Small remarks on testing ch 5&6: gzip, util-linux, udev, readline

Jeremy Utley jeremy at jutley.org
Thu Sep 9 05:46:03 PDT 2004


On Thu, 09 Sep 2004 12:17:36 +0200
Laurens Blankers <lblankers at xs4all.nl> wrote:

> Jeremy Utley wrote:
> 
> > Unless something has changed recently in testing, the filesystems are
> > mounted inside chroot.  The mount commands for these prior to chroot are
> > "fake" mounts to update the hosts /etc/fstab.  The real mounting is done
> > inside chroot.
> 
> No, you are right. I mean the fake mounts. What I was trying to say was:
> 
> Before entering chroots 3 fake mounts are performed. Below these fake 
> mounts is a paragraph explaining that the message 'can't open 
> /etc/fstab: No such file or directory' can be safely ignored. But if I 
> understand correctly the /etc/fstab used by the fake mounts (outside 
> chroot) is that of the host system.
> 
> The next line states:
> 
> This file___/etc/fstab___has not been created yet
> 
> This means that the file /etc/fstab on the _host_ system hasn't been 
> created yet. Well it hasn't and it shouldn't because it is most likely 
> already there, created by the install of the host system.
> 
> So I don't understand what this paragraph is doing here. It would make 
> more sense if it was located after the real mounts in the chroot 
> environment, because then we are talking about the /etc/fstab in the LFS 
> system, which indeed hasn't been created yet.

Ahhh! I see now!  That's a remnant from when we mounted them outside chroot.  Gerard, can you perhaps move this wording to it's proper location later in chapter 6?

> 
> >>chmod 755 /usr/lib/lib{readline,history}.*.5.0
> > 
> > Possibly - can you verify that this updated command will have the EXACT
> > same effect?
> 
> I think it has the same effect, but I can't verify that because I 
> already moved on in building LFS. So it is possible that:
> 
> chmod 755 /usr/lib/*.5.0
> 
> results in other libraries besides libreadline and libhistory having 
> permission 755.
> 
> If prefer not to make changes the command at this point in testing you 
> can leave it as it is. I don't think the current chmod line changes 
> permissions on any other library besides the intended ones at this time, 
> that however may change in the future. So may be this could be included 
> in unstable.

Something to check - I'll watch for this on my next build, and see what happens :)

Thanks for the input!

-J-



More information about the lfs-dev mailing list