From owner-freebsd-current Sun Jan 11 23:22:58 1998 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id XAA09933 for current-outgoing; Sun, 11 Jan 1998 23:22:58 -0800 (PST) (envelope-from owner-freebsd-current) Received: from sax.sax.de (sax.sax.de [193.175.26.33]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id XAA09917 for ; Sun, 11 Jan 1998 23:22:50 -0800 (PST) (envelope-from j@uriah.heep.sax.de) Received: (from uucp@localhost) by sax.sax.de (8.8.8/8.8.8) with UUCP id IAA02388; Mon, 12 Jan 1998 08:22:29 +0100 (CET) (envelope-from j@uriah.heep.sax.de) Received: (from j@localhost) by uriah.heep.sax.de (8.8.8/8.8.5) id IAA05468; Mon, 12 Jan 1998 08:20:22 +0100 (MET) Message-ID: <19980112082022.30571@uriah.heep.sax.de> Date: Mon, 12 Jan 1998 08:20:22 +0100 From: J Wunsch To: =?iso-8859-1?Q?=E1=CE=C4=D2=C5=CA_=FE=C5=D2=CE=CF=D7?= Cc: Peter Wemm , FreeBSD-current , info-cvs@gnu.ai.mit.edu Subject: Re: CVS DIFF fix for review (-L added) Reply-To: Joerg Wunsch References: <199801111457.WAA06475@spinner.netplex.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Mailer: Mutt 0.88 In-Reply-To: =?iso-8859-1?Q?=3CPine=2EBSF=2E3=2E96=2E980111232840=2E17312A-100000=40l?= =?iso-8859-1?Q?sd=2Erelcom=2Eeu=2Enet=3E=3B_from_=E1=CE=C4=D2=C5=CA_=FE?= =?iso-8859-1?Q?=C5=D2=CE=CF=D7_on_Sun=2C_Jan_11=2C_1998_at_11=3A35=3A40P?= =?iso-8859-1?Q?M_+0300?= X-Phone: +49-351-2012 669 X-PGP-Fingerprint: DC 47 E6 E4 FF A6 E9 8F 93 21 E0 7D F9 12 D6 4E Sender: owner-freebsd-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk As Андрей Чернов wrote: > > If you can convince the info-cvs folks to make the changes, then I'll go > > along. Otherwise, IMHO the new patch and/or "standard" is busted. It > > wouldn't be the first time that POSIX have screwed something up. CVS and > > patch have been doing it like this since as far back as 1993. > > Please, note that it _not_ POSIX invention, POSIX just document > _existent_practice_ for years! Well, Peter, indeed. This only incidentally worked as intended as long as the -L option to diff/rcsdiff didn't exist, so at least one of the pathnames in the context diff header was invalid by the time patch has been called. Now, with -L and the basename of the file in both lines of the context diff header, it's very likely that this file exists in the current directory (like `Makefile', for a popular example), and patch will try patching the wrong file. The idea of Index: was great, but there has never been a patch utility supporting it the way CVS did expect. :-O Again, all the pre-existant ``support'' for Index: lines taking precedence over context-diff headers was _incidentally_. -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)