Date: Wed, 31 Mar 2004 00:27:44 +0400 From: Andrey Chernov <ache@nagual.pp.ru> To: "Brian F. Feldman" <green@FreeBSD.ORG> Cc: Dag-Erling Smorgrav <des@FreeBSD.ORG> Subject: Re: cvs commit: src/etc/mtree BSD.local.dist BSD.usr.dist Message-ID: <20040330202743.GA3407@nagual.pp.ru> In-Reply-To: <200403301627.i2UGRx9q030892@green.homeunix.org> References: <20040330155233.GA71146@nagual.pp.ru> <200403301627.i2UGRx9q030892@green.homeunix.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Mar 30, 2004 at 11:27:58AM -0500, Brian F. Feldman wrote: > Okay, now I understand what the situation is that would make this not work > well. Why not just warn and ignore it, then? Warnings can be too verbose, in case too many dirs affected, so they can hide real errors. In opposite situation, i.e. no warnings at all, some errors could be skipped. So this thing needs a proper fix, not just a workaround. > The proper fix would be to > resolve, at the end of reading the mtree spec file, any type=link (when -L) > and fail if they cannot be dereferenced internally. It can be kinda of two-passes mode, if -L is specified: 1st pass checks everything excepting type=link, in -L mode 2nd pass checks type=link only, in -P mode And remember that both workaround and proper fix will make us very different from any other *BSD. -- Andrey Chernov | http://ache.pp.ru/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040330202743.GA3407>