[blfs-support] Make 3.82 Dev86src triple bug fixes patch

Bruce Dubbs bruce.dubbs at gmail.com
Fri Feb 17 08:58:51 PST 2012


Michael Shell wrote:
> On Thu, 16 Feb 2012 16:24:34 -0600
> Bruce Dubbs <bruce.dubbs at gmail.com> wrote:
> 
>> Personally, I don't think LFS has an obligation to support 16 year
>> old code that is needed for a boot loader that we don't support.
> 
> I understand that. However, as far as lilo goes, they seem to have
> restarted development in 2010 (and I also wonder why this is - what is
> not liked about current grub?):

They don't like it becuaue most distros use GRUB2 is a complex way. 
Actually, if they'd get away from the automatic rewriting of grub.cfg, 
it's not that hard to understand.

> and I wanted to see what they've done. 

AFAIK, you still have to have the program recalculate locations after 
any change.  For instance, I've made one line changes to the kernel 
config and then overwritten the kernel in /boot.  In lilo, it doesn't 
automatically pick that up.  GRUB does.  Also, IIRC, lilo offers no boot 
time interface to fix errors.

> Heck they even added a graphical Tux boot screen. 

I've always thought that a waste of programmer time.

> It is my understanding that Make 3.82 did tighten the rules with regard
> to the use of tabs rather than spaces which causes some existing
> Makefiles to break, but that is not a bug in Make, but rather its
> documented behavior and so it is the affected Makefiles that need
> correction. Eric S. Raymond even once wrote in one of his books that he
> wished the original author of make had never relied on tabs in the
> first place.

No kidding.  Relying on tabs was 'less than thoughtful'.  If you can't 
tell a tab from a string of spaces by just looking at it, it's a 
mistake.  I wish the tab key would be banned from the keyboard.  Tabs in 
program code suck too.

   -- Bruce



More information about the blfs-support mailing list