This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: Symbol compression?, linking problems.
- To: Shane Nay <shane at minirl dot com>
- Subject: Re: Symbol compression?, linking problems.
- From: Alexandre Petit-Bianco <apbianco at cygnus dot com>
- Date: Mon, 14 May 2001 16:40:57 -0700 (PDT)
- Cc: java at gcc dot gnu dot org
- References: <0105150750150A.00804@compiler>
- Reply-To: apbianco at cygnus dot com
Shane Nay writes:
> So.., what I noticed was everytime a "compression match" was found
> for a specific symbol that was related to a native function, things
> got ugly.
Can you send us a small test case?
> Basically, "gcj" when mangling the symbol names to encode object type
> parameters mangled them differently than g++ when compiling my native
> code..., so they wouldn't link.
Yes it won't link. Can you find out what c++filt thinks of the two
mangled symbols?
> Is there something new in the ABI that allows for references to
> objects encoded in symbols to be compressed somehow?
Not that I know. The symbols (hence the object) would have to be
regenerated from source using updated compilers. Both gcj and g++ from
the trunk and the branch are in sync with respect to the new mangling,
minus bugs that might exist in both implementation but most likely in
gcj since it's the newer of both implementations (but this only a test
case would tell us.)
./A