This is the mail archive of the
mailing list for the GCC project.
Re: Small change to expand_expr CONSTRUCTOR handling
- From: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- To: bosch at gnat dot com
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 22 May 03 17:17:34 EDT
- Subject: Re: Small change to expand_expr CONSTRUCTOR handling
Could you be a more explicit? Was there a compilation failure or
wrong code generated? What is the nature of the bug and why is
this a good fix?
Wrong code generated. The issue was that the elaboration code for the
initialization of library-level data put pointers to the stack of the
elaboration routine into the library-level data, which is always wrong.
The reason it did so was that it didn't treat a CONSTRUCTOR as static.
However, if you're looking for a constant address, giving the address
of a static CONSTRUCTOR is best.