plans and wishes

Greg Schafer gschafer at zip.com.au
Tue Jan 13 15:11:06 PST 2004


On Tue, Jan 13, 2004 at 11:15:25PM +0100, Alex Groenewoud wrote:
> Here's most of the changes I have in mind for the book, and things that
> I'd like to see happen sooner or later, in no particular order.  Only
> some of these points plan to do before branching for a 5.1 release.

Alex, I'd suggest extreme caution here. You don't want to "upset the apple
cart" too much, as you should have already learned. Most of your suggestions
are most definitely not suitable for 5.1 IMHO.

Not only that, you've hardly provided any rationale for any of it.

> a) Move the sections on $LFS, SBUS and test suites from chapter 2 to the
> beginning of chapter 5.  This info is now given at the beginning of the
> book but needed only three chapters later.  It seems better to put it
> just before it is becomes relevant.  (In reality it isn't important
> information at all: it won't hurt you one bit skipping it alltogether,
> mostly because after the creation of the /tools dir the LFS variable
> isn't used any more in the whole of chapter 5.)

$LFS is currently used in Ch 3 in "preparing a new partition". Therefore no
dice. SBUS and test suites I'd agree with.

> b) Move the section on Asking for help to an appendix, and refer to it 
> from somewhere at the beginnig of chapter 5.

I'd object to this, unless you can provide rationale.

> c) The beginning of this chapter 5 will now look so crowded that the
> preparational stuff warrants a chapter on its own.  Essentially chapters
> 2, 3 and 4 do a rotate-left, with a carry-in from 5. 

Chapters 5 are 6 are hardcoded in all of our brains. Please avoid changing
the meanings of these chapters at all costs :-)

> d) Move Conventions, Resources and Acknowledgements to the Preface.  
> Move Book version to the beginning of the Changelog, and all of that to 
> an appendix.  Combine 'How things' with 'Structure' in the Preface.  
> Chapter 1 is gone, all numbers drop.

Sounds reasonable in theory.

> e) Bring the preparational stuff of the late chapter 6 together in its
> own preceeding chapter.

Very unclear. You need to provide more detail and rationale.

> f) Move the unpacking of the bootscripts to chapter 7, so it sits all 
> together and one can skip it in one go, if necessary.

Sounds reasonable in theory.

> g) Move the creation of fstab somewhere else, as it has nothing to do 
> with the kernel.

It's part of "making the system bootable" and has everything to do with
that. I object.

> h) Move the stripping of the binaries from chapter 9 to the end of 
> chapter 6.

Sounds reasonable in theory. But doesn't really "feel" right to me.

> i) Alphabethize the order of the packages as far as possible, mostly 
> requiring only an extra /bin/gawk and /bin/sed symlink.

Yes, the bugzilla is open. Why haven't you entered your research results
there yet? My gsbuild script will help here.

> j) Change from using $LFS to /lfs.  If we can make a symlink on the
> host's root dir, we can also make a mount point.

I object. We already make a mount point i.e. /mnt/lfs. What would this change
buy us? Please provide rationale.



More information about the lfs-dev mailing list