Re: [Ltib] LFS version

[Stuart Hughes]

Hi Jurgen,

Some points:

First although it's named lfs-5.1, you should see that as a starting
point (or origin), many package have been updated.

Secondly, there are many packages that are out of date, and it could be
nice to upgrade.  However this is a big undertaking unless it can be
automated, there are 400 packages and even if it only took 2 hours per
package (and that's way too low), it's a lot of effort.  Currently there
is no one sponsoring any work for the community for LTIB (except some by
Zee2 who I have an association with).

Finally, and more philosophically in many ways LTIB in terms of its
package set has grown way beyond where I originally designed it.  The
idea was not to make a distribution, but a tool that builds enough
packages to allow you to bootstrap a new platform and from there if you
want build a full distribution.  There are plenty of other full
distributions that have enough resources to do a proper job of
maintenance (Debian).  The niche for LTIB is really that it can scale
down to a tiny package set (around 5) which is something few others can
offer.  In this context having old packages doesn't matter so much.

If there were a sponsor, here are the things that could be done:

# ARM QEMU target emulator support (for graphics)
# Web interface to build/add-packages/edit/release
# Documentation/website refresh
# Distribution update - direct from upstream or Debian sources
# Distribution rationalisation - relevant embedded packages
# Patch clearance and release interlock tool
# Use of QEMU to provide pseudo native builds - cross compile enhancement
# Re-factoring BSPs to be self-contained
# Dynamic host support package dependency installs
# Common git based kernel builds with per project branch management
# Better multi-core support
# Pre-built binary rpm package feeds
# Investigate supporting non-Linux real-time kernels (RTEMS for example)

Regards, Stuart


On 03/03/11 12:46, Jürgen Lambrecht wrote:
> Hello,
> 
> why does ltib stay with LFS v5.1, and not upgrade to 6.7?
> Is there a reason to?
> And if I would want to upgrade, how?
> 
> Thanks,
> Jürgen
>