On 3/13/13 12:43 PM, Rich Johnston wrote:
> On 03/13/2013 09:57 AM, Eric Sandeen wrote:
>> SCRATCH_DEV_POOL processing actually takes the first
>> device out for SCRATCH_DEV and leaves the rest in
>> SCRATCH_DEV_POOL.
>>
>> I'm not totally sold on that behavior, but for now,
>> at least don't populate SCRATCH_DEV_POOL with newlines.
>
>>
>> Cc: linux-btrfs@xxxxxxxxxxxxxxx
>> Signed-off-by: Eric Sandeen <sandeen@xxxxxxxxxx>
>> ---
>> common.config | 2 +-
>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/common.config b/common.config
>> index ed0f44c..5ac58bf 100644
>> --- a/common.config
>> +++ b/common.config
>> @@ -258,7 +258,7 @@ if [ ! -z "$SCRATCH_DEV_POOL" ]; then
>> exit 1
>> fi
>> SCRATCH_DEV=`echo $SCRATCH_DEV_POOL | awk '{print $1}'`
>> - SCRATCH_DEV_POOL=`echo $SCRATCH_DEV_POOL | awk '{ for (i = 2; i <= NF; i++) print $i}'`
>> + SCRATCH_DEV_POOL=`echo $SCRATCH_DEV_POOL | awk '{ ORS=" "; for (i = 2; i <= NF; i++) print $i}'`
>> fi
>>
>> echo $SCRATCH_DEV | grep -q ":" > /dev/null 2>&1
>>
>
> Is the purpose of this patch is so future improvements will be able to parse/use SCRATCH_DEV_POOL easier from within bash?
Yep, and just generally since it's used as an argument, feeding a string w/ newlines to any command is just a bit odd.
-Eric
> Looks good Eric.
>
> Reviewed-by: <rjohnston@xxxxxxx>
>
> Regards
> --Rich
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
