java/1199: Internal Compiler Error bootstrapping CVS

java/1199: Internal Compiler Error bootstrapping CVS

[gchambers]

>Number:         1199
>Category:       java
>Synopsis:       Internal Compiler Error bootstrapping CVS
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    green
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Dec 20 12:17:28 PST 2000
>Closed-Date:    Sun Feb 13 18:01:00 PST 2000
>Last-Modified:  Sun Feb 13 18:16:01 PST 2000
>Originator:     gchambers@provsol.com
>Release:        CVS as of thursday, Jan 27
>Organization:
>Environment:
Redhat 6.0 Linux on i486 chip
>Description:
Current CVS as of Thursday 27 Jan dies with internal
compiler error compiling java/lang/Math.java from 
current run-time CVS as of same date.

Reported to gcc-bugs@gcc.gnu.org at
http://gcc.gnu.org/ml/gcc-bugs/2000-01/msg00870.html
>How-To-Repeat:
Get current CVS trees and compile.  Configured as per gcc-bug report...
>Fix:

>Release-Note:

>Audit-Trail:

Formerly PR gcj/142

Responsible-Changed-From-To: apbianco->green
Responsible-Changed-By: green
Responsible-Changed-When: Sun Feb 13 10:17:43 2000
Responsible-Changed-Why:
    I'll deal with this one.
State-Changed-From-To: open->analyzed
State-Changed-By: green
State-Changed-When: Sun Feb 13 10:17:43 2000
State-Changed-Why:
    Bugs come and go in the snapshot.
    
    I did notice one interesting thing, which was that you
    were using sjlj exceptions.  The libgcj configuration
    code detaults to sjlj when it doesn't recognize the
    target.  Unfortunately, we had not told it about i486 systems,
    which is what you're using (thanks for the output btw).
    
    I've fixed this in libgcj.  Try updating libgcj and building it
    again.  If the problem doesn't go away, try updating
    your compiler and building again.  If it still doesn't 
    go away then we'll have to look at the compiler.

From: green@cygnus.com
To: apbianco@cygnus.com, gchambers@provsol.com, green@cygnus.com,
  java-gnats@sourceware.cygnus.com
Cc:  
Subject: Re: gcj/142
Date: 13 Feb 2000 18:17:43 -0000

 Synopsis: Internal Compiler Error bootstrapping CVS
 
 Responsible-Changed-From-To: apbianco->green
 Responsible-Changed-By: green
 Responsible-Changed-When: Sun Feb 13 10:17:43 2000
 Responsible-Changed-Why:
     I'll deal with this one.
 State-Changed-From-To: open->analyzed
 State-Changed-By: green
 State-Changed-When: Sun Feb 13 10:17:43 2000
 State-Changed-Why:
     Bugs come and go in the snapshot.
     
     I did notice one interesting thing, which was that you
     were using sjlj exceptions.  The libgcj configuration
     code detaults to sjlj when it doesn't recognize the
     target.  Unfortunately, we had not told it about i486 systems,
     which is what you're using (thanks for the output btw).
     
     I've fixed this in libgcj.  Try updating libgcj and building it
     again.  If the problem doesn't go away, try updating
     your compiler and building again.  If it still doesn't 
     go away then we'll have to look at the compiler.
 
 http://sourceware.cygnus.com/cgi-bin/gnatsweb.pl?cmd=view&pr=142&database=java

From: Glenn Chambers <gchamber@bright.net>
To: java-gnats@sourceware.cygnus.com, gchambers@provsol.com, green@cygnus.com
Cc:  
Subject: Re: gcj/142
Date: Sun, 13 Feb 2000 20:47:51 -0500

 (Checking in from home; ignore bogus return address, I've got my
 browser set to maximal paranoia.)
 
 I did know that snapshots were going to be unstable, but
 then again, bugs don't get fixed if you don't report
 them <grin>.
 
 I did two different things; I'm not sure which fixed it.
 
 First, I did a make bootstrap-lean on the compiler, then deleted tree.o
 and did a make CFLAGS=-fno-builtin, in case the issue was the problem
 that's been killing Godmar Beck's builds.
 
 Second, I did a CVS update in LIBGCJ, and did the make.  I did see your
 change adding i386 and i486 to the configure.hosts file.
State-Changed-From-To: analyzed->closed
State-Changed-By: green
State-Changed-When: Sun Feb 13 18:01:00 2000
State-Changed-Why:
    Glenn reports that the problem has gone away.

From: green@cygnus.com
To: gchambers@provsol.com, green@cygnus.com, java-gnats@sourceware.cygnus.com
Cc:  
Subject: Re: gcj/142
Date: 14 Feb 2000 02:01:00 -0000

 Synopsis: Internal Compiler Error bootstrapping CVS
 
 State-Changed-From-To: analyzed->closed
 State-Changed-By: green
 State-Changed-When: Sun Feb 13 18:01:00 2000
 State-Changed-Why:
     Glenn reports that the problem has gone away.
 
 http://sourceware.cygnus.com/cgi-bin/gnatsweb.pl?cmd=view&pr=142&database=java
>Unformatted: