Re: Bash scripts

2011-04-24 Thread Paul Fox
kevin wrote:
  Sascha:
  
  
  Continued to play with this.  Went into nautilus on an ext2 formatted
  card, and there is a checkbox in the panel for allowing execution of
  files.  The owner and other permission boxes didn't seem to do
  anything; but, clicking that on *did* work. Also, you were correct
  about it needing to not be FAT, checking that box, even though it sill
  displays, didn't 'stick' for FAT. Thanks.

i've lost track of where exactly this issue sits, but if it's
a regression, i suggest opening a ticket.  then it can be resolved
in a documented fashion (whether fixed as bug or wontfixed as
a feature).

paul

  
  Cheers,
  
  
  KG
  
  On Monday, April 18,
  2011, Kevin Gordon kgordon...@gmail.com wrote:
   Sascha:
  
   The file system actually had no bearing on the issue I was having, whether 
  ext2,
   ext3, or FAT32, the symptoms were identical - recent versions of udisks 
   now 
  does not allow
   'direct' execution of scripts from auto-mounted removable media.
  
   Also,
   there is some debate as to whether putting a journalling fs onto an SD
   or USB drive is wise, as it might half its life by in essence doubling the 
  number of writes.  In
   general,  I tend to stick with the factory default unless I need
   multiple partitions, symbolic link, or specific linux-swap support, since 
   I 
  presume the
   manufacturer has formatted it with the right number of blocks, units,
   etc to best match their controller/memory config.  If I need those, I will 
  still use ext2.  Call me optimistic :-)
  
   Cheers,
  
   KG
  
   On Mon, Apr 18, 2011 at 4:13 AM, Sascha Silbe 
  sascha-ml-reply-to-201...@silbe.org wrote:
  
   Excerpts from Kevin Gordon's message of Mon Apr 18 00:36:26 +0200 2011:
  
   But, since my main use of this technique is to
   semi-automate the process of installing a slew of custom activities and
   rpm's upon initial build and deployment, having to manually change every
   machine manually to basically avoid 5 keystrokes, was sort of
   counter-productive  :-)
  
   If you're only using this USB stick with Linux machines, why don't you
   just format it using a file system with POSIX semantics, i.e. ext3?
  
   Sascha
  
   --
   http://sascha.silbe.org/
   http://www.infra-silbe.de/
  
  
  ___
  Devel mailing list
  Devel@lists.laptop.org
  http://lists.laptop.org/listinfo/devel

=-
 paul fox, p...@laptop.org
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-18 Thread Sascha Silbe
Excerpts from Kevin Gordon's message of Sun Apr 17 17:19:36 +0200 2011:

 any build before.  Log in as root (su -)   Write a little script - it can be
 a one liner, echo Hello World  is good enough.  Save it as test.sh to a
 USB stick which has been formatted FAT32 and has a volume name FEDORA. Do a
 Chmod +x /media/FEDORA/test.sh.

Let me provide another piece of the puzzle: FAT is the file system that
was designed for DOS. Unlike the older Unix systems DOS didn't have the
concept of permissions, so neither has FAT. Assignment of ownership and
permissions for FAT file systems is always based on heuristics,
controlled by mount options. For your case the options exec/noexec (file
system independent), umask, fmask and showexec (FAT only) are relevant.

If you want real POSIX permissions support on a FAT file system, you
need to use umsdos [1]. Make sure you understand the principles of
operation and the associated drawbacks first, esp. if you use the same
USB stick with other systems that do not run umsdos. Nobody I know ever
went to the trouble of using it, not even when most of our systems were
still running OS/2.

Sascha

[1] http://tldp.org/HOWTO/UMSDOS-HOWTO.html
-- 
http://sascha.silbe.org/
http://www.infra-silbe.de/


signature.asc
Description: PGP signature
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-18 Thread Sascha Silbe
Excerpts from Kevin Gordon's message of Mon Apr 18 00:36:26 +0200 2011:

 But, since my main use of this technique is to
 semi-automate the process of installing a slew of custom activities and
 rpm's upon initial build and deployment, having to manually change every
 machine manually to basically avoid 5 keystrokes, was sort of
 counter-productive  :-)

If you're only using this USB stick with Linux machines, why don't you
just format it using a file system with POSIX semantics, i.e. ext3?

Sascha

-- 
http://sascha.silbe.org/
http://www.infra-silbe.de/


signature.asc
Description: PGP signature
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-18 Thread James Cameron
On Mon, Apr 18, 2011 at 10:13:56AM +0200, Sascha Silbe wrote:
 If you're only using this USB stick with Linux machines, why don't you
 just format it using a file system with POSIX semantics, i.e. ext3?

I don't know about Kevin, but for me some USB flash drives become slower
at writing of large files, when converted from factory formatted FAT to
ext3.  It is as if the flash translation layer in the USB drive was
optimised for FAT.  Adjusting the starting position of the partition can
do this as well.  ext3 also has this journal thing, which presumably
increases the number of writes.

-- 
James Cameron
http://quozl.linux.org.au/
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-18 Thread Kevin Gordon
Sascha:

The file system actually had no bearing on the issue I was having, whether
ext2, ext3, or FAT32, the symptoms were identical - recent versions of
udisks now does not allow 'direct' execution of scripts from auto-mounted
removable media.

Also, there is some debate as to whether putting a journalling fs onto an SD
or USB drive is wise, as it might half its life by in essence doubling the
number of writes.  In general,  I tend to stick with the factory default
unless I need multiple partitions, symbolic link, or specific linux-swap
support, since I presume the manufacturer has formatted it with the right
number of blocks, units, etc to best match their controller/memory config.
If I need those, I will still use ext2.  Call me optimistic :-)

Cheers,

KG

On Mon, Apr 18, 2011 at 4:13 AM, Sascha Silbe 
sascha-ml-reply-to-201...@silbe.org wrote:

 Excerpts from Kevin Gordon's message of Mon Apr 18 00:36:26 +0200 2011:

  But, since my main use of this technique is to
  semi-automate the process of installing a slew of custom activities and
  rpm's upon initial build and deployment, having to manually change every
  machine manually to basically avoid 5 keystrokes, was sort of
  counter-productive  :-)

 If you're only using this USB stick with Linux machines, why don't you
 just format it using a file system with POSIX semantics, i.e. ext3?

 Sascha

 --
 http://sascha.silbe.org/
 http://www.infra-silbe.de/

___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-17 Thread Ismael Luceno
El Sun, 17 Apr 2011 11:19:36 -0400
Kevin Gordon kgordon...@gmail.com escribió:
 Perhaps one could conjecture that this is some added security, or new
 property for execution of scripts resident on portable drives that
 has been implemented.  However, as for 'filing a bug report', it
 looks like it's something way upstream from my pay-class, since it
 isn't limited to OLPC, but occurs across Fedora 14 updated machines
 too.

I've not checked, but probably the filesystem is being mounted with
the noexec option.

Anyway, it's a feature, not a bug.


signature.asc
Description: PGP signature
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-17 Thread Kevin Gordon
Jerry Vonau For the WIN!

A yum downgrade udisks reinstalls the 1.0.1-4 package and makes it behave
just like before.  Nice little green file-names on teh ls output, the
ability to run scripts from USB drives without the bash command returns, and
tab complete functions again..

Again, since this package is used all through fedora, (and on Ubuntu too, I
think), I doubt little old me should be telling the upstream world how it
should work.

Cheers,

KG





On Sun, Apr 17, 2011 at 12:54 PM, Jerry Vonau jvo...@shaw.ca wrote:

 On Sun, 2011-04-17 at 11:19 -0400, Kevin Gordon wrote:
  Martin and the Gang:
 
  Since last we spoke, I've been working on this new issue I'd
  mentioned, where when I try to run a bash shell script resident on a
  USB drive inserted into an OS16 OLPC, I  always get Permission
  Denied.  But now, I'm not sure it's a bug; and if so, I'm now
  pretty sure it isn't an OLPC bug.
 
  Here are the details to reproduce.  Go to terminal on a 10.1.3 Build
  860, or any build before.  Log in as root (su -)   Write a little
  script - it can be a one liner, echo Hello World  is good enough.
  Save it as test.sh to a USB stick which has been formatted FAT32 and
  has a volume name FEDORA. Do a Chmod +x /media/FEDORA/test.sh.   One
  can then take that USB stick to any previous-to-Fedora 14 OLPC, and
  execute the script by simply installing the drive, logging into root
  (su -), going to the terminal prompt, and entering
  /media/FEDORA/test.sh, and see the old 'Hello World' come up - not
  Permission Denied.  One can also tab complete through that command,
  and it sees and auto-completes the test.sh portion too,
 
  So, I tried on a vanilla little lenovo with a fresh unaltered
  installation of the Fedora14 Live-Desktop.  The script runs fine from
  the USB drive.  Then after accepting and installing the 422 recommened
  Software Updates on the box, this box too now gives Permission
  Denied. I repeated this same procedure twice more with fresh
  installs, then updates.  Very consistent,, works before updates, does
  not work after.
 
  On the OLPC side, I did a more binary search, and the script continues
  to work right up to and including os5 from the 11.2 builds. At os6,
  and after, is where the Permission Denied starts to occur when
  attempting to execute.  Not surprisingly, and concurrently, the tab
  complete capability on the test.sh portion ceases to function at
  this point as well.
 
  Alas, someone upstream might now consider this a feature, not a bug.
  Also, someone upstream might be able to figure out which of those 422
  packages is responsible, if they should care to.

 Partial diff between os5.packages.txt os6.packages.txt

 -udisks-1.0.1-4.fc14.i686
 +udisks-1.0.2-2.fc14.i686


 Jerry


___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-17 Thread Martin Langhoff
On Sun, Apr 17, 2011 at 4:28 PM, Kevin Gordon kgordon...@gmail.com wrote:
 Again, since this package is used all through fedora, (and on Ubuntu too, I
 think), I doubt little old me should be telling the upstream world how it
 should work.

Bingo. It's a security feature -- avoids the situations where you plug
in a usb drive and accidentally double-click on something and kaboom.

Your workaround #2 (bash /media/foo/myscript.sh) is what I'd do.

You can also mount -o remount,exec /media/foo

cheers,


m
-- 
 martin.langh...@gmail.com
 mar...@laptop.org -- Software Architect - OLPC
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Bash scripts

2011-04-17 Thread Kevin Gordon
Thanks all, I'll reinstall back up to the OS16 version of udisks and enter
'bash ' before the script name.  Just FYI, there were some other novel
suggestions involving policy kit oerrides, fstab entries, and even some
other funky flag settings.  But, since my main use of this technique is to
semi-automate the process of installing a slew of custom activities and
rpm's upon initial build and deployment, having to manually change every
machine manually to basically avoid 5 keystrokes, was sort of
counter-productive  :-)

Cheers and thanks again.

KG.

On Sun, Apr 17, 2011 at 6:05 PM, Martin Langhoff
martin.langh...@gmail.comwrote:

 On Sun, Apr 17, 2011 at 4:28 PM, Kevin Gordon kgordon...@gmail.com
 wrote:
  Again, since this package is used all through fedora, (and on Ubuntu too,
 I
  think), I doubt little old me should be telling the upstream world how it
  should work.

 Bingo. It's a security feature -- avoids the situations where you plug
 in a usb drive and accidentally double-click on something and kaboom.

 Your workaround #2 (bash /media/foo/myscript.sh) is what I'd do.

 You can also mount -o remount,exec /media/foo

 cheers,


 m
 --
  martin.langh...@gmail.com
  mar...@laptop.org -- Software Architect - OLPC
  - ask interesting questions
  - don't get distracted with shiny stuff  - working code first
  - http://wiki.laptop.org/go/User:Martinlanghoff

___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel