Hi Joe,
Joseph J. VLcek wrote:
> Great Jan!
>
> I have one comment below.
>
>>
>>>
>>>
>>> usr/src/cmd/installadm/setup-sparc.sh
>>> -------------------------------------------------------------
>>>
>>> 97 dns_servers=`$GREP "$NAMESERVER_STRING"
>>> $RESOLV_CONF_FILE 2>/dev/null | \
>>> 98 $AWK '{printf("%s ", $2)}'`
>>> 99 dns_domain=`$GREP "$DOMAIN_STRING" $RESOLV_CONF_FILE
>>> 2>/dev/null | \
>>> 100 $HEAD -1 | $AWK '{printf("%s", $2)}'
>>>
>>> This code could benefit from the ksh93 (/sbin/sh) built in extended
>>> regular expression pattern matching. For examples see:
>>>
>>> http://installzone-wiki.central.sun.com/wiki/index.php/Ksh93_Tips#Use_built_in_extended_regular_expression_pattern_matching
>>>
>>>
>>
>> I am not sure if I could take advantage of this feature in this
>> particular
>> case, as I need to extract particular line from file, not only find
>> the match.
>> Could it be also used in such scenario ?
>
>
> This would work but in this case I am not convinced it is actually
> better than what you already have.
>
> $
> dn=""
>
>
> $
> line=""
>
>
> $ line=`$GREP "${DOMAIN_STRING}" ${RESOLV_CONF_FILE}
> 2>/dev/null` $ if [[ "${line}" ==
> ~(E)^domain.* ]]; then ^Jdn="${line#domain }"^Jfi
> $ echo
> $dn
>
>
> sfbay.sun.com
>
> So perhaps you should leave it as it is.
Looking at the implementation, I concur with you that original
version is better, as I think it is more readable.
Thanks for digging into this !
Jan
