SOC - SGD based on grub2 - Menu and scripting status

SOC - SGD based on grub2 - Menu and scripting status

[adrian15]

I've reading the G'SOC document:

http://www.gnu.org/software/grub/grub-soc.html

and I've been tented to add a new proposal for grub2.

As you all should have guessed it would a Super Grub Disk ( 
http://adrian15.raulete.net/grub/ ) but not based on grub legacy but grub2.

	If I want the project to be feasible I need that the grub2 api relative 
to menues and scriptings is fixed. That means that the name of the 
commands and its behavior don't change since I begin the project to the 
end of it.

	So here are some questions that have arised...

1) Marco Gerards: Do you think that commands/functions/variables related 
to menues files are going to change a lot?
But I don't mean RIGHT NOW. I mean in the beginnings of SOC... which I 
don't know... let's see... 23th May. (in one month time).

	If these commands are available where should I get a list of them and 
their behaviour?

2) Do you like the Super Grub2 CD name... or should it be Grub2 CD or 
Grub CD and that's all? Grub Rescue CD perhaps?

3) What features would you like to see apart from the ones that are 
already implemented in Super Grub Disk ( 
http://adrian15.raulete.net/grub/tiki-index.php?page=En )?

4) What is the best way for dealing with languages and translations when 
developing?

5) I need to define "something" (In my grub legacy fork is a menu.lst) 
that lets me choose the harddisk-partition which I want to work on. Can 
I make it easily?

Can Grub2 api write me in an array or a similar the partitions that can 
read in a given hard disk? So that if there are 3 partitions they are 
saved into an array of size 3?

adrian15

Re: SOC - SGD based on grub2 - Menu and scripting status

[Marco Gerards]

adrian15 <adrian15@raulete.net> writes:

Hi,

> As you all should have guessed it would a Super Grub Disk (
> http://adrian15.raulete.net/grub/ ) but not based on grub legacy but
> grub2.

Oh, nice/

> 	If I want the project to be feasible I need that the grub2 api
> relative to menues and scriptings is fixed. That means that the name
> of the commands and its behavior don't change since I begin the
> project to the end of it.

I can not make such promise.  But I am open to feature suggestions as
far as I did not make any promises or so yet.

> 	So here are some questions that have arised...
>
> 1) Marco Gerards: Do you think that commands/functions/variables
> related to menues files are going to change a lot?
> But I don't mean RIGHT NOW. I mean in the beginnings of SOC... which I
> don't know... let's see... 23th May. (in one month time).

The will keep changing, I think.

> 	If these commands are available where should I get a list of
> them and their behaviour?

Not yet.  We still hope someone will write good user documentation.

> 2) Do you like the Super Grub2 CD name... or should it be Grub2 CD or
> Grub CD and that's all? Grub Rescue CD perhaps?

I personally do not have a strong opinion about this. :-)

> 4) What is the best way for dealing with languages and translations
> when developing?

We are thinking about i18n support.  This is something about this on
the wiki, I think.

> 5) I need to define "something" (In my grub legacy fork is a menu.lst)
> that lets me choose the harddisk-partition which I want to work
> on. Can I make it easily?

You mean as a variable?  That's possible.

> Can Grub2 api write me in an array or a similar the partitions that
> can read in a given hard disk? So that if there are 3 partitions they
> are saved into an array of size 3?

No, at least not yet.  Scripting is in continuous development ATM.

--
Marco

Re: SOC - SGD based on grub2 - Menu and scripting status

[Yoshinori K. Okuji]

On Monday 24 April 2006 18:49, Marco Gerards wrote:
> > 4) What is the best way for dealing with languages and translations
> > when developing?
>
> We are thinking about i18n support.  This is something about this on
> the wiki, I think.

I don't think this part is described well in the wiki. So I describe my idea 
here:

- GRUB needs to have a locale variable to determine a language.

- GRUB needs to have gettext or something similar to translate messages.

- GRUB needs to be able to generate .po files compatible with gettext.

- GRUB needs to support message translations by scripting as well as by C 
code.

These are not implemented at all yet, but not difficult. I expect that this 
would take 2 weeks, if one works in full time.

Okuji

Re: SOC - SGD based on grub2 - Menu and scripting status

[Marco Gerards]

"Yoshinori K. Okuji" <okuji@enbug.org> writes:

> On Monday 24 April 2006 18:49, Marco Gerards wrote:
>> > 4) What is the best way for dealing with languages and translations
>> > when developing?
>>
>> We are thinking about i18n support.  This is something about this on
>> the wiki, I think.
>
> I don't think this part is described well in the wiki. So I describe my idea 
> here:
>
> - GRUB needs to have a locale variable to determine a language.
>
> - GRUB needs to have gettext or something similar to translate messages.
>
> - GRUB needs to be able to generate .po files compatible with gettext.

Will this mean Adrian can use it for menu entries as well?  I think
that is what he wants to do.

--
Marco

Re: SOC - SGD based on grub2 - Menu and scripting status

[Yoshinori K. Okuji]

On Monday 24 April 2006 20:36, Marco Gerards wrote:
> Will this mean Adrian can use it for menu entries as well?  I think
> that is what he wants to do.

In my idea, yes. I can say anything before implementing. ;)

Okuji

Re: SOC - SGD based on grub2 - Menu and scripting status

[Vesa Jääskeläinen]

Marco Gerards wrote:
> "Yoshinori K. Okuji" <okuji@enbug.org> writes:
> 
>> On Monday 24 April 2006 18:49, Marco Gerards wrote:
>>>> 4) What is the best way for dealing with languages and translations
>>>> when developing?
>>> We are thinking about i18n support.  This is something about this on
>>> the wiki, I think.
>> I don't think this part is described well in the wiki. So I describe my idea 
>> here:
>>
>> - GRUB needs to have a locale variable to determine a language.
>>
>> - GRUB needs to have gettext or something similar to translate messages.
>>
>> - GRUB needs to be able to generate .po files compatible with gettext.
> 
> Will this mean Adrian can use it for menu entries as well?  I think
> that is what he wants to do.

I think this is one of the possibilities. Giving support to script in
form like this:

menuentry $gettext("My menu entry text") {

}

Though gettext is not perfect in all places, but could be enough for us.

I am still wondering what are the features of the scripting language.
Marco could you write some summary of your ideas somewhere and some
syntaxical examples?

Thanks,
Vesa Jääskeläinen

Re: SOC - SGD based on grub2 - Menu and scripting status

[Marco Gerards]

Vesa Jääskeläinen <chaac@nic.fi> writes:

> Marco Gerards wrote:
>> "Yoshinori K. Okuji" <okuji@enbug.org> writes:
>> 
>>> On Monday 24 April 2006 18:49, Marco Gerards wrote:
>>>>> 4) What is the best way for dealing with languages and translations
>>>>> when developing?
>>>> We are thinking about i18n support.  This is something about this on
>>>> the wiki, I think.
>>> I don't think this part is described well in the wiki. So I describe my idea 
>>> here:
>>>
>>> - GRUB needs to have a locale variable to determine a language.
>>>
>>> - GRUB needs to have gettext or something similar to translate messages.
>>>
>>> - GRUB needs to be able to generate .po files compatible with gettext.
>> 
>> Will this mean Adrian can use it for menu entries as well?  I think
>> that is what he wants to do.
>
> I think this is one of the possibilities. Giving support to script in
> form like this:
>
> menuentry $gettext("My menu entry text") {
>
> }
>
> Though gettext is not perfect in all places, but could be enough for us.

That sounds sane to me.

> I am still wondering what are the features of the scripting language.
> Marco could you write some summary of your ideas somewhere and some
> syntaxical examples?

Sure.  I wrote a lot of emails about it.  But it might be nice to have
something more coherent on the wiki.  I'll work on
that... someday. :)

--
Marco

Re: SOC - SGD based on grub2 - Menu and scripting status

[adrian15]

>>5) I need to define "something" (In my grub legacy fork is a menu.lst)
>>that lets me choose the harddisk-partition which I want to work
>>on. Can I make it easily?
> 
> 
> You mean as a variable?  That's possible.

I will explain this point in more detail if you don't mind. Here goes 
copy-pasted
1) the first the "lst" for hard disks and then
2)the second one the "sh" that makes the "lst" for partitions.
3) An "lst" for making easier to call the selection of partitions from 
inside a function or whatever.
4) An example of how this menues/selectors are called. (lst)

I'm not asking you to develop a command for giving me this lst on the 
fly adapted to the user pc. You know what I was telling before: If the 
users hard disk has only three partitions make an array with only three 
values which are the partitions.

I'm asking you to think about new features that scripting should have so 
that the developing of this command is far easy.

OFFTOPIC NOTE: I am working on grub legacy call and back commands so 
that what you see below actually works. I think this night it will work 
ok but who knows.

adrian15


hard.disk.lst:
==============

title $(choose_title)
pause
title NATURAL	LINUX-IDE LINUX-SCSI GNU/GRUB
pause
title 1. 	hda 	  sda 	     hd0
set out_device=hd0
set out_hd=hd0
set out_lide_hd=hda
set out_lscsi_hd=sda
back

title 2. 	hdb 	  sdb 	     hd1
set out_device=hd1
set out_hd=hd1
set out_lide_hd=hdb
set out_lscsi_hd=sdb
back

title 3. 	hdc 	  sdc 	     hd2
set out_device=hd2
set out_hd=hd2
set out_lide_hd=hdc
set out_lscsi_hd=sdc
back

title 4. 	hdd 	  sdd 	     hd3
set out_device=hd3
set out_hd=hd3
set out_lide_hd=hdd
set out_lscsi_hd=sdd
back

generate_part_lst.sh:
======================

SDG_PARTITIONS_LIST="0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17";



cat << EOF

title \$(choose_title)
pause
title NATURAL	LINUX-IDE	LINUX-SCSI	GNU/GRUB	GNU/HURD
pause
EOF


for n_partition in $SDG_PARTITIONS_LIST; do

ln_partition=$[$n_partition + 1];


cat << EOF


title ${ln_partition}	\$(out_lide_hd)${ln_partition} 
\$(out_lscsi_hd)${ln_partition}	(\$(out_hd),${n_partition}) 
\$(out_hurd_hd)s${ln_partition}
set out_device=(\$(out_hd),${n_partition})
set out_part=(\$(out_hd),${n_partition})
set out_lide_part=\$(out_lide_hd)${ln_partition}
set out_lscsi_part=\$(out_lscsi_hd)${ln_partition}
set out_hurd_part=\$(out_hurd_hd)s${ln_partition}

EOF

done

partition.lst:
===============
default 0
timeout 0
title  Choose hard disk and partition
call $(grub_device)/boot/grub/choose/hard_disk.lst
call $(grub_device)/boot/grub/choose/part.lst
back

test.lst:
===========
title THIS IS A TEST FOR CALLING hard disk
set choose_title="TEST FOR CHOOSING HARD DISK TITLE"
call $(grub_device)/boot/grub/choose/hard_disk.lst
title THIS IS A TEST FOR CALLING part.lst
set choose_title="TEST FOR CHOOSING PART ONLY TITLE"
call $(grub_device)/boot/grub/choose/part.lst
title THIS IS A TEST FOR CALLING partition (Both hard disk and partition)
set choose_title="TEST FOR CHOOSING PART AND HD TITLE"
call $(grub_device)/boot/grub/choose/partition.lst

This is the begginnings of the generated part.lst and you know it is 9 
times bigger:
=======================================================================
title $(choose_title)
pause
title NATURAL	LINUX-IDE	LINUX-SCSI	GNU/GRUB	GNU/HURD
pause


title 1	$(out_lide_hd)1	$(out_lscsi_hd)1	($(out_hd),0)	$(out_hurd_hd)s1
set out_device=($(out_hd),0)
set out_part=($(out_hd),0)
set out_lide_part=$(out_lide_hd)1
set out_lscsi_part=$(out_lscsi_hd)1
set out_hurd_part=$(out_hurd_hd)s1



title 2	$(out_lide_hd)2	$(out_lscsi_hd)2	($(out_hd),1)	$(out_hurd_hd)s2
set out_device=($(out_hd),1)
set out_part=($(out_hd),1)
set out_lide_part=$(out_lide_hd)2
set out_lscsi_part=$(out_lscsi_hd)2
set out_hurd_part=$(out_hurd_hd)s2


adrian15

Re: SOC - SGD based on grub2 - Menu and scripting status

[adrian15]

> On Monday 24 April 2006 20:36, Marco Gerards wrote:
> 
>>Will this mean Adrian can use it for menu entries as well?  I think
>>that is what he wants to do.
> 
> 
> In my idea, yes. I can say anything before implementing. ;)
> 
> Okuji

Yes... The thing that when meaning translation I was thinking about SGD, 
which is kind of a GUI for Grub so... I did not think about Grub 
messages but to menu entries as Marco suggested.

However you're right Grub messages could also be translated.

Currently I have SGD (only menu entries) translated to various languages 
with a three-level structure... one for algorithm... one for "kind of 
universal language" and one for the actual languages. It's implemented 
with bash variables and uses sed and other unix tools and it is not 
perfect...

I know there was this gettext thing and in the first place I didn't get 
to work it with bash (I think I needed to recompile it) so I discarded 
the idea.

Maybe I will take a look at gettext again for SG2D.

adrian15

Re: SOC - SGD based on grub2 - Menu and scripting status

[Marco Gerards]

adrian15 <adrian15@raulete.net> writes:

>>>5) I need to define "something" (In my grub legacy fork is a menu.lst)
>>>that lets me choose the harddisk-partition which I want to work
>>>on. Can I make it easily?
>> You mean as a variable?  That's possible.
>
> I will explain this point in more detail if you don't mind. Here goes
> copy-pasted
> 1) the first the "lst" for hard disks and then
> 2)the second one the "sh" that makes the "lst" for partitions.
> 3) An "lst" for making easier to call the selection of partitions from
> inside a function or whatever.
> 4) An example of how this menues/selectors are called. (lst)
>
> I'm not asking you to develop a command for giving me this lst on the
> fly adapted to the user pc. You know what I was telling before: If the
> users hard disk has only three partitions make an array with only
> three values which are the partitions.
>
> I'm asking you to think about new features that scripting should have
> so that the developing of this command is far easy.

I was thinking about an iterator function.  Perhaps that might be the
easiest to use...  Please watch CVS closely.  I am actively working on
scripting, as much as time permits.

--
Marco

Re: SOC - SGD based on grub2 - Menu and scripting status

[adrian15]

>>>>5) I need to define "something" (In my grub legacy fork is a menu.lst)
>>>>that lets me choose the harddisk-partition which I want to work
>>>>on. Can I make it easily?
>>>
> 
> I was thinking about an iterator function.  Perhaps that might be the
> easiest to use...  Please watch CVS closely.  I am actively working on
> scripting, as much as time permits.

An iterator function among the various partitions of a hard disk... that 
could be interesting. Unfortunately I can't watch CVS closely so I will 
wait till you announce it in this list. :)

The iterator function should increment a variable for knowing the total 
amount of partitions on a given hard disk.

adrian15

Re: Changing the default OS

[adrian15]

> It's very interesting why people ask questions about GRUB Legacy here again 
> and again, although I have written "Do not use this list for GRUB Legacy" in 
> everywhere.
> 
> Okuji

Is it so difficult to rename mailing-lists ?

Rename bug-grub to grub
Rename grub-devel to grub2 or grubbeta and you will see how there are 
some changes.

adrian15

Re: Changing the default OS

[Marco Gerards]

adrian15 <adrian15@raulete.net> writes:

>> It's very interesting why people ask questions about GRUB Legacy
>> here again and again, although I have written "Do not use this list
>> for GRUB Legacy" in everywhere.
>> Okuji
>
> Is it so difficult to rename mailing-lists ?

It's besides the point.  A development list is not appropriate for
non-development stuff, I think that is quite obvious.  A lot of
projects even don't allow non project members to post on the devel
list.

--
Marco