Subject: Re: HEADS UP: migration to fully dynamic linked "base" system
To: Johnny Billquist <firstname.lastname@example.org>
From: Greg A. Woods <email@example.com>
Date: 08/27/2002 22:40:46
[ On Tuesday, August 27, 2002 at 02:38:56 (+0200), Johnny Billquist wrote: ]
> Subject: Re: HEADS UP: migration to fully dynamic linked "base" system
> My point is basically that Unix as such don't know anything about shared
No, but it doesn't mean it can't be taught about them at a low level either.
Remember AT&T UNIX System V shared libraries. They avoid many of the
problems discussed so far in this thread by making the situations that
cause them against the rules -- SysV shared libraries are specially
coded, by design.
SysV shared libraries are inherently more secure too since IIRC the
kernel does the loading and linking, not some user-replacable user-land
program (though there's still the executable storage space issue of course).
SysV shared libraries also actually make programs quicker, not slower,
or at least I recall they did for many things on my trusty old 3B2
(where a few CPU cycles on such a slow machine for resolving the jump
tables was far quicker than the extra disk operations that might be
necessary to read in the equivalent static object code for any program
that used any significant portion of the library in question -- leaving
only the issue of total memory use vs. things like the text sticky bit
to work out in the equation).
Greg A. Woods
+1 416 218-0098; <firstname.lastname@example.org>; <email@example.com>
Planix, Inc. <firstname.lastname@example.org>; VE3TCP; Secrets of the Weird <email@example.com>