New pkg-message[.in] guideline idea

Doug Barton dougb at
Wed Jun 23 19:43:26 UTC 2010

On 06/23/10 12:19, Philip M. Gollucci wrote:
> On 06/23/10 19:15, Doug Barton wrote:
>>> I think we need 2 things:
>>>> 1) All pkg-message whether .in or not go
>>>>    through the 'sed' that SUB_LIST/PLIST_SUB do.
>> I like this idea better than forcing them all to be in /files, and I
>> can't see any reason not to do it.
> The other benefit is you can carry over the PLIST logic with @comment to
> conditionalize lines in the message and thus eliminate all that crap
> from Makefiles.

Yes, simpler is better. :)

>>>> 2) You collect them in to /var/db/pkg and loop
>>>>    and display at end.
>>>> As a consequence all formatting should be removed from the individual
>>>> pkg-message[.in] files and added in #2.
>> I'm ambivalent about this. My first thought was that the formatting
>> should happen in step #1. What portmaster does is build each port one at
>> a time, and it makes a note if a port has a pkg-message. Then it does
>> what you suggest in #2 by cat'ing them all to $PAGER. However, thinking
>> more about it I could see how not doing the formatting until step 2
>> could work, so however it turns out to be "easiest" should be fine.
> I believe we are agreeing. I didn't say anything about
> portmaster/portupgrade.  The pkg/make infrastructure needs to do it at
> the end so that it works via pkg_add, make, or portmaster et al.

Yes, we are agreeing. I was simply adding my perspective as a tool
author to your proposal. I also agree that the ports infrastructure
itself should be as robust as possible, and include features like this
whenever it can.




	... and that's just a little bit of history repeating.
			-- Propellerheads

	Improve the effectiveness of your Internet presence with
	a domain name makeover!

More information about the freebsd-ports mailing list