Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-23 Thread Christian Perrier
Quoting Khalid Aziz ([EMAIL PROTECTED]):

 Since one of your proposals was to change from type Choice to
 Boolean for the debconf variable, I am trying to figure out how to
 handle upgrade from previous version. I might choose to simply shelve
 the type change issue for now and update the package with new template
 and new version first since there are other serious issues in the
 current version of kexec-tools that need to be fixed.


That seems fair to me



signature.asc
Description: Digital signature


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-22 Thread Khalid Aziz

On Sat, 2008-04-19 at 20:01 +0200, Christian Perrier wrote:
 Quoting Christian Perrier ([EMAIL PROTECTED]):
 
  Your proposal seems fair by me as long as it is turned into:
  
  _Description: Should kexec-tools handle reboots? 
   If you choose this option, a system reboot will trigger a restart
   into a kernel loaded by kexec instead of going through the full
   system bootloader process.
  
  
  (note the if you choose instead of if you accept)
 
 
 Has there been any progress on that issue ?
 

Since one of your proposals was to change from type Choice to
Boolean for the debconf variable, I am trying to figure out how to
handle upgrade from previous version. I might choose to simply shelve
the type change issue for now and update the package with new template
and new version first since there are other serious issues in the
current version of kexec-tools that need to be fixed.

-- 
===
Khalid Aziz
[EMAIL PROTECTED]




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-19 Thread Christian Perrier
Quoting Christian Perrier ([EMAIL PROTECTED]):

 Your proposal seems fair by me as long as it is turned into:
 
 _Description: Should kexec-tools handle reboots? 
  If you choose this option, a system reboot will trigger a restart
  into a kernel loaded by kexec instead of going through the full
  system bootloader process.
 
 
 (note the if you choose instead of if you accept)


Has there been any progress on that issue ?



signature.asc
Description: Digital signature


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-15 Thread Christian Perrier
Quoting Khalid Aziz ([EMAIL PROTECTED]):

 I like this wording except for specifying shutdown -r. I would suspect
 a number of users simply issue a reboot command or use reboot option
 from gdm menu, and shutdown -r may not be as familiar to all users as
 reboot might be. Plus I do not want to try to enumerate all possible
 commands that could result in kexec reboot. How about this instead:
 

Your proposal seems fair by me as long as it is turned into:

_Description: Should kexec-tools handle reboots? 
 If you choose this option, a system reboot will trigger a restart
 into a kernel loaded by kexec instead of going through the full
 system bootloader process.


(note the if you choose instead of if you accept)




signature.asc
Description: Digital signature


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-14 Thread Khalid Aziz

On Sat, 2008-04-05 at 10:08 +0100, Justin B Rye wrote:
 Khalid Aziz wrote:
  Please choose whether to enable kexec-tools to load a new kernel
  automatically on reboots and kexec into the new kernel instead of
  rebooting into firmware.
  
  Any thoughts?
 
 Mainly, again, what's new about them? - kexec _can_ be used to
 load a specified kernel, but that's not the behaviour you get from
 automatic reloads handled by kexec-tools.  What it does by default
 is load /vmlinuz, and that's usually an _old_ kernel.
 
 Also, I'm not keen on firmware (too technical) or reboots (when
 you mean restarts that avoid the bootstrapping process).  And this
 is too much work for one sentence.
 
 How about:
 
   Should kexec-tools handle reboots? If you accept this option,
   invocations of shutdown -r will trigger a restart into a kernel
   loaded by kexec instead of going through the full system bootloader
   process.
 

I like this wording except for specifying shutdown -r. I would suspect
a number of users simply issue a reboot command or use reboot option
from gdm menu, and shutdown -r may not be as familiar to all users as
reboot might be. Plus I do not want to try to enumerate all possible
commands that could result in kexec reboot. How about this instead:

  Should kexec-tools handle reboots? If you accept this option,
  a system reboot will trigger a restart into a kernel loaded by 
  kexec instead of going through the full system bootloader
  process.

-- 
===
Khalid Aziz
[EMAIL PROTECTED]




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-14 Thread Khalid Aziz

On Mon, 2008-04-14 at 10:48 +0200, Christian Perrier wrote:
 Quoting Christian Perrier ([EMAIL PROTECTED]):
  Quoting Justin B Rye ([EMAIL PROTECTED]):
  
  
   How about:
   
 Should kexec-tools handle reboots? If you accept this option,
 invocations of shutdown -r will trigger a restart into a kernel
 loaded by kexec instead of going through the full system bootloader
 process.
  
  
  Which would of course turn to:
  
  _Description: Should kexec-tools handle reboots? 
   If you accept this option, invocations of 'shutdown -r' will trigger
   a restart into a kernel loaded by kexec instead of going through the
   full system bootloader process.
 
 
 Khalid, what's the status of this suggestion/rewrite of the debconf
 template ?
 
 Once you adopt the new template, you should indeed add to your package
 tree the pending debconf l10n translations (#469960, #470499,
 #470921), then run debconf-updatepo and then use the
 podebconf-report-po utility to call for translation updates, leaving
 about 10 days to translators to cope with updates.
 
 Actually, I even suggest using 
 podebconf-report-po --call --languageteam so that a call for new
 translations is sent to debian-i18n and the language teams are
 notified of the needed changes (not only translators)
 
 Could you please keep me posted about this? Without news, I'll start
 an l10n NMU process..
 
 

Hi Christian,

I am getting back to this again now. I just sent another proposal out.
Hopefully we are starting to converge on the right wording for the
description.

I appreciate the guidance you have provided on this since this is the
first time I am dealing with debconf setup on a package. 

-- 
===
Khalid Aziz
[EMAIL PROTECTED]




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-14 Thread Christian Perrier
Quoting Christian Perrier ([EMAIL PROTECTED]):
 Quoting Justin B Rye ([EMAIL PROTECTED]):
 
 
  How about:
  
Should kexec-tools handle reboots? If you accept this option,
invocations of shutdown -r will trigger a restart into a kernel
loaded by kexec instead of going through the full system bootloader
process.
 
 
 Which would of course turn to:
 
 _Description: Should kexec-tools handle reboots? 
  If you accept this option, invocations of 'shutdown -r' will trigger
  a restart into a kernel loaded by kexec instead of going through the
  full system bootloader process.


Khalid, what's the status of this suggestion/rewrite of the debconf
template ?

Once you adopt the new template, you should indeed add to your package
tree the pending debconf l10n translations (#469960, #470499,
#470921), then run debconf-updatepo and then use the
podebconf-report-po utility to call for translation updates, leaving
about 10 days to translators to cope with updates.

Actually, I even suggest using 
podebconf-report-po --call --languageteam so that a call for new
translations is sent to debian-i18n and the language teams are
notified of the needed changes (not only translators)

Could you please keep me posted about this? Without news, I'll start
an l10n NMU process..




signature.asc
Description: Digital signature


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-05 Thread Justin B Rye
Khalid Aziz wrote:
 Please choose whether to enable kexec-tools to load a new kernel
 automatically on reboots and kexec into the new kernel instead of
 rebooting into firmware.
 
 Any thoughts?

Mainly, again, what's new about them? - kexec _can_ be used to
load a specified kernel, but that's not the behaviour you get from
automatic reloads handled by kexec-tools.  What it does by default
is load /vmlinuz, and that's usually an _old_ kernel.

Also, I'm not keen on firmware (too technical) or reboots (when
you mean restarts that avoid the bootstrapping process).  And this
is too much work for one sentence.

How about:

  Should kexec-tools handle reboots? If you accept this option,
  invocations of shutdown -r will trigger a restart into a kernel
  loaded by kexec instead of going through the full system bootloader
  process.

-- 
JBR with qualifications in linguistics, experience as a Debian
sysadmin, and probably no clue about this particular package



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-05 Thread Christian Perrier
Quoting Justin B Rye ([EMAIL PROTECTED]):


 How about:
 
   Should kexec-tools handle reboots? If you accept this option,
   invocations of shutdown -r will trigger a restart into a kernel
   loaded by kexec instead of going through the full system bootloader
   process.


Which would of course turn to:

_Description: Should kexec-tools handle reboots? 
 If you accept this option, invocations of 'shutdown -r' will trigger
 a restart into a kernel loaded by kexec instead of going through the
 full system bootloader process.


(using single quotes as we did in other reviews)



signature.asc
Description: Digital signature


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-04 Thread MJ Ray
Christian Perrier [EMAIL PROTECTED] wrote:
 You might need to explain us in more details what is a kexec'able kernel.

I think I would write either kexec-enabled or kexec-ready, depending
on whether it is the kernel that can kexec, or the kernel to be loaded
by kexec, respectively.

Hope that helps,
-- 
MJR/slef
My Opinion Only: see http://people.debian.org/~mjr/
Please follow http://www.uk.debian.org/MailingLists/#codeofconduct




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-04 Thread Khalid Aziz
On Fri, 2008-04-04 at 14:17 +0100, MJ Ray wrote:
 Christian Perrier [EMAIL PROTECTED] wrote:
  You might need to explain us in more details what is a kexec'able kernel.
 
 I think I would write either kexec-enabled or kexec-ready, depending
 on whether it is the kernel that can kexec, or the kernel to be loaded
 by kexec, respectively.
 
 Hope that helps,

I like this idea. OTOH I am wondering if this term can be dropped
altogether. kexec has been cleaned up enough that a standard kernel can
be kexec'd cleanly. It is only the kdump kernel that has to be built
specifically for that purpose. So even though kdump uses kexec to load
the kdump kernel, there is probably no need to let that influence the
kexec template. So with that it might be appropriate to change the
description to:

Please choose whether to enable kexec-tools to load a new kernel
automatically on reboots and kexec into the new kernel instead of
rebooting into firmware.

Any thoughts?

-- 
===
Khalid Aziz
[EMAIL PROTECTED]




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-03 Thread Christian Perrier
Quoting Khalid Aziz ([EMAIL PROTECTED]):
 On Tue, 2008-03-04 at 18:26 +0100, Christian Perrier wrote:
  
  Anyway, whether or not my suggestion is followed, the debconf
  templates need a serious rewrite to fit the commonly accepted writing style.
  Also, the long description is hard-formatted because of the use of
  leading double-spaces...which is not what should be done.
  
  I can propose rewritten templates if you wish.
 
 Hello Christian,
 
 I am getting ready to update kexec-tools package. Can you send me a
 proposal for template rewrite.


Could you repost the templates file to this bug report to make this
easier ?

I also suggest to not update the templates file without calling for
translation updates/new translations. That may take some time.




signature.asc
Description: Digital signature


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-03 Thread Khalid Aziz
On Thu, 2008-04-03 at 07:28 +0200, Christian Perrier wrote:
 Quoting Khalid Aziz ([EMAIL PROTECTED]):
  I am getting ready to update kexec-tools package. Can you send me a
  proposal for template rewrite.
 
 
 Could you repost the templates file to this bug report to make this
 easier ?
 
 I also suggest to not update the templates file without calling for
 translation updates/new translations. That may take some time.
 

Sure. Template file is attached. Yes, I will call for translations
before updating template file.

-- 
Khalid
===
Khalid Aziz
[EMAIL PROTECTED]
Template: kexec-tools/load_kexec
Type: select
_Choices: 1, 0
Default: 1
_Description: Enable automatic kexec reboots?
  You can enable kexec-tools to load a new kexec'able kernel 
  automatically upon reboot and kexec into the new kernel instead 
  of normal reboot. (1=enable, 0=disable)


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-03 Thread Christian Perrier
Quoting Khalid Aziz ([EMAIL PROTECTED]):

 Sure. Template file is attached. Yes, I will call for translations
 before updating template file.


I'd go for something liek the attached file.

But let's see what others on debian-l10n-english have to say.

You might need to explain us in more details what is a kexec'able kernel.

Template: kexec-tools/load_kexec
Type: boolean
Default: true
_Description: Enable automatic kexec reboots?
 Please choose whether you want kexec-tools to load a new kexec'able kernel 
 automatically on reboots and kexec into the new kernel instead 
 rather than rebooting normally.


signature.asc
Description: Digital signature


Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-03 Thread Justin B Rye
Christian Perrier wrote:
 But let's see what others on debian-l10n-english have to say.
 
 You might need to explain us in more details what is a kexec'able kernel.

 Template: kexec-tools/load_kexec
 Type: boolean
 Default: true
 _Description: Enable automatic kexec reboots?
  Please choose whether you want kexec-tools to load a new kexec'able kernel 
  automatically on reboots and kexec into the new kernel instead 
  rather than rebooting normally.

I don't like kexec'able, partly just for the apostrophe; you
don't kexec' a kernel.  In fact the normal usage at present (with
an example in the next line) is that you kexec into a kernel.  If
it isn't used as a transitive verb, it's a bit premature to coin a
fooable adjective.  The closest I would risk is kexec-capable.

But this phrasing also has an ambiguity: if I have two kernels
currently installed, 

 /vmlinuz   = linux-image-2.6.25-sanskexec
 /vmlinuz.old   = linux-image-2.6.24-withkexec

and I activate automatic kexec reboots, it _won't_ automatically
pick the kexec-capable kernel when I reboot, will it?  Indeed, if
I've also got a freshly-installed

 /vmlinuz.new   = linux-image-2.6.25-alpha

then it will specifically try to kexec into an old kernel rather
than the new kernel.  So I'd suggest something more like:

   Please choose whether you want reboots to be handled by kexec-tools.
   If you accept this option, instead of performing normal reboots the
   system will attempt to use kexec to load /vmlinuz (if that kernel
   supports this).

Is this correct?
-- 
JBR with qualifications in linguistics, experience as a Debian
sysadmin, and probably no clue about this particular package



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-04-02 Thread Khalid Aziz
On Tue, 2008-03-04 at 18:26 +0100, Christian Perrier wrote:
 
 Anyway, whether or not my suggestion is followed, the debconf
 templates need a serious rewrite to fit the commonly accepted writing style.
 Also, the long description is hard-formatted because of the use of
 leading double-spaces...which is not what should be done.
 
 I can propose rewritten templates if you wish.

Hello Christian,

I am getting ready to update kexec-tools package. Can you send me a
proposal for template rewrite.

Thanks
-- 
Khalid
===
Khalid Aziz
[EMAIL PROTECTED]




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-03-04 Thread Christian Perrier
Package: kexec-tools
Severity: normal

Your newly added debconf template wrongly uses a select question to ask an
enable/disable question.

Please use a boolean type for this.

-- System Information:
Debian Release: lenny/sid
  APT prefers unstable
  APT policy: (500, 'unstable')
Architecture: i386 (i686)

Kernel: Linux 2.6.24-1-686 (SMP w/1 CPU core)
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-03-04 Thread Khalid Aziz
On Tue, 2008-03-04 at 09:14 +0100, Christian Perrier wrote:
 Package: kexec-tools
 Severity: normal
 
 Your newly added debconf template wrongly uses a select question to ask an
 enable/disable question.
 
 Please use a boolean type for this.
 
 -- System Information:
 Debian Release: lenny/sid
   APT prefers unstable
   APT policy: (500, 'unstable')
 Architecture: i386 (i686)
 
 Kernel: Linux 2.6.24-1-686 (SMP w/1 CPU core)
 Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8)
 Shell: /bin/sh linked to /bin/dash
 
 

Use of select is intentional. There are other possibilities besides just
enable/disable for kexec reboot that others have already suggested, for
example do not load a kexec'able kernel automatically but attempt to
kexec in case user has loaded a kexec kernel by hand etc. So I do not
want to lock myself into only two states.
 
===
Khalid Aziz
[EMAIL PROTECTED]




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#469256: kexec-tools: Please use a boolean question with debconf

2008-03-04 Thread Christian Perrier
Quoting Khalid Aziz ([EMAIL PROTECTED]):

  Your newly added debconf template wrongly uses a select question to ask an
  enable/disable question.
  
  Please use a boolean type for this.
  
  -- System Information:
  Debian Release: lenny/sid
APT prefers unstable
APT policy: (500, 'unstable')
  Architecture: i386 (i686)
  
  Kernel: Linux 2.6.24-1-686 (SMP w/1 CPU core)
  Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8)
  Shell: /bin/sh linked to /bin/dash
  
  
 
 Use of select is intentional. There are other possibilities besides just
 enable/disable for kexec reboot that others have already suggested, for
 example do not load a kexec'able kernel automatically but attempt to
 kexec in case user has loaded a kexec kernel by hand etc. So I do not
 want to lock myself into only two states.

Well, using 1/0 is confusing, then. Why not propose enable/disable
as choices. 

Anyway, whether or not my suggestion is followed, the debconf
templates need a serious rewrite to fit the commonly accepted writing style.
Also, the long description is hard-formatted because of the use of
leading double-spaces...which is not what should be done.

I can propose rewritten templates if you wish.





signature.asc
Description: Digital signature