Tennison]
> Thanks again to Martin Guy, I can confirm that the buggy behaviour does NOT 
> happen for genisoimage version 1.1.7.1-1 (sid current) on armel. Instead, 
> the behaviour is just as on i386 and amd64: the RR signatures are detected 
> and the iso successfully updated.
>
> Maybe this is support for the explanation as an endian problem?

No, it's not an endian problem.  i386, amd64, arm, and armel are _all_
little-endian.

My theory is a bad build due to a buggy gcc-4.2.  It is only a theory.
-- 
Peter Samuelson | org-tld!p12n!peter | http://p12n.org/
[signature.asc (application/pgp-signature, inline)]

Information forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to Tobias Frost <tobi@coldtobi.de>:
Extra info received and forwarded to list. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #25 received at 476846@bugs.debian.org (full text, mbox):

From: Tobias Frost <tobi@coldtobi.de>
To: Peter Samuelson <peter@p12n.org>
Cc: Barry Tennison <barry@ukph.org>, 476846@bugs.debian.org, martinwguy@gmail.com, Tim Small <tim@buttersideup.com>, Bill Gatliff <bgat@billgatliff.com>
Subject: Re: Bug#476846: Bug does NOT happen for sid version (9:1.1.7.1-1) of genisoimage on armel
Date: Sun, 20 Apr 2008 20:23:02 +0200

On Sun, 2008-04-20 at 10:24 -0500, Peter Samuelson wrote:

> No, it's not an endian problem.  i386, amd64, arm, and armel are _all_
> little-endian.

MMmmh, I always thought arm and armel differs in endianess...




Information forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to Tim Small <tim@buttersideup.com>:
Extra info received and forwarded to list. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #30 received at 476846@bugs.debian.org (full text, mbox):

From: Tim Small <tim@buttersideup.com>
To: Tobias Frost <tobi@coldtobi.de>
Cc: Peter Samuelson <peter@p12n.org>, Barry Tennison <barry@ukph.org>, 476846@bugs.debian.org, martinwguy@gmail.com, Bill Gatliff <bgat@billgatliff.com>
Subject: Re: Bug#476846: Bug does NOT happen for sid version (9:1.1.7.1-1) of genisoimage on armel
Date: Sun, 20 Apr 2008 19:33:37 +0100
Tobias Frost wrote:
> On Sun, 2008-04-20 at 10:24 -0500, Peter Samuelson wrote:
>
>   
>> No, it's not an endian problem.  i386, amd64, arm, and armel are _all_
>> little-endian.
>>     
>
> MMmmh, I always thought arm and armel differs in endianess...
>   

Nope, some packing, and syscall, and FP differences.  If you have a 
sid/armel install, or an etch/arm install you can install a sid/arm 
chroot under it (using debootstrap), and test that if you suspect armel 
vs arm....

Ta,

Tim.




Information forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to Rick Thomas <rbthomas55@pobox.com>:
Extra info received and forwarded to list. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #35 received at 476846@bugs.debian.org (full text, mbox):

From: Rick Thomas <rbthomas55@pobox.com>
To: 476846@bugs.debian.org
Cc: Arm List Debian <debian-arm@lists.debian.org>, Barry Tennison <barry@ukph.org>
Subject: Re: Puzzling difference between debian-arm and debian-i386 re growisofs
Date: Mon, 21 Apr 2008 00:13:05 -0400
On Apr 18, 2008, at 1:14 PM, Barry Tennison wrote:

>
> So I think it's definitely in the genisoimage -M code, and COULD BE  
> an endian issue there.
>

For what it proves, if anything, I tried to recreate the bug

> cd <somewhere safe>
> mkdir tmp1
> touch tmp1/nullfile-with-long-name
> genisoimage -R tmp1/ >RR-arm-1.iso
> # note: this succeeds, and then RR-arm-1.iso will loop-mount perfectly
> #  well with Rock Ridge active, so nullfile-with-long-name shows with
> #  the right name
> genisoimage -M RR-arm-1.iso -C 0,0 -R tmp1/ >RR-arm-2.iso


on a PowerMac G3 running Etch, and had no errors at all.  So simple  
endian-ness bugs seem less likely.  Complex bugs (endian or  
otherwise) are still possible, of course.

Enjoy!

Rick




Information forwarded to debian-bugs-dist@lists.debian.org, Joerg Jaspert <joerg@debian.org>:
Bug#476846; Package genisoimage. Full text and rfc822 format available.

Acknowledgement sent to "Martin Guy" <martinwguy@yahoo.it>:
Extra info received and forwarded to list. Copy sent to Joerg Jaspert <joerg@debian.org>. Full text and rfc822 format available.

Message #40 received at 476846@bugs.debian.org (full text, mbox):

From: "Martin Guy" <martinwguy@yahoo.it>
To: "Barry Tennison" <barry@ukph.org>
Cc: 476846@bugs.debian.org
Subject: Re: Bug does NOT happen for sid version (9:1.1.7.1-1) of genisoimage on armel
Date: Mon, 21 Apr 2008 19:09:52 +0100
>  Maybe this is support for the explanation as an endian problem?

arm and armel are little-endian. The main weirdnesses of plain arm are
a bizarre middle-endian floating point format (look for things trying
to read FP values off a disk file or network stream) and unusual
structure-packing behaviour (look for filling a struct with values and
dumping it whole to a file or stream or vice versa).

Good luck!

    M




Send a report that this bug log contains spam.


Debian bug tracking system administrator <owner@bugs.debian.org>. Last modified: Thu Apr 17 00:03:16 2014; Machine Name: buxtehude.debian.org

Debian Bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.