Re: [lfs-dev] Systemd branch created

2013-02-14 Thread cybertao
I just finished building this, ironed out my mistakes (journald barfing
because there's no machine-id is a good one!), and couldn't be more pleased
with myself.  And immensely grateful for all the work that went into this.

Thanks.


On 6 February 2013 10:34, Matt Burgess  wrote:

> Hi all,
>
> I've just created a systemd branch for LFS.  The intention behind this
> is to provide a version of the book with systemd fully integrated,
> without the need for post-lfs changes.  This in no way means that
> systemd will make it into the book at any point, but it should help
> folks out that want to make use of systemd and also help LFS and BLFS
> editors assess the full impact that any potential migration to systemd
> might have.
>
> The book can be checked out from
> svn://svn.linuxfromscratch.org/LFS/branches/systemd/BOOK.  An online
> copy is available from http://www.linuxfromscratch.org/lfs/view/systemd/
> If you wish to raise any bugs/defects/issues that are specific to this
> version, there is a 'systemd' version identifier in Trac (if your bug is
> also present in 'trunk' please raise it against the 'SVN' version as
> usual; I'll pick up any fixes from trunk and merge them into the systemd
> branch as and when necessary).
>
> At the moment, trunk and the systemd branch are identical, save for the
> version identifiers.  I intend to start work on the branch very shortly.
>
> Regards,
>
> Matt.
>
> --
> http://linuxfromscratch.org/mailman/listinfo/lfs-dev
> FAQ: http://www.linuxfromscratch.org/faq/
> Unsubscribe: See the above information page
>
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] Systemd branch created

2013-02-18 Thread cybertao
Yeah, I think there were a couple of things out of place or missing.  It's
been a while since I compiled LFS so was sort of fuddling through it
anyway.  The post-LFS hint the and how-to someone posted in the blfs
mailing list helped a lot.
I'm about to try it again, this time in 32-bit.  I'll write a log as I do
and see if I can come up with some constructive feedback.
On Feb 15, 2013 10:28 PM, "Matthew Burgess" 
wrote:

> On Fri, 15 Feb 2013 17:31:05 +1300, cybertao  wrote:
> > I just finished building this, ironed out my mistakes (journald barfing
> > because there's no machine-id is a good one!), and couldn't be more
> > pleased
> > with myself.  And immensely grateful for all the work that went into
> this.
>
> Thanks for the quick feedback!  I wasn't expecting that at all.  I'm still
> working through my own build, and am actually surprised that you got it
> working at all.  I'm sure there's some configuration information and other
> polish missing.
>
> If you have any notes on what you needed to do above what is currently in
> the
> book, I'll put those on my todo list.
>
> Thanks,
>
> Matt.
>
> --
> http://linuxfromscratch.org/mailman/listinfo/lfs-dev
> FAQ: http://www.linuxfromscratch.org/faq/
> Unsubscribe: See the above information page
>
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] Problems setting up a systemd build

2013-02-27 Thread cybertao
I've been using Arch as a reference and configuring the network by writing
a network.service script that simply executes 'ip' commands ('ip addr add
...', 'ip route ...', 'ip link set ...', etc.)
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] 7.3 check-0.9.9

2013-03-06 Thread cybertao
I had the same issue when using 32-bit Ubuntu 12.04.  I got around it at
the time with:
make LDFLAGS=-lthreads


On 6 March 2013 19:21, Billy O'Connor  wrote:

> Matt Burgess  writes:
>
> > On Tue, 2013-03-05 at 14:05 -0600, Bruce Dubbs wrote:
> >> Billy O'Connor wrote:
> >> > I dl'd the book sources and rendered it, then built a new LFS.
> >> > Looks great, as always.  I must have fat-fingered something in chapter
> >> > 5, though, because the configure script for check-0.9.9 didn't pick up
> >> > the pthreads library and I had to add -lpthread to the Makefile by
> hand
> >>
> >> Compare to
> >>
> http://www.linuxfromscratch.org/lfs/build-logs/7.3/core2duo/logs/041-check
> >>
> >> It sounds like a host/configure issue.
> >
> > Sounds more like the issue that Pierre reported on 25/11/12 ("Error with
> > check-0.9.9"), re-reported by him again on 13/12/12 ("Potential
> > pollution of toolchain by the host") and confirmed, in that thread, by
> > Yaacov-Yosev Weiss.
> >
> > So far, we have reports of this failing on Debian Sid (Pierre) and
> > 64-bit Suse LiveCD (Yaacov-Yosev).  Billy, what host are you on, out of
> > interest?
>
> Fresh install of Ubuntu 12.10 32 bit, on a 64 bit core2 duo dell laptop.
> --
> http://linuxfromscratch.org/mailman/listinfo/lfs-dev
> FAQ: http://www.linuxfromscratch.org/faq/
> Unsubscribe: See the above information page
>
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3: Check-0.9.9 compile error fix

2013-03-11 Thread cybertao
I'm interested.  I'd definitely need the help and guidance, but have the
time and enthusiasm.
The previous Live CD was an invaluable tool to me over the years, not just
as a platform for building LFS.  Arch has replaced my need for it, but a
Live CD primed and ready to go increases the accessibility for people new
to LFS and the backend of Linux.
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3: ISO discussion

2013-03-12 Thread cybertao
Okay, I'm on the job.  Since I've been playing with the systemd branch I
need to compile vanilla 32-bit and 64-bit releases, which I'm currently
doing using jhalfs.
I just realised I chose the development branch without much consideration,
when the stable version my be more appropriate.  If only for the
convenience of release cycle guidelines.  I focus on making a bootable CD
and USB image for now using the vanilla-SVN build and replace it with a
release build later if it's more appropriate.
i686 and x86_64 on the same image, like Arch, is a good idea isn't it?  It
becomes complicated to build twice if things such as X are expected to be
included as standard though, and space becomes tight having to have two
builds of everything on the same disk.  The Arch disk splits the filesystem
off into SquashFS (or similar) files for i686, x86_64, and
architecture independent files.  This could also allow for additional
archives (such as BLFS software) to extend the bootdisk, much like Puppy.
To be honest, I'm not prepared to commit to maintaining non-essential
software such as X for now.  My ambition is a terminal environment with the
tools to build LFS, SSH, screen, jhalfs, and build the book.  Any bells and
whistles can be added by anyone through some sort of SquashFS or UnionFS
shenanigans (or perhaps BLFS using jhalfs).  If users feel more comfortable
using a GUI then they should be encouraged to use a virtual machine or SSH
into a dedicated box.
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3: ISO discussion

2013-03-12 Thread cybertao
> Either one would be great, though I think a 7.3 stable ISO would be
> perfect.
> If a dual i686 and x86_64 is doable, then great, otherwise two ISOs for the
> two platforms shouldn't be that much harder then the combo. At least it
> seems that way to me.


Looking at how Arch do it, it shouldn't be too difficult for a combined
image.  They have separate i686 and x86_64 kernels and separate compressed
filesystems.  The bootloader allows the user to select which to load.  This
is a super option for a recovery CD, and anyone who needs space for a
custom image can delete the architecture they don't want.  But I still have
to get my hands dirty and see if it's realistic.
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] 7.3 check-0.9.9

2013-03-12 Thread cybertao
Sorry, my mistake.  I got around it with make LDFLAGS='-lpthread', as Billy
stated in his first post.
The problem also crops up using the 64-bit version of Ubuntu 12.04 as well.

On 7 March 2013 07:38, cybertao  wrote:

> I had the same issue when using 32-bit Ubuntu 12.04.  I got around it at
> the time with:
> make LDFLAGS=-lthreads
>
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3: ISO discussion

2013-03-12 Thread cybertao
I elected to use 'Os' optimisations on my current compilation round.  I
understand it to be safe option and can only assume it help when it comes
to compressing the filesystem into an archive on a limited media choice.
 But I am no expert on the subject and would appreciate suggestions.
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3: ISO discussion

2013-03-16 Thread cybertao
I'd be interested to take a look at your work.

I've been taking my time as a lot of it is pretty new to me.  There is no
escaping the fact an initrd file is needed, for the bootup on multiple
systems by taking advantage of modules, and to mount a root from CD that
can be overlaid with a ram disk.  All that's needed is cpio and the
mkinitramfs files (with tweaking, I'm still playing with them) already
included in BLFS.
http://www.linuxfromscratch.org/blfs/view/svn/postlfs/initramfs.html
There's no need to complicate the process by adding other packages such as
busybox, the initramfs is a minimal collection of tools from the LFS
filesystem.
After making the root system into a volume that can reside on an ISO9660
filesystem (squashfs?), a bootable ISO image can be generated using
grub-rescue.  It's already included in the base system and can also be a
useful tool for recovery disk use.
I've added OpenSSH, wget, and a couple of text editors.  After installing
jhalfs and dependencies I'll be good to go for a minimal x86_64 base system.
The process is all fairly simple and think a chapter could be created in
BLFS to automate building a fresh ISO using the LFS filesystem; a
regenerating customisable LiveCD.

The list of software the old LiveCD on the LFS website is impressive,
though I don't see much of it being useful personally.  Does anyone think
anything is essential?  Braille and speech output are interesting options
that I would like to see included, if only out of curiosity and the
interests of accessibility.
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3 ISO discussion

2013-03-17 Thread cybertao
On 18 March 2013 14:08, Bruce Dubbs  wrote:

> I agree, but at least that is a good baseline that checks the host
> requirements that are currently published.


Focusing on the more modern, I hadn't considered that priority for the
development of the book.  The Host System Requirements and version script
inform the user if they are wasting their time or not on their chosen
distro.  Dependency on a bleeding edge LiveCD is something that should be
actively avoided.  Are there any issues such as newer packages not being
able to compile older ones or anything else significant that could get in
the way of automating the build of a test distribution?
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3 ISO discussion

2013-03-20 Thread cybertao
On 21 March 2013 10:01, William Harrington  wrote:

As far as it recreating ssh keys, I don't think you'd want to publish a
> livecd and distribute it and then everyone on the planet using the same ssh
> key if they have it open to the public vie their network connection. I
> don't know if that would happen often, but I'm sure plenty of people use
> livecd's as servers.
>

Curious.   I installed my OpenSSH on  my project and set a root password.
 I too thought about the security risk, but didn't think it was an issue
for a LiveCD with the purpose of building LFS.  Personally, I'd like to put
the CD in a box without a monitor or keyboard (BIOS dependent!) and ssh
into it.  The box will only be a security risk to the local network it
resides on.  There's no need for a root login apart from convenience,
having a standard user defined with sudo permissions would be exactly the
same risk.
I also weighed up the possibility of the CD being installed to a hardrive,
but LFS is not a full distribution that hold your hand and updates/secures
the system so it's inappropriate to treat the CD as anything but a platform
for building LFS.
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page


Re: [lfs-dev] LFS 7.3 ISO discussion

2013-03-20 Thread cybertao
So, what's up with the linux kernel and overlayed filesystems?  There's no
support.
I notice the Arch CD initramfs creates RAM disks for volatile directories
and copy the contents into from archives into them.  That's no where near
as elegant or flexible.


On 21 March 2013 10:20, cybertao  wrote:

>
>
>
> On 21 March 2013 10:01, William Harrington  wrote:
>
> As far as it recreating ssh keys, I don't think you'd want to publish a
>> livecd and distribute it and then everyone on the planet using the same ssh
>> key if they have it open to the public vie their network connection. I
>> don't know if that would happen often, but I'm sure plenty of people use
>> livecd's as servers.
>>
>
> Curious.   I installed my OpenSSH on  my project and set a root password.
>  I too thought about the security risk, but didn't think it was an issue
> for a LiveCD with the purpose of building LFS.  Personally, I'd like to put
> the CD in a box without a monitor or keyboard (BIOS dependent!) and ssh
> into it.  The box will only be a security risk to the local network it
> resides on.  There's no need for a root login apart from convenience,
> having a standard user defined with sudo permissions would be exactly the
> same risk.
> I also weighed up the possibility of the CD being installed to a hardrive,
> but LFS is not a full distribution that hold your hand and updates/secures
> the system so it's inappropriate to treat the CD as anything but a platform
> for building LFS.
>
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page