This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [patch] for PR 17531

From: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>
To: Richard Henderson <rth at redhat dot com>
Cc: dje at watson dot ibm dot com, gcc-patches at gcc dot gnu dot org
Date: Thu, 23 Sep 2004 23:03:57 +0200
Subject: Re: [patch] for PR 17531
References: <20040923121639.GB6592@atrey.karlin.mff.cuni.cz> <200409231430.i8NEUFV34074@makai.watson.ibm.com> <20040923180251.GB19335@atrey.karlin.mff.cuni.cz> <20040923204014.GF8766@redhat.com>

Hello,

> > it is wrapped in NOP_EXPR.  NOP_EXPR is dropped in expander.
> 
> Very enlightening.

I was a bit in a hurry, but still I think I answered the question I got.
Sorry, I cannot read your mind...

> Specifically, I asked why the convert_to_mode around expr.c:7309
> isn't taking care of this.  Adding another convert_to_mode to
> work around this one not working is dumb.

Not really.  expand_binop does not assume that its operands have the
same type as result, so the patch would be a proper fix for the function
anyway.

As for your question: the expression in question looks like

(short int) &something.field

expr.c:7292 asks for &something.field in HImode, and this expansion
is the one that aborts.  The line you ask for is not reached.

Zdenek

Follow-Ups:
- Re: [patch] for PR 17531
  - From: Paolo Bonzini

References:
- [patch] for PR 17531
  - From: Zdenek Dvorak
- Re: [patch] for PR 17531
  - From: David Edelsohn
- Re: [patch] for PR 17531
  - From: Zdenek Dvorak
- Re: [patch] for PR 17531
  - From: Richard Henderson

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]