Subject: Re: setq x setf
From: Erik Naggum <erik@naggum.no>
Date: 2000/06/15
Newsgroups: comp.lang.lisp
Message-ID: <3170070479999597@naggum.no>

* Paul Dietz
| Nitpick: in theory, yes, in practice there can be a trivial but
| nonzero difference in some implementations.

  If practice differs from theory, you have found a bug.  Report bugs
  as bugs to be removed, not as some nasty thing programmers need to
  be made _aware_ of.  Sheesh!  Such is the worst of all possible ways
  to deal with problems.  If you don't get a fix within a reasonable
  amount of time and the bug causes you to have to circumvent good
  engineering decisions and write bad code that increases the overall
  complexity of the system and introduces portability and maintenance
  issues that are otherwise unwarranted, complain _vociferously_.
  Accept no substitute for fixes to demonstrable bugs!

  (Well, in this here day and age, one _acceptable_ "substitute" for
  fixes to other people's mistakes is a reduced purchase price of the
  buggy product, such as you getting the source code and assuming full
  responsibilit for any problems.  To some of us, that is unworkable.)

| For example, if *x* is a global variable then the forms (setq *x* t)
| and (setf *x* t) compile to different code in ACL 5.0.1 ...

  I have been unable to reproduce this as stated.

#:Erik
-- 
  If this is not what you expected, please alter your expectations.