chaos-devel Mailing List for The chaos Operating System

> > I think just the needed basics should be left in, with all complex
> > chaos stuff configured already. One image for the compilation+tftp and
> > one for the testing, that are setup so it's enough to just do a
> > compile and then everything runs. The develop can then install what
> > ever environment he likes whether it's command line with vi, or Gnome
> > and xemacs.
>
> Yep...  That sounds like the way I have done it.  :-)  Was it easy to
> set it up (if we disregard your special Samba configuration and that
> stuff, and keep to the other stuff)?  Install/configure vmWare etc...

Well, I don't remember how I did it, so I probably didn't have to put
too much thought into it, so it must have been easy. =)

// Anders

to 2006-07-27 klockan 13:33 +0200 skrev Anders Öhrt:
> I think just the needed basics should be left in, with all complex
> chaos stuff configured already. One image for the compilation+tftp and
> one for the testing, that are setup so it's enough to just do a
> compile and then everything runs. The develop can then install what
> ever environment he likes whether it's command line with vi, or Gnome
> and xemacs.

Yep...  That sounds like the way I have done it.  :-)  Was it easy to
set it up (if we disregard your special Samba configuration and that
stuff, and keep to the other stuff)?  Install/configure vmWare etc...
-- 
Best regards,
Per Lundberg

> We could make a vmware image that has X and GNOME etc, but I don't know,
> perhaps keeping it simple is a good idea.  I think it is nice to have it
> be small for people with low-speed broadband etc.

I think just the needed basics should be left in, with all complex
chaos stuff configured already. One image for the compilation+tftp and
one for the testing, that are setup so it's enough to just do a
compile and then everything runs. The develop can then install what
ever environment he likes whether it's command line with vi, or Gnome
and xemacs.


> If we make it too
> simple, perhaps we get too many newbies trying to be kernel developers
> with all the fatal consequences that will bring.  :-P

Hehe...

// Anders

sö 2006-07-23 klockan 23:39 +0200 skrev Anders Öhrt:
> > Great to see that you're coming somewhere.  :)
> Yes, the environment is pretty complex, but works nice anyway. I can
> use Total Commander together with Xemacs, so I don't need to change
> all the tools I'm used to and don't need to be in linux textmode more
> that necessary. =)

We could make a vmware image that has X and GNOME etc, but I don't know,
perhaps keeping it simple is a good idea.  I think it is nice to have it
be small for people with low-speed broadband etc.  If we make it too
simple, perhaps we get too many newbies trying to be kernel developers
with all the fatal consequences that will bring.  :-P
-- 
Best regards,
Per Lundberg

> > Got back just a few hours ago, we went shopping in Ullared (Gek=E5s) an=
d
> > then stayed with a friend in Halmstad, the beach in Tyl=F6sand was
> > amazing.
>
> Ah. You missed me with a couple of days. ;) (GK)

Too bad, I could have used someone else to talk to besides my
'father-in-law' for the hours where my girlfriend was hunting
clothes... =3D)

// Anders

On 7/23/06, Anders =D6hrt <and...@gm...> wrote:

> > I hope you're having a good trip.
>
> Got back just a few hours ago, we went shopping in Ullared (Gek=E5s) and
> then stayed with a friend in Halmstad, the beach in Tyl=F6sand was
> amazing.

Ah. You missed me with a couple of days. ;) (GK)

--=20
6accd=E613eff7i319n4o4qrr4s8t12ux

> Great to see that you're coming somewhere.  :)

Yes, the environment is pretty complex, but works nice anyway. I can
use Total Commander together with Xemacs, so I don't need to change
all the tools I'm used to and don't need to be in linux textmode more
that necessary. =3D)


> I hope you're having a good trip.

Got back just a few hours ago, we went shopping in Ullared (Gek=E5s) and
then stayed with a friend in Halmstad, the beach in Tyl=F6sand was
amazing.


>  Lule=E5 btw (that you visited recently)
> is fairly close from here...  only 260 km up north.

I've never been in Ume=E5 except passing through, but we (Berendsen) do
have a laundry there somewhere, so I'm a bit familiar with it. =3D)

// Anders

on 2006-07-19 klockan 21:57 +0200 skrev Anders Öhrt:

Great to see that you're coming somewhere.  :)

> > I've got another problem with this now... I want to share my ~/chaos
> > via samba, so I can do the editing in my windows environment. Once I
> > manage (or give up) on setting that up, I'll fix the compile issues.
> I've done these things and wont get further for now (going away on a
> trip tomorrow, back sunday). I need the chaos-ramdisk and fstab entry
> and I'm to lazy to figure them out by myself, so can you provide them?
> =)

Something like this:

/home/chaos/chaos-from-anoncvs/modules/block/initial_ramdisk/ramdisk.image /mnt/chaos-ramdisk minix loop,rw,user,noauto 0 0

/mnt/chaos-ramdisk needs to be created for this to work.  When this line
is in place, the build system should take care of the rest (mounting it
when needed and unmounting afterwards) to copy in the boot program and
other programs.  (but AFAIK everything else is commented out right now,
the modplayer doesn't work unfortunately.  :)

Regarding the CVS account: I'll gladly fix that but you'll need to
create a user account on SourceForge first. 

https://sourceforge.net/account/newuser_emailverify.php

Do that and email me the account name (privately) and I'll fix the rest.

Regarding CVS, you should probably be able to use a Windows CVS client
(such as TortoiseCVS) as long as cons doesn't barf about the Windows
line endings (CR+LF instead of LF only).  Try it and see, if you are
already familiar with Tortoise it is easier to get going with of course.
Otherwise, take a look at this book: http://cvsbook.red-bean.com/

Or a five-minute tutorial:
- cvs checkout (or co) - checkout source.  See
http://sourceforge.net/cvs/?group_id=2981 for info on this.  You need to
specify your user name on SourceForge when doing the checkout.
- cvs update - update a previously checked out source tree.
- cvs diff -u to see your changes.
- cvs commit <file> to commit one file or just cvs commit to commit
everything.

Everything (almost) else you need to know is in the CVS Book.  ;)

I hope you're having a good trip.  Luleå btw (that you visited recently)
is fairly close from here...  only 260 km up north.
-- 
Best regards,
Per Lundberg

> I've got another problem with this now... I want to share my ~/chaos
> via samba, so I can do the editing in my windows environment. Once I
> manage (or give up) on setting that up, I'll fix the compile issues.

I've done these things and wont get further for now (going away on a
trip tomorrow, back sunday). I need the chaos-ramdisk and fstab entry
and I'm to lazy to figure them out by myself, so can you provide them?
=)

// Anders

> Okay... I haven't tried that actually, running Debian sarge here on my
> server (the same version that was in the disk image).  I compiled chaos
> with no big problems at all actually in the disk image so perhaps
> reverting to the one from me would be an idea?  Unless you've done a lot
> of work with the setup of course...  But then again, of course we need
> to fix it to make it compile with gcc 4.1 sooner or later.  :-)

I've got another problem with this now... I want to share my ~/chaos
via samba, so I can do the editing in my windows environment. Once I
manage (or give up) on setting that up, I'll fix the compile issues. I
guess I'll have to figure out CVS too, never really used the command
line version much. Can you give me an account?


> Great!  One small tip: you could set up remote debugging, vmware seems
> to be able to support it.  Create a new serial port with a named pipe,
> and use one virtual machine as the "server" and the other one as a
> client.  Then compile stormG3 with some special parameters and you
> should be able to debug it with gdb which is of course very helpful,
> being able to do proper backtraces, show contents of memory etc...  I
> used to work like that in 2002 and it was very convenient.  (but I was
> using a null modem cable back then.  :)

There really is quite a lot of messy stuff to setup in order to get a
good environment to work with this. I'll probably settle for managing
to code, compile and testboot for a while. I might buy an old computer
or two for this later, I'll have to see if I'm to old and senile to
get back into this first. =)

// Anders

On Tue, 2006-07-18 at 22:53 +0200, Anders Öhrt wrote:
> I've gotten the debian stuff to work, but I have problems compiling
> chaos. I upgraded to unstable, so I got gcc 4.1 and it had lots of
> complaints about signedness and aliasing. I fixed most of them but got
> stuck on some...

Okay... I haven't tried that actually, running Debian sarge here on my
server (the same version that was in the disk image).  I compiled chaos
with no big problems at all actually in the disk image so perhaps
reverting to the one from me would be an idea?  Unless you've done a lot
of work with the setup of course...  But then again, of course we need
to fix it to make it compile with gcc 4.1 sooner or later.  :-)

> Except for this, it appeared to work pretty good. Once I've gotten it
> compiling I think the rest should be easy.

Great!  One small tip: you could set up remote debugging, vmware seems
to be able to support it.  Create a new serial port with a named pipe,
and use one virtual machine as the "server" and the other one as a
client.  Then compile stormG3 with some special parameters and you
should be able to debug it with gdb which is of course very helpful,
being able to do proper backtraces, show contents of memory etc...  I
used to work like that in 2002 and it was very convenient.  (but I was
using a null modem cable back then.  :)
-- 
Best regards,
Per Lundberg

I've gotten the debian stuff to work, but I have problems compiling
chaos. I upgraded to unstable, so I got gcc 4.1 and it had lots of
complaints about signedness and aliasing. I fixed most of them but got
stuck on some...

Except for this, it appeared to work pretty good. Once I've gotten it
compiling I think the rest should be easy.

// Anders

> > I only had time to try and get Ubuntu installed and apparently there
> > is some problem with 6.06 and VMWare which I've found a fix for but
> > not had time to try yet (I just got back from a trip to Lule=E5). I'll
> > continue with it tomorrow.
>
> That's strange; I am running Ubuntu 6.06 (dapper) with no problems in
> vmWare.  But that's the 64-bit version so perhaps the problem is only
> with the 32-bit one.

It's something with the kernel, you need to change from the default
386 to the 686 one. Since you use the 64-bit one you're not affected.


> Anyway, use my disk image that I mentioned in my last email instead

I will...

// Anders

ti 2006-07-18 klockan 00:11 +0200 skrev Anders Öhrt:
> I only had time to try and get Ubuntu installed and apparently there
> is some problem with 6.06 and VMWare which I've found a fix for but
> not had time to try yet (I just got back from a trip to Luleå). I'll
> continue with it tomorrow.

That's strange; I am running Ubuntu 6.06 (dapper) with no problems in
vmWare.  But that's the 64-bit version so perhaps the problem is only
with the 32-bit one.

Anyway, use my disk image that I mentioned in my last email instead and
you won't have to bother.  It is only 1 gig big so there's not so much
space available but you could easily add in another disk image of your
own and mount it as /usr or similar to be able to install X Windows etc,
if desirable.
-- 
Best regards,
Per Lundberg

> > Wow, that's a lot to setup. I'll give it a try, but from previous
> > experience, I'll probably get stuck on the way...
>
> Was that how it ended up?

I only had time to try and get Ubuntu installed and apparently there
is some problem with 6.06 and VMWare which I've found a fix for but
not had time to try yet (I just got back from a trip to Lule=E5). I'll
continue with it tomorrow.

// Anders

må 2006-07-17 klockan 14:52 +0200 skrev Per Lundberg:
> If/when I'll have the time to do this, I'll also post a floppy image
> that can be used to netboot with GRUB and vmware.

It has been done.  The vmWare/GRUB compatibility seems a bit less than
100% but perhaps it would be improved if direct PCI access is used
instead of BIOS; I'll have to try recompile GRUB with that flag and see
if it makes a difference.  For now, what you have to do is when it
"hangs" in the boot, press Ctrl-C and then press Enter twice to get back
to the menu and reactivate the same option.  It loads perfectly then,
but for some reason, a page fault :-) is caused which causes other page
faults so it (chaos) doesn't seem to work 100% in vmWare.  I think we
have had problems with chaos in vmWare before as well actually, but it
works well in Bochs so it shouldn't be unsolvable by any means.

If you or anyone else wanna try it out anyway, here are the URLs.

http://chaos.lundcom.se/download/vmware.img - chaos image for netbooting
in vmWare.  Set up a vmWare guest (in vmWare) system with this as A: and
networking enabled using vmnet8 (NAT networking).

http://chaos.lundcom.se/download/Debian.vmdk
http://chaos.lundcom.se/download/Debian-flat.vmdk.bz2 - vmWare disk for
my small Debian installation.  Obviously, the bz2 one needs to be
decompressed.  Use this as the C: with SCSI (BusLogic as the controller
type) and you should be able to boot Debian with all the neccessary
stuff.  I realize now that I didn't actually install xemacs or any other
editor but it should be just an apt-get away...  Oh, use the same
network config for this one (i.e. vmnet8, NAT networking).  That way,
the chaos guest can access the Debian one and TFTP boot from it.

Logon as root/chaos or chaos/chaos.  The chaos user has the chaos
sources in his home directory so the best way is probably to use this
one, either renaming it to something more personal or keeping it like
that.

Any comments?  :-)  Of course, you or anyone else are very welcome to
improve on this, I have only done a quick and dirty work to get it
running.  And obviously, if someone finds and fixes the reason for the
page faults, be my guest...

Now, back to the homework.  :)
-- 
Best regards,
Per Lundberg

on 2006-07-12 klockan 19:56 +0200 skrev Anders Öhrt:
> > get vmWare Server from www.vmware.com.  It is a free (yes, as
> > in $0 cash, not freedom unfortunately)
> Cool, I didn't know that.

It is fairly new, the "sharp" version was released between these two
emails (yours and mine) actually.

> > [...]
> Wow, that's a lot to setup. I'll give it a try, but from previous
> experience, I'll probably get stuck on the way...

Was that how it ended up?  I realized after writing that Bochs is not
neccessary; you should be able to run both the development VM as well as
the testing VM in vmWare.  I have started a minimal installation of
Debian in vmWare now which I'll post a URL for to the list when it's
ready.  The idea is to provide a minimal development environment with no
X windows but with xemacs, gcc etc...  and perhaps a README with info on
how to check out from CVS either using a "real" user account or
anonymous (for people without CVS accounts).

If/when I'll have the time to do this, I'll also post a floppy image
that can be used to netboot with GRUB and vmware.

Now, of course it would be great if you've already done all of
this...  ;-)
-- 
Best regards,
Per Lundberg

> Hello Anders, great to see you're interested!  :)

Well, I need something to relieve the boredom now that I'm on vacation
and "not allowed" to work. ;)


> get vmWare Server from www.vmware.com.  It is a free (yes, as
> in $0 cash, not freedom unfortunately)

Cool, I didn't know that.


> [...]

Wow, that's a lot to setup. I'll give it a try, but from previous
experience, I'll probably get stuck on the way...

I'll keep you posted on my progress.

// Anders

on 2006-07-12 klockan 00:20 +0200 skrev Anders Öhrt:

Hello Anders, great to see you're interested!  :)

> > (If there are other people out there willing to give some time as well,
> > just let me know and I can arrange CVS access for you)
> I might give it a go, but I'm in 100% windows environment right now,
> and it was a really long time since I did any tinkering with other
> OS'es, boot loaders and such... What do I need to compile and test
> chaos at the moment?

* GNU cons (a "make" replacement written in Perl, which is superior in
some aspects).
* A working binutils/gcc toolchain with support for creating ELFs.
* Loopback mounting of the initial ramdisk (it is loopback mounted and
then the boot program is copied into it).  You could get around this by
not using the initial ramdisk and instead working on modules or similar.
This is probably the bit that would be hardest to get working in a
Cygwin environment because the initial ramdisk uses the Minix
filesystem... which is unsupported on Windows AFAIK.

You're probably easiest away with an Ubuntu or Debian installation
(Ubuntu is a bit better than Debian IMO, I am moving over towards it).
I suggest: get vmWare Server from www.vmware.com.  It is a free (yes, as
in $0 cash, not freedom unfortunately) version of vmWare that you can
use to run Ubuntu or whatever on top of Windows.  If you have 512 megs
of RAM or more, it should work pretty OK but don't let the virtual
machine get too much memory; 192 megs is probably max.  You can probably
give it less if you skip X and just run in textmode (for the ultimate
geeky/hacker feeling.  ;-)

You can then setup rarpd (RARP daemon, better than bootp or dhcp because
it is less likely to conflict with other DHCP servers on your network)
and tftp in this Ubuntu install and run Bochs in Windows to netboot from
this Ubuntu server.  :)  Use the bridged networking approach in the
setup of vmware so you can access the Internet from the virtual machine
as well, and access the CVS server.

What I think I will do, but I cannot promise when, is to create an
Ubuntu disk image setup this way, so that the only thing needed to get
it running is:

1) setup vmware.
2) boot the virtual machine
3) create your own user
4) download the sources from CVS
5) compile

But as I said, I cannot promise when I can take time for doing this so
if you're having your summer vacation or similar with a lot of free time
available, you might be better of doing it yourself...  :)  It is not
too hard to do anyway and just post to the list if you run into
problems.

Oh, and for the Bochs setup: use this floppy image.
http://chaos.lundcom.se/download/grub-netboot-floppy.image which I might
have mentioned on the mailing list before.  It is preconfigured with a
GRUB config that works with the ne2k support in bochs and tries to load
the menu over the network (it presumes it to be
in /tftpboot/chaos/config/grub/menu).

More info on the netbooting in bochs approach:
http://chaos.lundcom.se/index.pl?chaos/Development_environment
-- 
Best regards,
Per Lundberg

> (If there are other people out there willing to give some time as well,
> just let me know and I can arrange CVS access for you)

I might give it a go, but I'm in 100% windows environment right now,
and it was a really long time since I did any tinkering with other
OS'es, boot loaders and such... What do I need to compile and test
chaos at the moment? I seem to remember messing with GCC under cygwin
to get it to compile ELF and not succeding... To boot, Bochs + Grub?

// Anders

Ladies and gentlemen,

I am highly honored to be able to tell you ;-) that I have now posted a
new snapshot on SourceForge, for the first time in 4½ years...  I guess
you can say that the project has been a bit dead since then. (See later
in my email for the full changelog of everything that has been changed;
the biggest new feature since I started looking into the code again is
probably the .bss support I implemented in the stormG3 elf loader)

The snapshot is mostly just stating to the world that the project is
alive; not so much is new but the new thing is that at least 1 person
has been committing code again.

(If there are other people out there willing to give some time as well,
just let me know and I can arrange CVS access for you)

So what about the talks about a chaos vision and so forth?  Yes, those
things are important as well but it must not hinder people from hacking
freely things that they are interested in.  I mean, we don't need to sit
down and have a meeting about whether we should have a
system_call_exit() or not...  

Things like those are fairly obvious, at least to me.  The things that
need to be discussed and debated are on a high-level plane and don't so
much affect the underlying structure and the inner workings of the
kernel etc.  With a slight risk of being a bit technocratic here: Those
questions (about whether to go micro-, macro- or mellon kernel and
similar) are perhaps best settled by people actively doing the code
because those will at least on a short-term be most affected by the
decision, and because they are implementing it they are likely to
understand the pros and cons of different ways of doing it.

Just my $0.02 of course.  :-)

And now for what you have all been waiting for... Here comes the
changelog.  ;-)


2006-07-11 22:19  Per Lundberg <pe...@gm...>

	* README: Updated a little

2006-07-11 22:14  Per Lundberg <pe...@gm...>

	* programs/boot/boot.c: Stupid non-compiling commit, fixed. :-)

2006-07-11 22:12  Per Lundberg <pe...@gm...>

	* programs/boot/boot.c: [no log message]

2006-07-11 22:02  Per Lundberg <pe...@gm...>

	* snapshot.sh: Added script for creating snapshots

2006-07-11 15:10  Per Lundberg <pe...@gm...>

	* ChangeLog, usermap: 2 new files

2006-07-11 14:52  Per Lundberg <pe...@gm...>

	* libraries/common/memory/memory.h: Added file

2006-05-25 21:08  Per Lundberg <pe...@gm...>

	* programs/Construct: Added library path

2006-05-25 21:06  Per Lundberg <pe...@gm...>

	* modules/system/boot/boot.c: Made the code go through and fixed a
	  construct in regard to strict aliasing, added a comment

2006-05-25 21:04  Per Lundberg <pe...@gm...>

	* modules/Construct: Turned on -Werror again

2006-05-25 21:03  Per Lundberg <pe...@gm...>

	* modules/: sound/soundblaster/soundblaster.c,
	  filesystem/vfs/assign.c: Valid code with strict aliasing

2006-05-25 21:02  Per Lundberg <pe...@gm...>

	* modules/system/console/console.c: Added Ctrl-Tab switching as
	  well (good for Bochs)

2006-05-22 16:26  Per Lundberg <pe...@gm...>

	* programs/install.sh: Added FIXME

2006-05-20 23:39  Per Lundberg <pe...@gm...>

	* modules/filesystem/vfs/vfs.c,
	  stormG3/include/storm/ia32/memory.h, stormG3/source/ia32/elf.c:
	  The problem was because of .bss sections being more or less
	  unimplemented which is now solved... :)

2006-05-18 23:27  Per Lundberg <pe...@gm...>

	* modules/filesystem/vfs/vfs.c: Localized the problem with the boot
	  module not being able to come through properly. It seems that the
	  ELF loading of modules is not properly functioning when it comes
	  to the .data section; I will look more into this at some other
	  time. But default values of variables are therefore not always
	  working 100% right now which could cause any kind of weird
	  bugs...

2006-05-13 22:24  Per Lundberg <pe...@gm...>

	* compile.sh: Fixed the code to be able to use the INSTALL_PREFIX
	  properly.

2006-05-13 22:23  Per Lundberg <pe...@gm...>

	* libraries/common/unicode/unicode.c: Commented out code because of
	  gcc warnings

2006-05-13 22:22  Per Lundberg <pe...@gm...>

	* programs/Construct: Commented out modplay because it doesn't
	  compile right now

2006-05-13 22:21  Per Lundberg <pe...@gm...>

	* modules/Construct: Disabled -Werror because of the strict
	  aliasing stuff

2006-05-13 22:21  Per Lundberg <pe...@gm...>

	* modules/system/boot/boot.c: Tried to localize the reason for a
	  segfault in the init process. Haven't located it yet though, it
	  is after the while (TRUE) I've added at least

2006-05-13 22:20  Per Lundberg <pe...@gm...>

	* modules/system/keyboard/: keyboard.c, keyboard.h: Fix to make it
	  compile

2006-05-13 22:17  Per Lundberg <pe...@gm...>

	* modules/filesystem/vfs/: assign.c, mount.c, open.c: Some
	  workarounds to the strict aliasing rule

2006-05-13 22:05  Per Lundberg <pe...@gm...>

	* libraries/module/Conscript: Added vfs and video since they now
	  compile

2006-05-13 22:05  Per Lundberg <pe...@gm...>

	* libraries/module/: vfs/Conscript, video/Conscript: Fixes to get
	  it compiling

2002-12-20 21:51  Per Lundberg <pe...@gm...>

	* libraries/module/log/return_values.h,
	  libraries/module/video/defines.h,
	  libraries/program/console/console.h,
	  libraries/program/console/functions.h,
	  libraries/program/console/return_values.h,
	  libraries/program/console/types.h,
	  libraries/program/kernel/functions.h,
	  libraries/program/kernel/return_values.h,
	  libraries/program/kernel/types.h,
	  libraries/program/log/return_values.h,
	  libraries/program/memory/defines.h,
	  libraries/program/memory/functions.h,
	  libraries/program/memory/inlines.h,
	  libraries/program/memory/memory.h,
	  libraries/program/memory/return_values.h,
	  libraries/program/memory/types.h,
	  libraries/program/sound/return_values.h, modules/Construct,
	  programs/boot/boot.c, programs/modplay/modplay.c: Last checkin
	  from me this year, so to all of you: happy new year! (This is
	  just a big checkin to get the repository in touch with my working
	  version; no big changes have been made. If someone wants to fix
	  the mod player, be my guest)

2002-12-20 21:49  Per Lundberg <pe...@gm...>

	* protocol/video.h: Last checkin this year: happy new year!

2002-12-17 23:05  Per Lundberg <pe...@gm...>

	* programs/modplay/modplay.c: Added chaos header.

2002-12-17 23:03  Per Lundberg <pe...@gm...>

	* programs/modplay/modplay.h: New file.

2002-12-17 23:03  Per Lundberg <pe...@gm...>

	* programs/Construct: Change the include path properly.  Build
	  modplay.

2002-12-17 23:02  Per Lundberg <pe...@gm...>

	* programs/modplay/modplay.c: Some minor patches.

2002-12-17 22:54  Per Lundberg <pe...@gm...>

	* libraries/program/: console/console.h, console/defines.h,
	  console/functions.h, console/return_values.h, console/types.h,
	  kernel/functions.h, kernel/return_values.h, kernel/types.h,
	  sound/functions.h, sound/return_values.h, sound/types.h: New
	  files.

2002-12-17 22:54  Per Lundberg <pe...@gm...>

	* libraries/program/log/return_values.h: Added a missing semicolon.

2002-12-17 22:48  Per Lundberg <pe...@gm...>

	* libraries/program/log/functions.h: New file.

2002-12-17 22:48  Per Lundberg <pe...@gm...>

	* libraries/program/log/: log.c, log.h, return_values.h, types.h:
	  New files.

2002-12-16 22:58  Per Lundberg <pe...@gm...>

	* libraries/Construct: Added common/Conscript to the list of
	  sub-scripts.

2002-12-16 22:57  Per Lundberg <pe...@gm...>

	* libraries/common/Conscript: Removed string/Conscript
	  (non-existent).

2002-12-16 22:57  Per Lundberg <pe...@gm...>

	* libraries/common/unicode/Conscript: Updated to comply with new
	  location (bug #55).

2002-12-16 22:56  Per Lundberg <pe...@gm...>

	* libraries/common/Conscript: New file.

2002-12-16 22:36  Per Lundberg <pe...@gm...>

	* libraries/module/video/: Conscript, defines.h, functions.h,
	  return_values.h, types.h, video.c, video.h: Moved the video
	  library to module/video.

2002-12-14 12:22  Johannes Lundberg <joh...@gm...>

	* modules/Construct: Added $build_directory/../libraries/common to
	  the include-path.

2002-12-14 08:25  Per Lundberg <pe...@gm...>

	* libraries/common/unicode/: Conscript, functions.h,
	  return_values.h, types.h, unicode.c, unicode.h: Moved the unicode
	  library to common/unicode (bug #55).

2002-12-14 08:23  Per Lundberg <pe...@gm...>

	* libraries/common/string/: functions.h, inlines.h,
	  return_values.h, string.h: Moved the string library to
	  common/string (bug #55).

2002-12-14 08:21  Per Lundberg <pe...@gm...>

	* libraries/module/skeleton/: defines.h, functions.h,
	  return_values.h, skeleton.c, skeleton.h, types.h: Moved the
	  skeleton library to module/skeleton (bug #55).

2002-12-12 23:10  Per Lundberg <pe...@gm...>

	* programs/modplay/modplay.c: Started working on making the mod
	  player work in the new architecture; quite a bit of work remains
	  (it's not compiling yet).

2002-12-12 23:10  Per Lundberg <pe...@gm...>

	* modules/sound/soundblaster/Conscript: Link with -llog and
	  -lsound.

2002-12-12 23:09  Per Lundberg <pe...@gm...>

	* modules/sound/soundblaster/soundblaster.c: Register a sound
	  service.

2002-12-12 23:08  Per Lundberg <pe...@gm...>

	* protocol/keyboard.h: Moved from the libraries, since it's
	  actually a protocol definition that will be used both from
	  modules and from programs like cluido.

2002-12-12 23:08  Per Lundberg <pe...@gm...>

	* libraries/program/: kernel/kernel.h, sound/sound.h: New headers,
	  used by the non-existing program libraries.

2002-12-12 23:06  Per Lundberg <pe...@gm...>

	* libraries/module/: Conscript, block/Conscript,
	  checksum/Conscript, console/Conscript, console/console.h,
	  exec/Conscript, filesystem/Conscript, log/Conscript: Changed the
	  conscripts to work at the new location of the libraries.

2002-12-12 23:06  Per Lundberg <pe...@gm...>

	* libraries/module/sound/: functions.h, return_values.h, sound.c,
	  sound.h, types.h: New files.

2002-12-12 23:05  Per Lundberg <pe...@gm...>

	* libraries/module/sound/Conscript: New file, used for compiling
	  the sound library.

2002-12-12 23:05  Per Lundberg <pe...@gm...>

	* libraries/module/vfs/: Conscript, defines.h, functions.h,
	  return_values.h, types.h, vfs.c, vfs.h: Moved from vfs to
	  module/vfs (bug #55).

2002-12-12 23:03  Per Lundberg <pe...@gm...>

	* libraries/module/memory/memory.h: Moved from memory to
	  module/memory (bug #55).

2002-12-12 22:59  Per Lundberg <pe...@gm...>

	* libraries/module/log/: Conscript, defines.h, functions.h, log.c,
	  log.h, return_values.h, types.h: Moved the log library to
	  module/log (bug #55).

2002-12-12 22:57  Per Lundberg <pe...@gm...>

	* libraries/module/filesystem/: Conscript, defines.h, filesystem.c,
	  filesystem.h, functions.h, return_values.h, types.h: Moved the
	  filesystem library to the module subdirectory (bug #55).

2002-12-12 22:53  Per Lundberg <pe...@gm...>

	* libraries/module/exec/: Conscript, exec.c, exec.h, functions.h,
	  return_values.h, types.h: Moved the exec library to module/exec
	  (bug #55).

2002-12-12 22:52  Per Lundberg <pe...@gm...>

	* libraries/module/console/: Conscript, console.c, console.h,
	  defines.h, functions.h, return_values.h, types.h: Moved the
	  console library to module/console (bug #55).

2002-12-12 22:51  Per Lundberg <pe...@gm...>

	* libraries/module/checksum/: Conscript, checksum.h, crc32.c,
	  defines.h, functions.h, md5.c, types.h: Moved the checksum
	  library to module/checksum (bug #55).

2002-12-12 22:49  Per Lundberg <pe...@gm...>

	* libraries/module/block/: Conscript, block.c, block.h,
	  functions.h, return_values.h, types.h: Moved the block library to
	  module/block (bug #55).

2002-12-12 22:48  Per Lundberg <pe...@gm...>

	* libraries/module/ethernet/: defines.h, ethernet.h, functions.h,
	  return_values.h, types.h: Moved ethernet to module/ethernet (bug
	  #55).

2002-12-08 22:54  Per Lundberg <pe...@gm...>

	* libraries/Construct: Made it compile with the sound and kernel
	  library (everything else still needs to be moved though).

2002-12-08 22:26  Per Lundberg <pe...@gm...>

	* libraries/: module/Conscript, program/Conscript: New file

2002-12-08 22:18  Per Lundberg <pe...@gm...>

	* libraries/program/README: README

2002-12-08 21:09  Per Lundberg <pe...@gm...>

	* libraries/module/kernel/: Conscript, functions.h, kernel.c,
	  kernel.h, return_values.h, types.h: Moved from libraries/kernel.

2002-12-08 21:04  Per Lundberg <pe...@gm...>

	* protocol/sound.h: New file.

2002-12-08 21:04  Per Lundberg <pe...@gm...>

	* modules/Construct: Add libraries/module to the include path.

2002-12-08 20:59  Per Lundberg <pe...@gm...>

	* libraries/Construct: Added module to the include path.

2002-12-08 20:39  Per Lundberg <pe...@gm...>

	* libraries/module/README: New file