On Tue, Nov 16, 2004, Ralf S. Engelschall wrote:
> On Tue, Nov 16, 2004, Michael Schloh wrote:
>>   -    ntp_pidfile="@l_prefix@/var/ntp/ntpd.pid"
>>   +    ntp_pidfile=`grep "^pidfile" @l_prefix@/etc/ntp/ntp.conf | sed -e 
>> "s;[^ ^t]*[ ^t]*\(.*\)$;\1;"`
>>
> Unfortunately there are three problems: the "^t" escape for a tab
> character in a sed(1) regex is not sufficiently portable, the
> backslashes in the sed(1) command would have to be escaped against
> the surrounding double-quotes and the capturing regex also would
> capture trailing whitespaces. I see no portable and at the same time
> generic enough way to solve this with sed(1). Instead awk(1) works far
> more reasonable here, hence I've fixed it with the construct: awk '{
> printf("%s", $2); }'
>
Yes, I'd considered these things and preliminary test proved the regex
functional. Adding a dependency to sed(1) as in other packages is less
appealing however, so I like your awk(1) based idea better as well. Thanks
for the improvement.

-- 
Michael Schloh von Bennewitz <[EMAIL PROTECTED]>
Development Team, Operations Northern Europe
Cable & Wireless Telecommunications Services
Tel +49-89-92699-227, Fax +49-89-92699-808

Attachment: pgp66yAiYJZS2.pgp
Description: PGP signature

Reply via email to