After some discussion with Gerard, he has requested I prepare a proposal
to the LFS community concerning the Cross-LFS book.
Up to this point work on Cross-LFS has been done with the idea that,
eventually, its features would be merged into the main LFS book. I would
like at this time to propose that we create a separate project for
Cross-LFS, like ALFS, HLFS and BLFS. There are many reasons for wanting
to do so:
* Keeping the books as separate projects allows users learning the LFS
process to focus on what is needed and avoid confusion with
techniques/patches.
* Toolchain maintenance. Due to the amount of platforms Cross-LFS is
geared towards, it is possible that in the future the toolchain will
need to be changed to include/remove something that is not necessary in
the LFS toolchain. By keeping these two books as separate projects we
can remove the need for dependencies on other projects much like the
LFS/BLFS bootscript compatibility issues.
* The ability of two separate platforms as test platforms (a straight
build and a cross build) will offer more opportunity to find
bugs/problems with packages in more scenarios, thus helping version
stability/usability on packages. The reverse can also be said about
logging faults. Why waste the LFS community's time searching for a bug
in the build process when it is a Cross-LFS bug only? Cross-LFS will
tend to see more build issues due to its build nature; LFS can then
investigate to see what affects them.
Pros:
1 Multi-architecture build - We can build on any platform.
2 Keeping Cross-LFS related bugs and support questions off of the LFS
List, since only a few of us
currently can support it.
3 Book will always be under constant development due to the nature of
it's build scenario and the
multiple architectures it supports. We have a lot of community input.
4 Techniques used in this book could be used to enhance LFS and HLFS
build process.
5 Cross-LFS could potentially see issues before LFS. We also seem to
find all the weird glitches
with programs due to the nature of the book.
6 Multilib support for architectures that can support it.
Cons:
1 Multi-architecture build - The book is designed around building a
complete toolchain. So when
compiling x86 to x86 for example will still build a complete
toolchain..
2 Support. Cross-LFS maintainers will have to support more than one
architecture.
--
------
[EMAIL PROTECTED]
[EMAIL PROTECTED]
LFS User # 2577
Registered Linux User # 299986
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
