From mboxrd@z Thu Jan 1 00:00:00 1970 From: Darren Hart Subject: Re: [PATCH 0/3 V2] ext2fsprogs: Symlink support Date: Fri, 04 Jan 2013 16:26:43 -0800 Message-ID: <50E77343.4010807@infradead.org> References: <1357329660-28639-1-git-send-email-dvhart@infradead.org> <20130105001521.GD6501@thunk.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: linux-ext4@vger.kernel.org, adilger@dilger.ca, sgw@linux.intel.com, darrick.wong@oracle.com To: Theodore Ts'o Return-path: Received: from mga11.intel.com ([192.55.52.93]:29624 "EHLO mga11.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754934Ab3AEA0o (ORCPT ); Fri, 4 Jan 2013 19:26:44 -0500 In-Reply-To: <20130105001521.GD6501@thunk.org> Sender: linux-ext4-owner@vger.kernel.org List-ID: On 01/04/2013 04:15 PM, Theodore Ts'o wrote: > On Fri, Jan 04, 2013 at 12:00:57PM -0800, Darren Hart wrote: >> The following series adds symlink support the libext2fs and adds the command to >> debugfs. Basic testing demonstrates functional duplication of a small test >> directory including fastlinks and slowlinks in both the root and a single >> subdirectory. > > One thing I'd really appreciate is if you could write a regression > test for this new code. You can use tests/d_loaddump and > tests/f_dup4/script as a model; I'd suggest calling it d_symlink, and > it would be good if it created a short symlink and a long symlink, and > that the script verified that you get the expected output when you > stat the small and long symlinks using debugfs. Then run e2fsck on > the file system and make sure the file system is consistent. > > Note the use of the debugfs commands: > > set_current_time 200704102100 > set_super_value lastcheck 0 > set_super_value hash_seed null > set_super_value mkfs_time 0 > > and the environment variable: > > E2FSCK_TIME=200704102100 > > in f_dup4's script. These are used to control various variable > (random or time-specific) values such that so that you can use a > simple comparison of the output of programs such as debugfs, mke2fs, > and dumpe2fs using "diff". > > Thanks again!! > > - Ted > > P.S. Actually, while you're at it, if you could also have the script > create some block and character device files using the debugfs mknod > command, and perhaps call the test d_create_spec_files instead of > d_symlink, that would be great! (Since I don't think those functions > currently any good test coverage.) Ah, great. I'm all for built-in tests. I'll get on it next week. Thanks Ted. -- Darren Hart Intel Open Source Technology Center