Re: java/247

Re: java/247

[apbianco]

Synopsis: ICE compiling methods which handle Double values @ -O2

State-Changed-From-To: open->feedback
State-Changed-By: apbianco
State-Changed-When: Tue Jun  5 21:44:52 2001
State-Changed-Why:
    I can't reproduce the bytecode to native failures on x86/linux.
    Alpha has one failure with -O, two with -O2 and non with -O3.
    PPC doesn't show any failures. And the DTest class builds on
    all three platforms from -O to -O3.

http://sources.redhat.com/cgi-bin/gnatsweb.pl?cmd=view&pr=247&database=gcc

Re: java/247

[Bryce McKinlay]

The following reply was made to PR java/247; it has been noted by GNATS.

From: Bryce McKinlay <bryce@waitaki.otago.ac.nz>
To: bryce@gcc.gnu.org, gcc-gnats@gcc.gnu.org, warnes@queenbee.fhcrc.org,
        apbianco@redhat.com, Jeff Sturm <jeff.sturm@appnet.com>
Cc:  
Subject: Re: java/247
Date: Sat, 03 Mar 2001 16:07:10 +1300

 >  FWIW this might be the same ICE I'm seeing.  It only occurs in certain methods
 >  that handle double values, and only with -O and above on x86, but with the
 >  current compiler.
 >
 >  Here's a minimal test case:
 >
 >  public class DTest {
 >      static double strtod(String s) {
 >          double result = 0;
 >
 >          try {
 >              result = Double.valueOf(s).doubleValue();
 >          } catch (NumberFormatException e) {
 >          }
 >          if (result == Double.NaN) {
 >              return 0;
 >          }
 >          return result;
 >      }
 >  }
 >
 
 Thanks for the test case, Jeff. I can reproduce this bug with the
 current 3.0 branch compiler, so I'm going to update the PR name and
 reopen it.
 
 regards
 
   [ bryce ]
 
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=247&database=gcc
 
 
 

Re: java/247

[Jeff Sturm]

The following reply was made to PR java/247; it has been noted by GNATS.

From: Jeff Sturm <jeff.sturm@appnet.com>
To: bryce@gcc.gnu.org, gcc-gnats@gcc.gnu.org, warnes@queenbee.fhcrc.org
Cc:  
Subject: Re: java/247
Date: Tue, 09 Jan 2001 16:48:18 -0500

 FWIW this might be the same ICE I'm seeing.  It only occurs in certain methods
 that handle double values, and only with -O and above on x86, but with the
 current compiler.
 
 Here's a minimal test case:
 
 public class DTest {
     static double strtod(String s) {
         double result = 0;
 
         try {
             result = Double.valueOf(s).doubleValue();
         } catch (NumberFormatException e) {
         }
         if (result == Double.NaN) {
             return 0;
         }
         return result;
     }
 }
 
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=247&database=gcc
 
 --
 Jeff Sturm
 jeff.sturm@commerceone.com
>From harri@trolltech.com Sun Apr 01 00:00:00 2001
From: harri@trolltech.com
To: gcc-gnats@gcc.gnu.org
Subject: other/1697: bootstrapping fails: malformed option `-A system=posix'
Date: Sun, 01 Apr 2001 00:00:00 -0000
Message-id: <20010118171142.14918.qmail@sourceware.cygnus.com>
X-SW-Source: 2001-q1/msg00451.html
Content-length: 2159

>Number:         1697
>Category:       other
>Synopsis:       bootstrapping fails: malformed option `-A system=posix'
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jan 18 09:16:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     harri@trolltech.com
>Release:        2.97
>Organization:
>Environment:
i386-linux/2.95.2
>Description:
I can't bootstrap the current CVS (20010118) with gcc 2.95.2
and srcdir != objdir.

configure --enable-shared
make bootstrap
....
/usr/local/src/gcc/obj/gcc/xgcc -B/usr/local/src/gcc/obj/gcc/ -B/usr/local/i686-pc-linux-gnu/bin/ -B/usr/local/i686-pc-linux-gnu/lib/ -isystem /usr/local/i686-pc-linux-gnu/include -dumpspecs > tmp-specs
mv tmp-specs specs
echo "int xxy_us_dummy;" >tmp-dum.c
/usr/local/src/gcc/obj/gcc/xgcc -B/usr/local/src/gcc/obj/gcc/ -B/usr/local/i686-pc-linux-gnu/bin/ -B/usr/local/i686-pc-linux-gnu/lib/ -isystem /usr/local/i686-pc-linux-gnu/include -S tmp-dum.c
echo '/*WARNING: This file is automatically generated!*/' >tmp-under.c
if grep _xxy_us_dummy tmp-dum.s > /dev/null ; then \
  echo "int prepends_underscore = 1;" >>tmp-under.c; \
else \
  echo "int prepends_underscore = 0;" >>tmp-under.c; \
fi
/bin/sh /usr/local/src/gcc/gcc/gcc/move-if-change tmp-under.c underscore.c
rm -f tmp-dum.c tmp-dum.s
touch s-under
gcc -c  -DIN_GCC    -g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long  -DHAVE_CONFIG_H    -I. -I. -I/usr/local/src/gcc/gcc/gcc -I/usr/local/src/gcc/gcc/gcc/. -I/usr/local/src/gcc/gcc/gcc/config -I/usr/local/src/gcc/gcc/gcc/../include underscore.c -o underscore.o
underscore.c:0: malformed option `-A system=posix'
underscore.c:0: malformed option `-A cpu=i386'
underscore.c:0: malformed option `-A machine=i386'
make[1]: *** [underscore.o] Error 1
make[1]: Leaving directory `/usr/local/src/gcc/obj/gcc'
make: *** [all-gcc] Error 2

The 'spec' file in the gcc dir is picked up by my old 2.95
and not understood.
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
>From rodrigc@mediaone.net Sun Apr 01 00:00:00 2001
From: Craig Rodrigues <rodrigc@mediaone.net>
To: nobody@gcc.gnu.org
Cc: gcc-prs@gcc.gnu.org
Subject: Re: c++/1925
Date: Sun, 01 Apr 2001 00:00:00 -0000
Message-id: <20010210190600.10573.qmail@sourceware.cygnus.com>
X-SW-Source: 2001-q1/msg01116.html
Content-length: 672

The following reply was made to PR c++/1925; it has been noted by GNATS.

From: Craig Rodrigues <rodrigc@mediaone.net>
To: bh@techhouse.org, gcc-gnats@gcc.gnu.org, nobody@gcc.gnu.org,
   rodrigc@mediaone.net
Cc:  
Subject: Re: c++/1925
Date: Fri, 09 Feb 2001 22:39:19 -0500

 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=1925&database=gcc
 
 Sir,
 
 Compiling your testcase with the latest CVS snapshot of gcc yields this
 error message:
 
 a.cpp: In destructor `virtual A::~A()':
 a.cpp:6: abstract virtual `virtual void A::f()' called from destructor
 
 Does this address your issue?
 
 --
 Craig Rodrigues
 http://www.gis.net/~craigr
 rodrigc@mediaone.net