Movatterモバイル変換

This is the mail archive of thebinutils@sources.redhat.commailing list for thebinutils project.

Re: Does a zero-sized object in BSS go away?

• From: Thiemo Seufer <ica2_ts at csv dot ica dot uni-stuttgart dot de>
• To: Geoff Keating <geoffk at geoffk dot org>
• Cc: kazu at cs dot umass dot edu, binutils at sources dot redhat dot com
• Date: Fri, 27 Jun 2003 22:59:36 +0200
• Subject: Re: Does a zero-sized object in BSS go away?
• References: <20030619.114107.129788470.kazu@cs.umass.edu> <20030619.120003.127177937.kazu@cs.umass.edu> <jmisr2rlex.fsf@desire.geoffk.org> <20030619.141939.27391978.kazu@cs.umass.edu> <200306192213.h5JMDs021996@desire.geoffk.org> <20030627091031.GT31215@rembrandt.csv.ica.uni-stuttgart.de> <200306272041.h5RKflW07494@desire.geoffk.org>

Geoff Keating wrote:[snip]> > > ISO C programs can't create zero-size objects, but you can do it in> > > GCC.> > > > > > I think there's no way to tell if an object, which the program thinks> > > is zero-size, actually has non-zero size.  So you could just make the> > > common symbols have size 1 if the user asks for size 0...> > > > This may break pointer comparisions. (Not that I knew of a good reason> > to do this).> > I don't see how this could break pointer comparisons.Gcc allows to take the address of a zero-sized object:struct a {int b[0];int c[0];} d;int main(void){return &d.b == &d.c;}> You can't meaningfully do them with zero-size objects anyway...It's probably (hopefully :-) only a theoretical point and not apractical issue.Thiemo

• References:
  • Re: Does a zero-sized object in BSS go away?
    • From: Kazu Hirata
  • Re: Does a zero-sized object in BSS go away?
    • From: Geoff Keating
  • Re: Does a zero-sized object in BSS go away?
    • From: Kazu Hirata
  • Re: Does a zero-sized object in BSS go away?
    • From: Geoff Keating
  • Re: Does a zero-sized object in BSS go away?
    • From: Thiemo Seufer
  • Re: Does a zero-sized object in BSS go away?
    • From: Geoff Keating