Re: [Freestandards-ldps] Re: LDPS 0.91 - even closer to release

[tytso@valinux.com]

   From: gk4@us.ibm.com
   Date: Wed, 26 Jul 2000 12:57:25 -0500

(Note, this is a LSB issue, not a LDPS issue, so this is on the wrong
list.  I've redirected replies to the lsb-spec list.)

   To be clear, which of the following is not a goal or is unrealistic:

   1) Consistent threads implementation for applications between GNU/Linux
   distributions for binary compatibility

This is realistic, but it may be very difficult to do at the
specification document level.  It's a lot easier to point to a reference
implementation here.  

   2) GNU/Linux threads applications to be sourcecode portable to other
   non-Linux kernel GNU systems.
   3) GNU/Linux threads applications to be sourcecode portable to other UNIX
   pthread enabled systems

It's probably not realistic to claim that Linux threads applications
will be source code portable to other Unix pthreads systems with out a
lot of #ifdef's, since our signalling model and file locking model are
quite different.  One can claim that it's *saner*, but it's at the same
time *different*.  

   4) GNU/Linux threads applications to be binary compatible on other Intel
   UNIX systems (ie., OpenBSD) running GNU emulation

Doubtful, without some explicit kernel support in the other forein Unix
system. 

   5) POSIX.1c compliance
   6) reasonable performance

Only with a lot of development work.   Part of the problem is that I
haven't seen anyone really passionate about Posix threads.  This is both
at the developer level, and at the application level.  

Given that we don't even have a customer for this work, it becomes hard
to get motivated about this.  If someone actual developer who *needs*
posix thredas were to step forward, like IBM's Java porting team, and be
willing to invest some energy in this, it would be different.  

I've always pushed for Posix threads because I figured that ISV's wanted
it, since they had programmers already trained in using the Posix
threading model (however broken it might be), and because they have
already existing code which uses Posix threads (again, no matter how
broken the standard might be, if you already have code using it, you
don't want to have to pry it open and change it).  

The problem is that we haven't been able to get any of these ISV's to
step forward, so people doubt whether they actually exist.  (In fact,
I'm beginning to have my doubts; maybe I was wrong.)  

							- Ted