Different CVS modules, different (outdated) states

Matthew Burgess matthew at linuxfromscratch.org
Sun Mar 21 09:36:03 PST 2004


On Sun, 21 Mar 2004 09:54:00 -0700
Gerard Beekmans <gerard at linuxfromscratch.org> wrote:

> Hi guys,
> 
> It's time to bring some order to the chaos. Or maybe it's just a
> perceived chaos on my part.

As evidenced by my activity today I think it's clear that something
needs to be done, although I wouldn't go so far as to call it "chaos"
exactly.
 
> We have a lfs-5.1-pre1 branch that's outdated, it needs to be brought
> up to the same level as current HEAD after Ryan tells us what Glibc to
> use in it.

Yep, agreed.  IIRC glibc-cvs-20040116 seems to be the best candidate
but I'll obviously let Ryan confirm/deny this.  I was going to upgrade
gcc to 3.3.3 as well but there's an issue with a failing test.  I'm not
too concerned about it as it appears (from CVS logs, etc.) to be a fault
in the testcase rather than a fault of the compiler itself.  I'd be
happy to just put a note in explaining that you *may* see the test fail,
but not to treat it as seriously as other test failures. Again, I'll
leave this up to Ryan to decide which course of action is best. 

> Matthew, can you explain with as much detail as you can what it will
> take to transform newxml into HEAD and keep using it?

The best advice I can give is to do a `ls -R1 LFS/BOOK` & `ls -R1
LFS/newxml` and diff the two outputs.  Essentially all that is
required is to use that output to determine what needs `cvs add`ing
and `cvs rm`ing.  Once it's in HEAD (or maybe before) then we'll have to
make sure that libxml2, libxslt, docbook-xml & docbook-xsl are all up to
date on belgrarath as well, so it can be processed.

Actually, I remember responding to something a few days ago on lfs-dev
whereby it was generally agreed that newxml should replace HEAD only
when the docbook-DTD is officially released in it's 4.3 version.  At the
moment we're referencing a candidate release (albeit one that very very
closely resembles the finally accepted version).

> We'll figure out what to do with the b5_1 branch later, we might
> upgrade it too to the newxml format if we decide to make the branch
> equal to HEAD and work from there. Basically that means start the
> branch from scratch.

I'd like to see HEAD updated to newxml, then cut a fresh branch as of
whenever that gets done.
 
> If you think newxml can be transformed to HEAD, then I would like to
> get this done today.

I'm prepared to do this, whenever it's agreed that it can go ahead. 
Close coordination will be required with other dev's to ensure they're
not working on anything that needs to be merged prior to the cutover.

> This also means update the /usr/bin/render-lfs-book.sh script and
> replace with whatever you use script wise.

Well, it's largely implemented in a Makefile now.  Calling `make`
will generate the HTML (by default in ~/lfs-book) and `make pdf`
will generate the PDF.

Cheers,

Matt.



More information about the lfs-book mailing list