| To: | Christoph Hellwig <hch@xxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: [PATCH 3/7] Test basic quota enforcement |
| From: | Jan Kara <jack@xxxxxxx> |
| Date: | Thu, 20 May 2010 19:21:15 +0200 |
| Cc: | Jan Kara <jack@xxxxxxx>, xfs@xxxxxxxxxxx |
| In-reply-to: | <20100520151130.GC27124@xxxxxxxxxxxxx> |
| References: | <1274276289-8839-1-git-send-email-jack@xxxxxxx> <1274276289-8839-4-git-send-email-jack@xxxxxxx> <20100520151130.GC27124@xxxxxxxxxxxxx> |
| User-agent: | Mutt/1.5.20 (2009-06-14) |
On Thu 20-05-10 11:11:30, Christoph Hellwig wrote:
> The golden output contains "/local/test1/file5" which appaears
> to be in your test fs. You need to do something like
>
> sed -e "s,$SCRATCH_MNT,SCRATCH_MNT,g"
>
> or
>
> sed -e "s,$T�T_MNT,TEST_MNT,g"
>
> to make the test portable. This probably applies to the other tests
> as well.
I have filter_scratch function for this in all tests. I just forgot to
filter output of touch(1) through it. It should be fixed now.
BTW: What is the purpose of 'status' variable? I set it to 1 if I want
the test to fail because diff found differences between accounted and
real usage. Is this correct? Actually, diff will also print the differences
to stdout so setting status probably shouldn't be needed. So is it preferred
to set status to 1 or leave it 0?
Honza
--
Jan Kara <jack@xxxxxxx>
SUSE Labs, CR
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [PATCH 1/7] xfstests: test extent size hints, Jan Kara |
|---|---|
| Next by Date: | Re: [PATCH 07/12] xfs: Improve scalability of busy extent tracking, Alex Elder |
| Previous by Thread: | Re: [PATCH 3/7] Test basic quota enforcement, Christoph Hellwig |
| Next by Thread: | [PATCH 1/7] xfstests: test extent size hints, Jan Kara |
| Indexes: | [Date] [Thread] [Top] [All Lists] |