Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm Sender: cygwin-owner AT sourceware DOT cygnus DOT com Delivered-To: mailing list cygwin AT sourceware DOT cygnus DOT com Message-Id: <3.0.5.32.19990329182129.00a12a40@pop.ma.ultranet.com> X-Sender: lhall AT pop DOT ma DOT ultranet DOT com X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.5 (32) Date: Mon, 29 Mar 1999 18:21:29 -0500 To: earnie_boyd AT yahoo DOT com, cygwin users From: "Larry Hall (RFK Partners, Inc)" Subject: Re: B20: mv deletes files on error (NT) In-Reply-To: <19990329231516.11977.rocketmail@web115.yahoomail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" At 03:15 PM 3/29/99 -0800, Earnie Boyd wrote: >--- "Larry Hall (RFK Partners, Inc)" wrote: >--8<-- >> One person's feature is another's bug! I guess now I understand what the >> goal of making this change was. I have to say, for the number of times that >> hard links are used (especially in Win 9x), I can't say that I think the >> change is for the better. Still, I have no problem with adding functionality >> to cygwin (in general or in this case specifically) so long as it doesn't >> remove functionality. I think if we need to make sure we can use hard links, > >> we need to be able to do this without sacrificing something that is >> functionally useful. At the very least, if nothing else can be found, we >> need a switch, even though I hate this idea in general. Mind you, I'm not >> against adding the ability to make use of NTFS hard links but why do we need >> to support this at the expense of existing functionality? I think this >> change needs to be revisited. >--8<-- > >It's easy enough to modify the code for mv.c. I know this (moving from one CASE >to the other case) works on NT but how about 9x? > >=== >- \\||// >-------------------o0O0--Earnie--0O0o------------------- >-- earnie_boyd AT yahoo DOT com -- >----------------------ooo0O--O0ooo---------------------- > >_________________________________________________________ Why wouldn't it? According to Corinna, the change that made the moves work in the first place was based on the hash calculation for the file's inode. This shouldn't be any different on NT than 9x. The ease of modifying one file's code or another's is not my point... Larry Hall lhall AT rfk DOT com RFK Partners, Inc. (781) 239-1053 8 Grove Street (781) 239-1655 Wellesley, MA, 02482-7797 http://www.rfk.com -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com