This is the mail archive of the
mailing list for the GDB project.
Re: Strange change to read_namespace
- From: David Carlton <carlton at kealia dot com>
- To: gdb-patches at sources dot redhat dot com
- Date: Mon, 29 Mar 2004 08:51:41 -0800
- Subject: Re: Strange change to read_namespace
- References: <20040327171044.GA9440@nevyn.them.org>
On Sat, 27 Mar 2004 12:10:44 -0500, Daniel Jacobowitz <firstname.lastname@example.org> said:
> When you checked in this patch:
> 2004-03-15 David Carlton <email@example.com>
> Fix for PR c++/1553:
> * dwarf2read.c (read_structure_type): Determine type name by
> calling determine_class_name.
> (determine_class_name): New.
> (determine_prefix): Look at TYPE_TAG_NAME and call
> determine_class_name when appropriate.
> (determine_prefix_aux, class_name): Delete.
> You also checked in this:
> @@ -3724,6 +3717,7 @@ read_namespace (struct die_info *die, st
> TYPE_TAG_NAME (type) = TYPE_NAME (type);
> new_symbol (die, type, cu);
> + die->type = type;
> if (is_anonymous)
> cp_add_using_directive (processing_current_prefix,
> Did you mean to do that? It seems plausible.
Whoops, sorry, I guess that slipped through the ChangeLog. Yes, I
meant to do that: that's so that determine_prefix can check to see if
the die it's trying to find the prefix for has a type and, if so, look
at its TYPE_TAG_NAME.
I'll go and update that ChangeLog entry:
(read_namespace): Set die->type.