Roboot's BOF

Roboot's BOF

[Eric Salomé]

Roboot's birth of feather

Dear all,

I’d like to talk about some ideas that came to me a few months ago when
I heard about Grub (legacy) and gave it a try on my PC (Linux +
Windows).
I thought (like others) Grub needed some enhancements and I see most of
them are already on their way with Grub 2.
Thoughts, ideas are like free software, they get better when shared, so
let’s try this one :

Grub legacy’s lack of scripting feature, but also a need for modularity,
expandability + a project I care for (and personal experience among
other things), mix that and it gets : 
why not try Grub + Forth to see how it goes ?

Let’s call it “Roboot”.

I’ve spend some free time doing a “thing that works for me” as a start.
Technically, It turns out as a boot loader (using Grub’s stage1 –
stage1.5) embedded with full ANSI Forth compiler/interpreter that calls
Grub’s library routines to do keyboard/screen/disk input/output, etc. 
Calling Grub’s menu functions or any registered Grub C functions can be
mixed in Forth scripts to get powerful scripting features. 
It computes 1 + 1 and it loads OS among other things.

Also works with underlying Linux OS on i386. And no, it doesn’t
read/execute Fcode device drivers yet (if there is any sparc fan among
us).

All of current Grub 2 development sources will be useful to next
versions of “Roboot” if any, including :
- better memory management
- modularity in supported input/output devices
- support for dynamic loading of C routines (Roboot already have Forth
modules dynamic loading)
- even scripting routines (though it won’t be used the same)

I still had some spare time to see how it goes. So that I have a running
Grub 2 module, built on “BigForth” by Bernd Paysan, that adds a Forth
engine to current version of Grub 2 development. It also runs on
grub-emu under i386-pc.
	
I kind of believe that “Roboot” tries might be useful to current/next
versions of Grub in some way. What’s your opinion ?

Best Regards,

Eric Salomé - Paris, France

Re: Roboot's BOF

[Stefan Reinauer]

* Eric Salomé <esalome@ctx.net> [060526 14:57]:
> Grub legacy's lack of scripting feature, but also a need for modularity,
> expandability + a project I care for (and personal experience among
> other things), mix that and it gets : 
> why not try Grub + Forth to see how it goes ?
> 
> Let's call it "Roboot".
 
Something very similar exists already. Its called OpenBIOS. See 
www.openbios.org for more details.

Stefan

-- 
coresystems GmbH • Brahmsstr. 16 • D-79104 Freiburg i. Br.
      Tel.: +49 761 7668825 • Fax: +49 761 7664613
Email: info@coresystems.de  • http://www.coresystems.de/

RE: Roboot's BOF

[Gregg C Levine]

Hello!
Stefan, Eric, when I saw the earlier message, to which that you Stefan made your contribution I was about to bring to the list's, and of course Eric's attention your OpenBIOS project at www.openbios.org . Now seeing that you have done so, I am going to call it a coincidence. And also comment that like minds do think alike.

Yes, Eric, you should indeed in your spare time, do look at the items hosted on the OpenBIOS project pages, they make for an interesting application of the Forth language to the world of the BIOS and its methods. 
--
Gregg C Levine hansolofalcon@worldnet.att.net
---
"Remember the Force will be with you. Always." Obi-Wan Kenobi 

> -----Original Message-----
> From: grub-devel-bounces+hansolofalcon=worldnet.att.net@gnu.org [mailto:grub-
> devel-bounces+hansolofalcon=worldnet.att.net@gnu.org] On Behalf Of Stefan
> Reinauer
> Sent: Friday, May 26, 2006 11:57 AM
> To: The development of GRUB 2
> Subject: Re: Roboot's BOF
> 
> * Eric Salomé <esalome@ctx.net> [060526 14:57]:
> > Grub legacy's lack of scripting feature, but also a need for modularity,
> > expandability + a project I care for (and personal experience among
> > other things), mix that and it gets :
> > why not try Grub + Forth to see how it goes ?
> >
> > Let's call it "Roboot".
> 
> Something very similar exists already. Its called OpenBIOS. See
> www.openbios.org for more details.
> 
> Stefan
> 
> --
> coresystems GmbH • Brahmsstr. 16 • D-79104 Freiburg i. Br.
>       Tel.: +49 761 7668825 • Fax: +49 761 7664613
> Email: info@coresystems.de  • http://www.coresystems.de/
> 
> 
> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/grub-devel

RE : Roboot's BOF

[Eric Salomé]

> > Let's call it "Roboot".
> 
> Something very similar exists already. Its called OpenBIOS. See
> www.openbios.org for more details.
> 
> Stefan

I've glanced over the Internet looking for OpenBIOS and other software. 
I've found an OpenBIOS version 1.1 that fullfills well its goals. 
(Can't open www.openbios.org on my PC :-( so I don't know what's new).

Using Forth at boot time isn't a fresh new idea, but I believe (as you
do) it's a very good one.

As a start, I thought it might be interesting to have a Forth engine
make use of the Grub 2's commands. 
Let's say it can be thought of as a script helper for situations that
won't be handled by the scripting language of Grub 2.

Well, it's done.

But there is more than that. In fact my belief is there is a life before
booting OS, and a life after the OS gives up (on user’s request or when
it can’t go on any more (windows’ blue screen)). A life full of many
things to do:
-         boot an (other) OS
-         connect to the net to request for Roboot’s Web assistance
-         partition disks and other flavours
-         examine current configuration, diagnostic misfits
-         add a new OS in the PC 
-         advertise central server of boot events
-         look for email incoming and decide not to boot if there’s
nothing new
-         have a calculator, play cards, etc.

You might be interested in ? 

I don't know what Marco, Yoshinori and others would like to do for Grub
3 to X, so I just ask.

Even if it happens to be two different projects, I kind of believe one
will gain from the other.

_________________________________________
Eric Salomé - Paris - France

Re: RE : Roboot's BOF

[Marco Gerards]

Eric Salomé <esalome@ctx.net> writes:

> But there is more than that. In fact my belief is there is a life before
> booting OS, and a life after the OS gives up (on user’s request or when
> it can’t go on any more (windows’ blue screen)). A life full of many
> things to do:
> -         boot an (other) OS
> -         connect to the net to request for Roboot’s Web assistance
> -         partition disks and other flavours
> -         examine current configuration, diagnostic misfits
> -         add a new OS in the PC 
> -         advertise central server of boot events
> -         look for email incoming and decide not to boot if there’s
> nothing new
> -         have a calculator, play cards, etc.

Personally, I do not see how this will add anything to what GRUB 2 is
capable of, or what we planned to change in GRUB.  With the scripting
support planned (and partially implemented), you can do most of these
things.  Some other more fancy things can be done by writing a module
in C or whatever.  Please tell me if I misunderstood your goals.

It would be easy to add a command to GRUB 2 that makes it possible to
parse other languages.  What you need is a new parser.  You could
create some new AST nodes or so and add all this to GRUB 2 from within
your module.  All this should integrate quite well.  This way, you
would be able to create new commands, functions, etc in another
language.

Personally I prefer to see an add-in module, even one maintained
outside of GRUB, than a complete fork.  But you would have to keep in
mind that we can make interface changes when we see the need for this.

> I don't know what Marco, Yoshinori and others would like to do for Grub
> 3 to X, so I just ask.

First I would like to see GRUB 2 being finished.  It is not there
yet. :-)

--
Marco

RE : RE : Roboot's BOF

[Eric Salomé]

> It would be easy to add a command to GRUB 2 that makes it possible to
> parse other languages.  What you need is a new parser.  You could
> create some new AST nodes or so and add all this to GRUB 2 from within
> your module.  All this should integrate quite well.  This way, you
> would be able to create new commands, functions, etc in another
> language.

I give this "integrate work" a try and so far so good. Had found nothing
that would have been *mandatory* to change in grub 2's kernel. If I find
on the way some good ideas to share I'll express them on this
mailing-list for comments. 

> Personally I prefer to see an add-in module, even one maintained
> outside of GRUB, than a complete fork.  But you would have to keep in
> mind that we can make interface changes when we see the need for this.

That's what roboot aim to be, an add-in module.
I understand we will have to face some interface changes from time to
time but I am ready for that too ... and that will make things even
better.

> > I don't know what Marco, Yoshinori and others would like to do for
Grub
> > 3 to X, so I just ask.
> 
> First I would like to see GRUB 2 being finished.  It is not there
> yet. :-)
  
> --
> Marco
>

Well, we are all working on that part too ! and I am full of hope we are
going to make it.
I am quite confused with drivers dealing with BIOS, but I don't want to
feel helpless, so I'll try to do something useful with grub-emu and let
you know if I can work on grub's kernel (is there anything missing you'd
like to add ?). 
By the way, is there anybody working on having the vbe interface working
on grub-emu ? Any hint to help me start work on it ?


Best regards,
_________________________________________
Eric Salomé - Paris - France

Re: RE : RE : Roboot's BOF

[Vesa Jääskeläinen]

Eric Salomé wrote:

> By the way, is there anybody working on having the vbe interface working
> on grub-emu ? Any hint to help me start work on it ?

Not that I know.

Write video driver for it. See video/ folder for more details and in
wiki (when it comes back):
http://grub.enbug.org/VideoSubsystem

Video API is still being fine tuned, so try to keep in sync of it.