PATCH: put setjmp warnings under control of -W

Marc Lehmann pcg@opengroup.org
Mon Nov 1 12:46:00 GMT 1999


On Mon, Nov 01, 1999 at 01:53:19PM +1100, Geoff Keating <geoffk@ozemail.com.au> wrote:
> It is as easy to fix this as it is to fix the normal uninitialized
> warnings; just make the affected variables 'volatile'.

This "fix" imposes a major speed penalty.

> is easy enough to make it acceptable to have this under -Wall, when
> the alternative is a change to the functionality of the compiler.

Forcing people to use highly suboptimal code is at least not acceptable to
me. A superflous initialization of a local might be acceptable (although I´d
use an attribute if it existed), compared to the procedure call overhead.

> In any case if you make this change you must change the documentation
> to match, and probably you should get the patch re-approved with the
> documentation changes.

The patch, however, is already in (it could easily be reverted)... I´ll
send an documentation patch soon.

> I believe the warning has been controlled by -Wuninitialized since at
> least 2.7.2.

Ok.. my working these is that the documentation _was_ right and the
implementation was wrong. We should clarify this point before more changes
happen.

-- 
      -----==-                                             |
      ----==-- _                                           |
      ---==---(_)__  __ ____  __       Marc Lehmann      +--
      --==---/ / _ \/ // /\ \/ /       pcg@opengroup.org |e|
      -=====/_/_//_/\_,_/ /_/\_\       XX11-RIPE         --+
    The choice of a GNU generation                       |
                                                         |


More information about the Gcc-patches mailing list