This is the mail archive of the java@gcc.gnu.org mailing list for the Java 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: GCJ problem on RS/6000

From: David Edelsohn <dje at watson dot ibm dot com>
To: "Martyn Honeyford" <MARTYNH at uk dot ibm dot com>
Cc: Tom Tromey <tromey at redhat dot com>, Joe Buck <jbuck at synopsys dot com>, java at gcc dot gnu dot org
Date: Fri, 08 Feb 2002 17:21:12 -0500
Subject: Re: GCJ problem on RS/6000
References: <OF549ED3A8.7FE8473B-ON80256B5A.003627BE@portsmouth.uk.ibm.com> <87u1ssdpku.fsf@creche.redhat.com> <200202081841.NAA24946@makai.watson.ibm.com>

	With a little more investigating, it appears that Sun's Java
Compiler set the precedent of using $ as the separator for inner classes.
I guess this is now part of some spec / ABI.  But that only defines the
names of symbols in class files, not symbols in object files and
executables.  Other than convenience, I do not see any reason that GCJ and
G++ need to generate $ in symbol names.  GCJ could translate everything
rot13 or EBCDIC because the spec cannot impose requirements at that level.

	Allowing $ in identifiers is one thing, but blindly copying them
through to assembly files when NO_DOLLAR_IN_LABEL is defined creates big
problems for portability of GCC, G++ and GCJ.

David

Follow-Ups:
- Re: GCJ problem on RS/6000
  - From: Tom Tromey
- Re: GCJ problem on RS/6000
  - From: Joe Buck

References:
- Re: GCJ problem on RS/6000
  - From: Martyn Honeyford
- Re: GCJ problem on RS/6000
  - From: Tom Tromey
- Re: GCJ problem on RS/6000
  - From: David Edelsohn

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