Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 09 Mar 1997 02:29:46 -0800
From:      "Jordan K. Hubbard" <jkh@time.cdrom.com>
To:        current@freebsd.org
Subject:   2 questions about C++ support in 2.2
Message-ID:  <24352.857903386@time.cdrom.com>

next in thread | raw e-mail | index | archive | help
I've noticed lately in trying to compile some of the more
complex C++ applications I've been playing with (one being
ports/devel/tvision to cite a specific example) that the
assembler now spits out quite a bit of stuff along these
lines:

/var/tmp/cc024167.s:10678: Warning: GOT relocation burb: `_Am_Standard_Widget_Abort_Method_proc__FG9Am_Object' should be global
/var/tmp/cc024167.s:10678: Warning: GOT relocation burb: `_Am_Standard_Widget_Stop_Method_proc__FG9Am_ObjectG8Am_Value' should be global
/var/tmp/cc024167.s:10678: Warning: GOT relocation burb: `_Am_Standard_Widget_Start_Method_proc__FG9Am_ObjectG8Am_Value' should be global
/var/tmp/cc024167.s:10678: Warning: GOT relocation burb: `_border_rectangle_draw_proc__FG9Am_ObjectP13Am_Drawonableii' should be global
/var/tmp/cc024167.s:10678: Warning: GOT relocation burb: `_do_accelerator_action_proc__FG9Am_Object' should be global

And so on (these messages are from a compilation of the Amulet 2.0
library, in case anyone actually cares).

So now the two questions:

1. What the hell is a "relocation burb" and how can one avoid them? :-)

2. Shouldn't these messages demange the symbol first, so you don't
   have to hassle with c++filt?  I thought we were here once before
   (though probably in connection with the loader) and everyone agreed
   that built-in symbol demanging was good, very user friendly and
   we should all do it.  Whatever happened to that idea? :-)

   If it's just something which has fallen on the floor, it annoys me
   enough that I'd be willing to pick it up myself.

Yours for a nicer (relatively speaking) C++ environment...

					Jordan

P.S. I don't expect for us to have anything improved for 2.2-RELEASE,
but maybe we could at least get it fixed up in time for 2.2.1? :)



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?24352.857903386>