xfs
[Top] [All Lists]

Re: [PATCH] xfstests: fix 108 golden output

To: aelder@xxxxxxx
Subject: Re: [PATCH] xfstests: fix 108 golden output
From: Boris Ranto <branto@xxxxxxxxxx>
Date: Fri, 10 Dec 2010 11:06:49 +0100
Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
In-reply-to: <1291929579.13355.30.camel@doink>
References: <20101110125855.GA18357@xxxxxxxxxxxxx> <1291929579.13355.30.camel@doink>
On Thu, 2010-12-09 at 15:19 -0600, Alex Elder wrote:
> On Wed, 2010-11-10 at 07:58 -0500, Christoph Hellwig wrote:
> > The new common scratch mount filter produces slightly different
> > whitespaces than the old home grown filter, so adjust the golden
> > output.
> > 
> > Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> 
> I found out why this is happening.  I get no
> such error, and Dave Chinner said he wasn't seeing
> it either.
> 
> xfs_quota is formatting its output such that the
> second and later columns always align at a certain
> offset from the beginning of the line.  As a result,
> path names assigned as SCRATCH_DEV of different
> lengths will produce different (filtered) output,
> with more or fewer blanks between SCRATCH_DEV and
> the usage number.
> 
> Boris Ranto has proposed a change that would
> add the "-b" flag to diff when checking output
> against golden output.  That should fix it, but
> it's possible we *want* differences in white
> space to matter in some cases.  Have to look
> into this.
> 
>                                       -Alex
> 

The change was designed to add the "-b" flag only for this test case.
Any other test case could also modify its diff options through
DIFF_OPTIONS variable in its config file ($seq.config).

-Boris

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