Am Donnerstag, den 09.04.2020, 20:28 +0200 schrieb Tim Tassonis via
blfs-dev:
> 
> <snip>
> > > I'm planning to do a full rebuild this year, if time permits, shall I 
> > > refrain from updating stuff until then?
> > 
> > Yes, I think you probably should do that.  I do not update my day-to-day 
> > system that oftern (maybe once a year), but I think most editors have 
> > separate development systems that have the latest packages.
> 
> Ok. I just bought another 16 GB of RAM for my development server, after 
> putting it in I should be able to start a new full build while leaving 
> my current system up and running.

Just as a side-note:  Sometimes, packages shows strange behaviour when
compiled on new tool chain. That happens sometimes when gcc and/or
glibc has been updated. Than, a newer gcc might complain about
constructs while older gccs does not or glibc hides definitions which
were visible previously so we encounter that a package might be needed
to be patched somehow. This is one of the reasons why the toolchain
should be kind of recent.

IMHO building on VMs is sufficient. Just prepare a VM (using
VirtualBox (i'm a big fan of that software) which also can be built
using LFS!) and you can try stuff just as you like. Create snapshots,
test something, rollback, clone it and throw it away and so on.
I use VMs (i686,x86_64) with three disks:
sda     /boot, /home, swap      (~32 GB)
sdb     System 1                (~20 GB)
sdc     System 2                (~20 GB)
So i can build Sys2 from Sys1, and than build Sys1 from Sys2 and so
on...
Using same layout for machines (again i686 and x86_64) installed with
the "LFS releases":
sda     /boot, /home, swap
sdb     lfs-9.0
sdc     lfs-9.1
Disk sizing quite similar to the devel-VMs. Next time, when 9.2 is
released, I'll install 9.2 on sdb using lfs-9.1.

If there is really something which needs the hardware directly, its
easy to tar up a built system and extract it on a physical partition
and boot it there.

Since the SBU is relative by nature, it is not that relevant that a VM
might be a tiny bit slower than the physical machine. 

Ok, backup is somewhat time consuming when transferring the vdi to
NAS, but thats something for which night+fcron is good for.

Just my 2ct ... and please don't stop taking tickets (on a recent
toolchain ;-) )!

--
Thomas

-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to