From owner-cvs-ports Wed Jan 21 03:50:04 1998 Return-Path: Received: (from daemon@localhost) by hub.freebsd.org (8.8.8/8.8.8) id DAA17081 for cvs-ports-outgoing; Wed, 21 Jan 1998 03:50:04 -0800 (PST) (envelope-from owner-cvs-ports) Received: from spinner.netplex.com.au (root@spinner.netplex.com.au [202.12.86.3]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id DAA17025; Wed, 21 Jan 1998 03:49:52 -0800 (PST) (envelope-from peter@netplex.com.au) Received: from spinner.netplex.com.au (peter@localhost [127.0.0.1]) by spinner.netplex.com.au (8.8.8/8.8.8/Spinner) with ESMTP id TAA22231; Wed, 21 Jan 1998 19:49:21 +0800 (WST) (envelope-from peter@spinner.netplex.com.au) Message-Id: <199801211149.TAA22231@spinner.netplex.com.au> X-Mailer: exmh version 2.0zeta 7/24/97 To: =?KOI8-R?B?4c7E0sXKIP7F0s7P1w==?= cc: cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG, cvs-ports@FreeBSD.ORG Subject: Re: cvs commit: ports/www/squid12 Makefile In-reply-to: Your message of "Wed, 21 Jan 1998 09:08:51 +0300." Date: Wed, 21 Jan 1998 19:49:20 +0800 From: Peter Wemm Sender: owner-cvs-ports@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk =?KOI8-R?B?4c7E0sXKIP7F0s7P1w==?= wrote: > On Tue, 20 Jan 1998, Peter Wemm wrote: > > > peter 1998/01/20 19:06:06 PST > > > > Modified files: > > www/squid12 Makefile > > Log: > > Update this for the benefit of the new broken patch. Tell it to look at > > Index: like posix says, rather than going into dummy mode. The distribut ed > > patches are generated with standard cvs and don't have ache's hack (which > > doesn't work when new files are added anyway). > > As I say elsewhere please stop fixing something for new -current patch, we > NOT decide yet what to do with it. I.e. there is a chance you'll need to > back out you fix in case new patch goes to ports. > > Our CVS hack is an extension, it is backward-compatible. Yes, but the 'distributed patches' are generated on SOMEBODY ELSE's unpatched cvs, not even a FreeBSD system, so what we do to our patch is irrelevant in this particular case. In order for a new patch to apply this patch as distributed by the authors, POSIXLY_CORRECT needs to be set so that it will look at the Index: lines. There is no ambiguity, none of the files listed in the +++/ --- entries exist, the only place the path exists is in the Index: header. However, patch-2.5 is not posix compatable enough by default. If it looks at Index as a last resort (as the standard says), then it can apply the patch. But by default, it can't, even though the diff itself is posix compliant. > What you mean exactly by the words "which doesn't work when new files are > added"? Could you send me some examples? An unpatched cvs (as used by the squid folks) produces this with 'cvs diff -cN': Index: lib/malloc-2.6.4.c =================================================================== RCS file: malloc-2.6.4.c diff -N malloc-2.6.4.c *** /dev/null Wed Jan 7 22:18:28 1998 --- /var/tmp/aaaa005bq Wed Jan 7 22:19:29 1998 *************** However, It seems that this was generated on an older version of cvs than is in our tree. I was mistaken, your tweak to cvs generates a real pathname in the +++ line rather than the bogus /var/tmp/aaaa*. Sorry for the false alarm. Cheers, -Peter