From owner-freebsd-emulation@FreeBSD.ORG Mon Sep 20 23:46:32 2010 Return-Path: Delivered-To: freebsd-emulation@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 7EC5E106564A; Mon, 20 Sep 2010 23:46:32 +0000 (UTC) (envelope-from barbara.xxx1975@libero.it) Received: from cp-out2.libero.it (cp-out2.libero.it [212.52.84.102]) by mx1.freebsd.org (Postfix) with ESMTP id 1A5448FC23; Mon, 20 Sep 2010 23:46:31 +0000 (UTC) Received: from wmail3.libero.it (172.31.0.146) by cp-out2.libero.it (8.5.107) id 4C830A1D01073444; Tue, 21 Sep 2010 01:46:29 +0200 Message-ID: <6674929.370141285026389185.JavaMail.defaultUser@defaultHost> Date: Tue, 21 Sep 2010 01:46:29 +0200 (CEST) From: Barbara To: MIME-Version: 1.0 Content-Type: text/plain;charset="UTF-8" Content-Transfer-Encoding: 7bit X-SenderIP: 87.20.218.39 Cc: freebsd-emulation@freebsd.org, jkim@FreeBSD.org Subject: R: Re: nspluginwrapper(-devel) and core dumps X-BeenThere: freebsd-emulation@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: Barbara List-Id: Development of Emulators of other operating systems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 20 Sep 2010 23:46:32 -0000 >> Date: Tue, 21 Sep 2010 00:12:22 +0200 (CEST) >> From: Barbara >> Sender: owner-freebsd-emulation@freebsd.org >> >> >> I have a question about nspluginwrapper and linux-f10-flashplugin10: >> is it possible to rebuild nspluginwrapper so it doesn't dump core files >> (npviewer.bin.core)? >> Is it possible to have it dying "silently"? >> >> The main reason it that when it starts dumping core, the whole browser freezes >> until the file it's completely written. >> In case of heavy disk I/O, e.g. upgrading some ports or world, it can take a >> lot of time. >> >> The other reason is that those files are really huge. I often get 300mb to 1.2 >> gb core file when I have a discrete number of tab opened. >> And I don't even play flash-games. And I have flashblock installed. I just >> watch at some videos on youtube or similar. >> I know that I could limit the size of coredumps, but this could prevent me >> getting backtrace from other sw. >> I'm fearing that the extra stress on hd could hurt it. >> But please tell me if they are still useful for debugging, despite being linux >> binaries. >> >> I was thinking about contacting the author to ask him. >> But the site reported in pkg-descr (http://gwenole.beauchesne. >> info/projects/nspluginwrapper/) in not reachable for me. >> And the last news on freshmeat is dated "02 Jan 2009". >> Does anyone know what happened to him? And what about his project? >> >> For what I can understand npviewer.bin and a couple of .so should be rebuilt >> on Fedora 10 i386. >> If that's a problem, I have a VM with F10. Can this help? >> Could someone suggest me the magic to rebuild the linux part of the port? > >Barbara, > >While I don't know of any way to make an executable not dump core when >it crashes, you can tell you shell to not do so. For tcsh (FreeBSD >default), "limit coredumpsize 0" should do the trick. Of course, no core >files will be dumped for any executable that crashes, but this may be >just what you want for the default. Similar commands are available for >other shells. I'm trying the suggestion Alexander Kabaev gave me on the previous reply. Obviously npviewer isn't crashing anymore! Even if I'm trying hard opening and closing furiously a lot of tab with flash content. > >BTW, If you have an F10 VM, you might find it VERY popular if you can >build a current Linux pango. There is a security vulnerability in the >last F10 RPM of this, so having a patched RPM for it would be nice. I would be happy to do that, but I don't know about .spec files, which I think are needed to build rpms. And probably I'll have a lot of other deps to build, also not available as rpm. I'll try to have a look during next weekend, but I think I am not be able to do that. I remember about a tool, "checkinstall" if I'm not wrong, I used many years ago to build rpms from source tarballs. But I don't know if it's still working. Could it be an idea? Barbara