From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 18617 invoked by alias); 20 Jul 2009 18:43:25 -0000 Received: (qmail 18608 invoked by uid 22791); 20 Jul 2009 18:43:25 -0000 X-SWARE-Spam-Status: No, hits=-2.6 required=5.0 tests=BAYES_00,SPF_PASS X-Spam-Check-By: sourceware.org Received: from mail-px0-f183.google.com (HELO mail-px0-f183.google.com) (209.85.216.183) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Mon, 20 Jul 2009 18:43:19 +0000 Received: by pxi13 with SMTP id 13so1939150pxi.12 for ; Mon, 20 Jul 2009 11:43:17 -0700 (PDT) Received: by 10.140.255.19 with SMTP id c19mr2388381rvi.24.1248115397159; Mon, 20 Jul 2009 11:43:17 -0700 (PDT) Received: from hotblack.localnet ([189.4.46.16]) by mx.google.com with ESMTPS id k37sm1516645rvb.30.2009.07.20.11.43.15 (version=SSLv3 cipher=RC4-MD5); Mon, 20 Jul 2009 11:43:16 -0700 (PDT) From: Thiago Jung Bauermann To: Tom Tromey Subject: Re: Value reference counting Date: Mon, 20 Jul 2009 19:45:00 -0000 User-Agent: KMail/1.11.4 (Linux/2.6.30-1-amd64; KDE/4.2.4; x86_64; ; ) Cc: gdb-patches@sourceware.org, Daniel Jacobowitz , Vladimir Prus References: <20090717184152.GA6863@caradoc.them.org> <200907201022.19758.thiago.bauermann@gmail.com> In-Reply-To: MIME-Version: 1.0 Content-Type: Text/Plain; charset="utf-8" Content-Transfer-Encoding: 8bit Content-Disposition: inline Message-Id: <200907201544.36156.thiago.bauermann@gmail.com> X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2009-07/txt/msg00493.txt.bz2 Em Segunda-feira 20 Julho 2009 13:45:55 Tom Tromey escreveu: > >>>>> "Thiago" == Thiago Jung Bauermann > >>>>> writes: > > Daniel> free_all_values can have a very long interval. But it looks like > Daniel> breakpoint commands always run it, so not unbounded user behavior. > Daniel> What about breakpoint conditions? Is anything released and free'd > Daniel> during a condition check going to linger until we stop? > > Thiago> I don't like the idea of having the GC being ran at unpredictably > long Thiago> intervals. Aren't we aiming at debugging big fat apps with big > insane Thiago> debugging sessions afterall? > > Running free_all_values at a long interval is potentially a problem. > However, any particular problem here can (most likely) be resolved > locally, using value_mark and value_free_to_mark in the code that causes > too much allocation. Ok, now we're all talking about the same issue. :-) If you guys feel that you can fix any potential performance issues for Real Big™ apps and debugging sessions, then I'm happy. -- []'s Thiago Jung Bauermann