Subject: Re: HEADS UP: migration to fully dynamic linked "base" system
To: Greywolf <>
From: Johnny Billquist <>
List: current-users
Date: 08/26/2002 21:02:06
On Mon, 26 Aug 2002, Greywolf wrote:

> On Mon, 26 Aug 2002, Jukka Marin wrote:
> # Err.  I apologize for opening this can of worms.  I should have known
> # better, having been reading these mailing lists for years. ;-)
> Meh, a good discussion is a healthy thing.  I am seeing things presented
> here that I cannot refute in any meaningful way, either due to lack of
> technical savvy or just because the points are good.

I actually think this is a good discussion, but I don't know if
current-users is the right place?
It started with a little mud slinging (for which I apologize), but I think
we've made some progress from there.

> Now I need to go look at dlopen().

For me, the issue seems clear. (or parts of it) should be linked into static binaries.
dlopen() should be available to statically linked binaries. If
dlopen() fails (because the file couldn't be found), things should go
ahead in a sensible way in the programs in /bin and /sbin.

For people who normally don't use/abuse english, it should still be an
acceptable behaviour that when the system is broken, or you are running in
single-user mode with no file systems mounted, you'll have english and
nothing else (that's the way I assume /crashandburn behave as well, by the

> I seem to remember, though, that even the old 4.3 distribution of BSD
> made some reference about ld being able to link an object file into an
> already running executable...or is this something so totally different
> that the sky might as well be purple with yellow spots?

I suspect you are thinking of something totally different, or that you are
late for that scheduled brain maintenance. ;-)


Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email:           ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol