This is the mail archive of the
libstdc++-prs@sourceware.cygnus.com
mailing list for the libstdc++ project.
RE: libstdc++/16
- To: bkoz at cygnus dot com
- Subject: RE: libstdc++/16
- From: Benjamin Kosnik <bkoz at cygnus dot com>
- Date: 20 Mar 2000 21:07:00 -0000
- Cc: libstdc++-prs at sourceware dot cygnus dot com,
- Reply-To: Benjamin Kosnik <bkoz at cygnus dot com>
The following reply was made to PR libstdc++/16; it has been noted by GNATS.
From: Benjamin Kosnik <bkoz@cygnus.com>
To: "Gibbons, Chad" <Chad_Gibbons@bmc.com>
Cc: libstdc++-gnats@sourceware.cygnus.com
Subject: RE: libstdc++/16
Date: Mon, 20 Mar 2000 12:57:48 -0800 (PST)
I guess you have not convinced me that it is a bug, unless I can reproduce
it. Furthermore, I believe keeping that initializer in the header is
necessary to conform to c++ static initializer rules.
I have, however, run across some weird solaris initializer problems when I
was fixing other parts of the library. First off, you need to use GNU ld.
Second, it looks like g++ on solaris is putting stuff into .bss instead of
.data:
http://gcc.gnu.org/ml/gcc-bugs/2000-03/msg00274.html
Unfortunately, you'll need to come up with a testcase before this can go
much further.
-benjamin
> A quick look at today's CVS archives shows the code with the bug hasn't been
> modified.
>
> The problem only manifests itself in very large projects, so it's difficult
> to reproduce. I have a patch available, but I don't think anyone was
> comfortable with the change unless there is a small test case to reproduce
> it.