From minimalist@rocklinux.org Fri Oct 31 11:51:24 2003 Received: from nerd.clifford.at (localhost [127.0.0.1]) by nerd.clifford.at (8.12.1/8.12.1/rock) with ESMTP id h9VApOej015466; Fri, 31 Oct 2003 11:51:24 +0100 Received: (from daemon@localhost) by nerd.clifford.at (8.12.1/8.12.9/Submit) id h9VApJ19015458; Fri, 31 Oct 2003 11:51:19 +0100 Received: from mail.is-vn.bg (IDENT:2005@ns.is-vn.bg [212.50.28.34]) by nerd.clifford.at (8.12.1/8.12.1/rock) with SMTP id h9VAm8ej015217 for ; Fri, 31 Oct 2003 11:48:10 +0100 Received: (qmail 2437 invoked from network); 31 Oct 2003 10:40:50 -0000 Received: from jimmy.is.net (HELO jimmy) (192.168.193.6) by ncp.is.net with SMTP; 31 Oct 2003 10:40:50 -0000 Date: Fri, 31 Oct 2003 12:39:38 +0200 References: <572663576.20031029201551@freebits.de> <958742883.20031030094549@freebits.de> From: Dimitar Zhekov Organization: Information Services - Varna Content-Type: text/plain; charset=windows-1251; format=flowed MIME-Version: 1.0 Message-ID: In-Reply-To: <958742883.20031030094549@freebits.de> User-Agent: Opera7.11/Win32 M2 build 2887 X-AntiVirus: scanned for viruses by AMaViS 0.2.1 (https://amavis.org/) To: Precedence: list X-Sender: minimalist@rocklinux.org Reply-To: rock-linux@rocklinux.org Subject: Re: [rock-linux] [BUG] bootdisk sets LIMITCXX, other targets don't unset it again List-Help: List-Subscribe: List-Unsubscribe: List-Post: List-Owner: X-List-Server: Minimalist v2.2(4.2) (Cocteau) TCR>>> the bootdisk target sets LIMITCXX to 1, but other TCR>>> targets (like generic) doesn't explicitely set it to TCR>>> 0 again. That causes possibly several build errors DZ>> That applies to all settings locked for whatever DZ>> reason. A global locked-value-restoration is needed... TCR> Who speaks about locking? I'm just advocating that TCR> this shouldn't be on by default, because that's a TCR> really non-standard option... It is not _on by default_: ./scripts/config.in: bool 'Disable exceptions and rtti in C++' ROCKCFG_LIMITCXX 0 and it is locked to 1 in bootdisk: target/bootdisk/config.in:ROCKCFGSET_LIMITCXX=1 as well as several other options, for space reasons - a c++ application compiled with rtti and expections may be up to 25% larger. Now if a 'bootdisk' package fails because of LIMITCXX, that's a different beast and would need unlock. Basicly, when you choose the 'bootdisk' target and then go back to 'generic', nothing will be restored. See 'Binary package format' for a really bad example. DZ>> my limited selective build is OK with -fno-rtti/ DZ>> exceptions. TCR> Jimmy, c'm on.. that's no argument, is it? Well... consider it a praise :-) -- E-gards: Jimmy -- To unsubscribe from this list: send a mail with the subject "unsubscribe rock-linux" to . For more information about ROCK Linux have a look at .