Re: Compiler Warnings, again

[spector@mimsy.umd.edu (Lee Spector)]

So much acrimony over hiding warnings! I generally write warning-less
code, but I think there's a fine reason to have a simple way to turn off all
warnings:

  You want to give someone some code that is still under development.
  You know that the warnings don't indicate real problems for the user
  in question, and you don't want the user to be faced with a screen full
  of warnings every time they load the code.  All of the warnings will
  be fixed at some point, but for now you just want to hand a user something
  that will load without warnings. Sometimes you get into this situation
  because someone else on your programming team has written warning-generating
  code, and you haven't yet coaxed them into cleaning it up. But it works,
  and you want to provide it to a user.

There was an earlier round of messages on warning-hiding -- I guess it was
on the info-mcl mail list last summer. Some without-warnings macros were
posted, but if memory serves there were cases in which they didn't work.
(I did some experiments, but I can't find them...) Anyway, I'd be interested
in info on the state-of-the-art in warning-hiding, and I'd second the motion
that such techniques be documented, whether they're standardized or not.

Lee Spector (spector@cs.umd.edu)
Department of Computer Science
University of Maryland
College Park, MD 20742
USA