xfs
[Top] [All Lists]

Re: REVIEW: ASCII CI support in xfsprogs

To: "Christoph Hellwig" <hch@xxxxxxxxxxxxx>
Subject: Re: REVIEW: ASCII CI support in xfsprogs
From: "Barry Naujok" <bnaujok@xxxxxxx>
Date: Fri, 02 May 2008 10:13:38 +1000
Cc: "xfs@xxxxxxxxxxx" <xfs@xxxxxxxxxxx>, xfs-dev <xfs-dev@xxxxxxx>
In-reply-to: <20080501131725.GB7435@infradead.org>
Organization: SGI
References: <op.uagkf9jl3jf8g2@pc-bnaujok.melbourne.sgi.com> <20080501131725.GB7435@infradead.org>
Sender: xfs-bounce@xxxxxxxxxxx
User-agent: Opera Mail/9.24 (Win32)
On Thu, 01 May 2008 23:17:25 +1000, Christoph Hellwig <hch@xxxxxxxxxxxxx> wrote:

On Thu, May 01, 2008 at 02:37:11PM +1000, Barry Naujok wrote:
I've reworked xfsprogs userspace tools from scratch with ASCII CI support
only. Much simpler than previous patches with Unicode CI.


One issue that sort of bugs me a bit is I've maintained the same output as
the original IRIX mkfs.xfs with CI mode:
naming =version 2 bsize=4096 mixed-case=Y|N


where "N" = CI enabled and "Y" = normal case-senstive filesystem.

Should this be changed to something clearer?

I'd prefer something cleaner. Like case-insensitive=Y|N.

Yes, I was going to do "ascii-ci=Y|N", got to keep it clear in only does CI for A-Z.

Otherwise
the patch looks fine, but I wonder if you shouldn't better backport the
xfs_dir* files from the kernel completely instead of applying the
changes selectively, otherwise we'll get totoally out of sync codebases.

I was thinking of that, but libxfs is pretty out of date anyway from all the cleanup changes that have occured and would like to sync them back at some stage in the near future. The sync would bring those other changes in at that time (Nathan can correct me if I misunderstand the sync process!)

Barry.







<Prev in Thread] Current Thread [Next in Thread>