lfs65 at cruziero.com
Thu Nov 5 09:15:55 PST 2015
> From: Michael Havens <bmike1 at gmail.com>
> Date: Thu, 5 Nov 2015 09:45:12 -0500
> To: LFS Support List <lfs-support at lists.linuxfromscratch.org>
> Subject: Re: [lfs-support] 5.33
> > Last thing, as a general point, stated again: it really really is normally
> > very useful to have built the book manually at least once, prior to trying
> > scripting it.
> That is exactly what I am going to do. I am copying the text out of the
> scripts and pasting them in.
Copy'n'paste the commands carefully from THE BOOK: NOT from your scripts.
I looked at your earlier pastebinned history ( http://pastebin.com/N5RamvxH
) : not a forensic analysis, so apols if am wrong on the following. Even
allowing for apparent gaps/discontinuities in history lists due to
e.g. changing user, it looks like an extremely garbled sequence. E.g.:
ref: numbered-line '426.' ff (== history-command '440' ff):
* you do glibc, then kernel-headers [wrong order];
* then glibc again (with duplicate 'make' command 473./474. , and three
'make-install' commands 475./479./484. , with a sprinkling of gcc
test-compile stuff );
* then some faffing around with mpfr/gmp [again, out of proper sequence];
* commands run together (e.g. 463./488./497.) with not always apparent
* re 'echo GOOD'/'tee ...log'/'exit $PIPESTATUS': NB that you are of course
changing at least some book commands; you still need to be sure that your
amended commands are valid; else just stick to the book commands literally.
I know you've since gone on to at least one 'fresh' approach: but it needs to
follow the book; and certainly not the likes of what's in that history list.
Bear in mind also:
More information about the lfs-support