From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 15357 invoked by alias); 7 Mar 2003 18:01:00 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 15350 invoked from network); 7 Mar 2003 18:00:59 -0000 Received: from unknown (HELO jackfruit.Stanford.EDU) (171.64.38.136) by 172.16.49.205 with SMTP; 7 Mar 2003 18:00:59 -0000 Received: (from carlton@localhost) by jackfruit.Stanford.EDU (8.11.6/8.11.6) id h27I0vL20532; Fri, 7 Mar 2003 10:00:57 -0800 X-Authentication-Warning: jackfruit.Stanford.EDU: carlton set sender to carlton@math.stanford.edu using -f To: Daniel Jacobowitz Cc: gdb-patches@sources.redhat.com Subject: Re: [dictionary] commit for 2003-03-06 References: <20030307144017.GA20963@nevyn.them.org> From: David Carlton Date: Fri, 07 Mar 2003 18:01:00 -0000 In-Reply-To: Message-ID: User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Common Lisp) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-SW-Source: 2003-03/txt/msg00173.txt.bz2 On 07 Mar 2003 08:53:00 -0800, David Carlton said: > And there might be uses elsewhere, too: I haven't yet audited all the > uses of lookup_minimal_symbol. I'm fairly sure that the symbol lookup > functions in symtab.c are safe, from memory, and there shouldn't be > _too_ many other places that want a natural name when looking at > minimal symbols, but there might be some. In fact, there are: the .y files all have likely candidates, and I want to look more closely at a handful of other uses (in remote.c, printcmd.c, objc-lang.c, and valops.c). So the demangled hash table definitely stays. Which raises another interesting issue: should whatever of those uses turn out to want natural names also accept linkage names? I'm reluctantly starting to think that they should. They typically call lookup_minimal_symbol with a string taken from user input; in an ideal world, I think it would be reasonable to not allow users to get at stuff via linkage names, but there have been enough C++ bugs that can be worked around if you know the linkage name that I don't think we're there yet. Hmm; I'll have to think about what internal interface I like, given that desire. David Carlton carlton@math.stanford.edu