On 9/3/16 6:55 AM, Ryan Schmidt wrote: > >> On Sep 3, 2016, at 2:18 AM, [email protected] wrote: >> >> Revision >> 152306 >> Author >> [email protected] >> Date >> 2016-09-03 00:18:15 -0700 (Sat, 03 Sep 2016) >> Log Message >> >> zebra: increment revision to rebuild with icu 55.1 >> Modified Paths >> >> • trunk/dports/databases/zebra/Portfile >> Diff >> >> Modified: trunk/dports/databases/zebra/Portfile (152305 => 152306) >> >> --- trunk/dports/databases/zebra/Portfile 2016-09-03 07:16:34 UTC (rev >> 152305) >> +++ trunk/dports/databases/zebra/Portfile 2016-09-03 07:18:15 UTC (rev >> 152306) >> >> @@ -5,6 +5,7 @@ >> >> >> >> name zebra >> >> version 2.0.59 >> >> +revision 1 > > If zebra uses icu, it should declare a dependency on it. That way, whoever > next updates icu will not forget to revbump zebra if necessary. > > > Yes, this is true for both pazpar2 and zebra. In each case, icu is mentioned in configure.ac but the builds succeeded because of yaz's dependency on icu. I only noticed that these binary archives were broken because I installed all the dependents of yaz before updating. Shows you can't rely on explicit dependencies alone when deciding what to revbump.
Dependencies added in r152316. Thanks for the reminder. _______________________________________________ macports-dev mailing list [email protected] https://lists.macosforge.org/mailman/listinfo/macports-dev
