Subject: Re: PAM
To: None <>
From: Dan Melomedman <>
List: current-users
Date: 09/26/2002 11:37:04
Jim Wise wrote:
> Hash: SHA1
> On Thu, 26 Sep 2002, Dan Melomedman wrote:
> >think it's being used where it shouldn't also. Statically-linked
> >binaries will load faster, so they're more suitable for busy fork/exec
> >servers. Linking statically when developing helps to show a programmer
> And since exec-chaining would turn _all_ servers into `busy
> fork-and-exec' servers, I see why you think this :-P

So you must be from the camp which thinks fork/exec are somehow
extraordianrily inherently heavier than other syscalls. You'd be
surprised if you actually measured the fork/exec latency as compared
to some other syscalls which your servers make rather than badmouth
this model, and bring personal insults into this.

Just because static linking results in a quicker process start doesn't
mean this is somehow related to the PAM discussion. I merely pointed out
where static linking makes sense. I also wrote dynamic linking has its
use too.