Re: finding a dependency chain

2014-09-05 Thread davidson 

On Thu, 4 Sep 2014, david...@ling.ohio-state.edu wrote:


On Wed, 3 Sep 2014, Rob Owens wrote:


- Original Message -

From: Kelly Clowers kelly.clow...@gmail.com

On Tue, Sep 2, 2014 at 1:44 PM, Rob Owens row...@ptd.net wrote:

I'm trying to figure out, for example, what causes brasero to
ultimately depend on systemd.  I found a utility called debtree,
but it produces too much output to be of use to me -- it shows all
dependency chains starting at brasero, but I am only interested in
the one that ends at systemd.

Can anybody suggest another utility, or maybe the proper syntax to
make debtree do what I want?

Thanks

-Rob


I just used vim to search through debtree's .dot output (for such a
complex thing, way easier than trying to look at the image file),
and then followed up by looking around in aptitude interactive
mode.

Thanks.  That is much easier than looking at the image file!

I'd still like to find a method to specify the start point and end
point, and get output of a single dependency chain.  If anybody
knows a way, please post it.


seems like this pipeline

$ apt-cache --recurse -i depends brasero |tac

yields the elements in question, and in a convenient order (children
before parents), but with a whole bunch of irrelevance interleaved.

piping it to something like the following clears away the irrelevant
parts:


note the careful choice of words above: *something like*



$ cat above_kludge
#!/bin/bash

pattrn=$1

while read line
do if [[ $line =~ ^${pattrn}$ ]]  # Found node C.
then echo $line   # Admit node C to output.
pattrn=$line  # Ground the pattern.
else if [[ $line =~ [[:blank:]]${pattrn}$ ]]  # Found node C listed
  # above its parent.
then pattrn='[[:alnum:][:punct:]]*'   # Seek the parent.
fi
fi
done


did some more tests this morning.  ouch.  above kludge has problems.

version below works better:

  #!/bin/bash

  seek_parent=1
  pattrn=$1

  while IFS='' read line
  do
  if [[ $seek_parent -eq 1 ]]  [[ $line =~ ^${pattrn}$ ]]
  then
  echo $line
  pattrn=$line
  seek_parent=0
  elif [[ $seek_parent -eq 0 ]]  [[ $line =~ [[:space:]]${pattrn}$ ]]
  then
  pattrn=[[:alnum:][:punct:]]+
  seek_parent=1
  fi
  done


so, like this:

$ apt-cache --recurse -i depends brasero |tac |above_kludge '.*systemd.*'


er, no.  that won't work in general.  for better results, replace '.'
on the line above with '[[:alnum:][:punct:]]'.

so, more like this:

  $ apt-cache --recurse -i depends brasero |tac |\
above_kludge '[[:alnum:][:punct:]]*systemd[[:alnum:][:punct:]]*'

-wes


libsystemd-login0
dbus
udisks
libgdu0
gvfs-daemons
gvfs
brasero

while in general there might be multiple paths, above kludge will find
just one of them.

-wes






--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Archive: 
https://lists.debian.org/alpine.deb.2.02.1409041945030.31...@brutus.ling.ohio-state.edu

Re: finding a dependency chain

2014-09-05 Thread Cindy-Sue Causey 
On 9/3/14, Rob Owens row...@ptd.net wrote:
 - Original Message -
 From: Kelly Clowers kelly.clow...@gmail.com

 I just used vim to search through debtree's .dot output (for such a
 complex thing, way easier than trying to look at the image file), and
 then followed up by looking around in aptitude interactive mode.

 Thanks.  That is much easier than looking at the image file!

 I'd still like to find a method to specify the start point and end point,
 and get output of a single dependency chain.  If anybody knows a way, please
 post it.


Found something the other day while scouring the Net to fix some or
another issue here. Am sure this is probably not doing quite what
you're seeking. Hoping it's still maybe a start for training some
program to do that since it's already (unverified so allegedly)
reporting ALL dependencies..

If it's a far lesser [function] than what you've already been chatting
about, I'd love to hear the difference. Really trying to learn this
stuff inside out.. Hoping to contribute sooner than later. :)

The command is:

apt-cache showpkg packageName

It outputs this (apologies, is a little extensive but decided to leave intact):

++

elf@northpole:~$ apt-cache showpkg wvdial
Package: wvdial
Versions:
1.61-4.1 
(/var/lib/apt/lists/mirrors.kernel.org_debian_dists_wheezy_main_binary-amd64_Packages)
 Description Language:
 File:
/var/lib/apt/lists/mirrors.kernel.org_debian_dists_wheezy_main_binary-amd64_Packages
  MD5: b8bf30c8dfd4d09e02af74bf497505d6
 Description Language: en
 File:
/var/lib/apt/lists/mirrors.kernel.org_debian_dists_wheezy_main_i18n_Translation-en
  MD5: b8bf30c8dfd4d09e02af74bf497505d6

1.61-4build1 
(/var/lib/apt/lists/us.archive.ubuntu.com_ubuntu_dists_precise_main_binary-amd64_Packages)
(/var/lib/dpkg/status)
 Description Language:
 File:
/var/lib/apt/lists/mirrors.kernel.org_debian_dists_wheezy_main_binary-amd64_Packages
  MD5: b8bf30c8dfd4d09e02af74bf497505d6
 Description Language: en
 File:
/var/lib/apt/lists/mirrors.kernel.org_debian_dists_wheezy_main_i18n_Translation-en
  MD5: b8bf30c8dfd4d09e02af74bf497505d6

Reverse Depends:
  wvdial:i386,wvdial
  wader-core,wvdial
  lubuntu-desktop,wvdial
  ichthux-desktop,wvdial
  gnome-ppp,wvdial
  usb-modeswitch,wvdial
  wvdial:i386,wvdial
  wader-core,wvdial
  usb-modeswitch,wvdial
  gnome-ppp,wvdial

Dependencies:
1.61-4.1 - ppp (2 2.3.0) debconf (18 0.5.00) cdebconf (0 (null)) libc6
(2 2.3.4) libgcc1 (2 1:4.1.1) libstdc++6 (2 4.1.1) libuniconf4.6 (0
(null)) libwvstreams4.6-base (0 (null)) libwvstreams4.6-extras (0
(null)) debconf (18 0.5) debconf-2.0 (0 (null)) wvdial:i386 (0 (null))
1.61-4build1 - ppp (2 2.3.0) debconf (18 0.5.00) cdebconf (0 (null))
libc6 (2 2.4) libstdc++6 (2 4.1.1) libuniconf4.6 (0 (null))
libwvstreams4.6-base (0 (null)) libwvstreams4.6-extras (0 (null))
debconf (18 0.5) debconf-2.0 (0 (null)) wvdial:i386 (0 (null))
Provides:
1.61-4.1 -
1.61-4build1 -
Reverse Provides:

++

Hoping that's on topic.. In particular, having that as a latest
personal find just showed me how much average users are probably not
tapping anywhere near what's available to us in Debian and its
derivatives..

Cindy :)

-- 
Cindy-Sue Causey
Talking Rock, Pickens County, Georgia, USA

* runs with duct tape *


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/CAO1P-kC_QfQMiVu4xFRM4-GHt9-yqHmk7+Sm7_e-UUor3bh2=g...@mail.gmail.com

Re: finding a dependency chain

2014-09-04 Thread davidson 

On Wed, 3 Sep 2014, Rob Owens wrote:


- Original Message -

From: Kelly Clowers kelly.clow...@gmail.com

On Tue, Sep 2, 2014 at 1:44 PM, Rob Owens row...@ptd.net wrote:

I'm trying to figure out, for example, what causes brasero to
ultimately depend on systemd.  I found a utility called debtree,
but it produces too much output to be of use to me -- it shows all
dependency chains starting at brasero, but I am only interested in
the one that ends at systemd.

Can anybody suggest another utility, or maybe the proper syntax to
make debtree do what I want?

Thanks

-Rob


I just used vim to search through debtree's .dot output (for such a
complex thing, way easier than trying to look at the image file),
and then followed up by looking around in aptitude interactive
mode.

Thanks.  That is much easier than looking at the image file!

I'd still like to find a method to specify the start point and end
point, and get output of a single dependency chain.  If anybody
knows a way, please post it.


seems like this pipeline

 $ apt-cache --recurse -i depends brasero |tac

yields the elements in question, and in a convenient order (children
before parents), but with a whole bunch of irrelevance interleaved.

piping it to something like the following clears away the irrelevant
parts:

 $ cat above_kludge
 #!/bin/bash

 pattrn=$1

 while read line
 do if [[ $line =~ ^${pattrn}$ ]]  # Found node C.
 then echo $line   # Admit node C to output.
 pattrn=$line  # Ground the pattern.
 else if [[ $line =~ [[:blank:]]${pattrn}$ ]]  # Found node C listed
   # above its parent.
 then pattrn='[[:alnum:][:punct:]]*'   # Seek the parent.
 fi
 fi
 done


so, like this:

 $ apt-cache --recurse -i depends brasero |tac |above_kludge '.*systemd.*'
 libsystemd-login0
 dbus
 udisks
 libgdu0
 gvfs-daemons
 gvfs
 brasero

while in general there might be multiple paths, above kludge will find
just one of them.

-wes


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Archive: 
https://lists.debian.org/alpine.deb.2.02.1409040835510.19...@brutus.ling.ohio-state.edu

Re: finding a dependency chain

2014-09-03 Thread Rob Owens 
- Original Message -
 From: Kelly Clowers kelly.clow...@gmail.com
 
 On Tue, Sep 2, 2014 at 1:44 PM, Rob Owens row...@ptd.net wrote:
  I'm trying to figure out, for example, what causes brasero to ultimately
  depend on systemd.  I found a utility called debtree, but it produces too
  much output to be of use to me -- it shows all dependency chains starting
  at brasero, but I am only interested in the one that ends at systemd.
 
  Can anybody suggest another utility, or maybe the proper syntax to make
  debtree do what I want?
 
  Thanks
 
  -Rob
 
 I just used vim to search through debtree's .dot output (for such a
 complex thing, way easier than trying to look at the image file), and
 then followed up by looking around in aptitude interactive mode.
 
Thanks.  That is much easier than looking at the image file!

I'd still like to find a method to specify the start point and end point, and 
get output of a single dependency chain.  If anybody knows a way, please post 
it.

-Rob


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1673216099.1327226.1409754647365.javamail.zim...@ptd.net

finding a dependency chain

2014-09-02 Thread Rob Owens 
I'm trying to figure out, for example, what causes brasero to ultimately depend 
on systemd.  I found a utility called debtree, but it produces too much output 
to be of use to me -- it shows all dependency chains starting at brasero, but I 
am only interested in the one that ends at systemd.

Can anybody suggest another utility, or maybe the proper syntax to make debtree 
do what I want?

Thanks

-Rob


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/1295244982.1164148.1409690663860.javamail.zim...@ptd.net

Re: finding a dependency chain

2014-09-02 Thread Kelly Clowers 
On Tue, Sep 2, 2014 at 1:44 PM, Rob Owens row...@ptd.net wrote:
 I'm trying to figure out, for example, what causes brasero to ultimately 
 depend on systemd.  I found a utility called debtree, but it produces too 
 much output to be of use to me -- it shows all dependency chains starting at 
 brasero, but I am only interested in the one that ends at systemd.

 Can anybody suggest another utility, or maybe the proper syntax to make 
 debtree do what I want?

 Thanks

 -Rob

I just used vim to search through debtree's .dot output (for such a
complex thing, way easier than trying to look at the image file), and
then followed up by looking around in aptitude interactive mode.


Looks like Brasero pulls in gvfs, which depends on gvfs-daemons, which
depends on libsystemd-login0 and udisks2. Udisks2 depends on a couple
systemd libs directly (but not the systemd package), and
libpam-systemd. libpam-systemd depends on systemd.

There is also a chain where gvfs-daemons recommends policykit-1-gnome,
which depends on policykit-1, which depends on libpam-systemd, which
as above depends on systemd

This is in Sid, though, your mileage may vary.

Huh, and from this I just found that in Experimental there is a new
version of policykit-1 that depends on systemd directly and also...
libmozjs??? That sounds odd...

Cheers,
Kelly Clowers


--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/CAFoWM=9ppp72cwHMb4eabPc6qWsETLb3Af=tiau2bjw347g...@mail.gmail.com

Re: finding a dependency chain

2014-09-02 Thread Brian 
On Tue 02 Sep 2014 at 14:21:06 -0700, Kelly Clowers wrote:

 Huh, and from this I just found that in Experimental there is a new
 version of policykit-1 that depends on systemd directly and also...
 libmozjs??? That sounds odd...

The libmozjs dependency is probably because .rules files have replaced
.pkla files.


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/02092014233145.0e843c3d2...@desktop.copernicus.demon.co.uk

Re: finding a dependency chain

2014-09-02 Thread Kelly Clowers 
On Tue, Sep 2, 2014 at 3:34 PM, Brian a...@cityscape.co.uk wrote:
 On Tue 02 Sep 2014 at 14:21:06 -0700, Kelly Clowers wrote:

 Huh, and from this I just found that in Experimental there is a new
 version of policykit-1 that depends on systemd directly and also...
 libmozjs??? That sounds odd...

 The libmozjs dependency is probably because .rules files have replaced
 .pkla files.

Ah, right you are, thanks!
Found this: 
http://davidz25.blogspot.com/2012/06/authorization-rules-in-polkit.html
Eh, I guess JS works... Lua might have been good too.


-- 
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org 
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/CAFoWM=_dqvgnmkprkyz4day-j8akjwqacu9trcswcqtq43g...@mail.gmail.com