From owner-cvs-all@FreeBSD.ORG Tue May 22 17:01:09 2007 Return-Path: X-Original-To: cvs-all@FreeBSD.org Delivered-To: cvs-all@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id CE79E16A41F; Tue, 22 May 2007 17:01:09 +0000 (UTC) (envelope-from rwatson@FreeBSD.org) Received: from cyrus.watson.org (cyrus.watson.org [209.31.154.42]) by mx1.freebsd.org (Postfix) with ESMTP id 81F0113C4B9; Tue, 22 May 2007 17:01:09 +0000 (UTC) (envelope-from rwatson@FreeBSD.org) Received: from fledge.watson.org (fledge.watson.org [209.31.154.41]) by cyrus.watson.org (Postfix) with ESMTP id B334D4731F; Tue, 22 May 2007 13:01:08 -0400 (EDT) Date: Tue, 22 May 2007 13:01:08 -0400 (EDT) From: Robert Watson X-X-Sender: robert@fledge.watson.org To: Warner Losh In-Reply-To: <20070522.101526.78768763.imp@bsdimp.com> Message-ID: <20070522125959.Q28780@fledge.watson.org> References: <200705192051.l4JKp0rF074703@repoman.freebsd.org> <20070522003856.GX21795@elvis.mu.org> <20070522112825.U50138@fledge.watson.org> <20070522.101526.78768763.imp@bsdimp.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Cc: cvs-src@FreeBSD.org, alfred@FreeBSD.org, andre@FreeBSD.org, cvs-all@FreeBSD.org, src-committers@FreeBSD.org Subject: Re: cvs commit: src/sys/kern uipc_syscalls.c X-BeenThere: cvs-all@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: CVS commit messages for the entire tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 22 May 2007 17:01:09 -0000 On Tue, 22 May 2007, Warner Losh wrote: >> On Mon, 21 May 2007, Alfred Perlstein wrote: >> >>> Is this the same bug that used to be in sendfile but got corrected some >>> years back? >> >> Dunno, but it sounds like we really need a comprehensive set of edge-case >> tests for sendfile. > > There are those that advocate writing a test case that fails, then fixing > the code until the test case succeeds, then committing the test case to the > test-bed at the same time the code is committed to the tree. With the > test-case gating the code. > > This works well for some thing, but we can't do it universally since some > test-cases would be only for 'rogue' hardware, which can't be automatically > tested. In this case, the problem is that we already have a piece of complex subsystem code without tests, so we bump into both regressions and new bugs with moderate frequency, and need to write basic functional tests. An interesting variable in test cases for sendfile, btw, is to use UNIX domain sockets as a transport, not just TCP Robert N M Watson Computer Laboratory University of Cambridge