conary should build out of the box, with plain contexts. in the past we were building conary against old groups, so that updates/migrations, worked flawlessly from setups with oldish glibc releases but that has been worked on strictly by runtime deps/reqs mangling.
meanwhile latest conary=@fl:2-devel is not building... in the past i was being reverting 3 upstream commits - '757a858', '2c2ca3e', 'bfbd22a' because with them applied i wasn't been able to get rmake to work at all [rmake was complaining about not being able to find 'suitable contexts' on startup] . *If* we take those reverts off conary would, naturally, build and work nicelly but rmake shouldn't work, as is, on any box with that conary on the rmake host. Michael, can you confirm that this is is still an issue or that it has been fixed en passant by recent conary commits, and if so take a look plz ? [ i don't see any moves in rmake side that would had helped in that front] All the best, António António Meireles -- Lead Developer, The Foresight Linux Project http://www.foresightlinux.org On Tue, Jan 7, 2014 at 9:35 PM, Michael K. Johnson <[email protected]>wrote: > There is a long-standing bug in which a system-model update would > fail with unresolved dependencies that were already satisfied by > installed troves. Michael Tharp recently discovered the reason > for this bug and fixed it. > > We'll want a new conary build. António, do you use special contexts > for fl:2-devel conary builds? I've committed the source update but > did not want to build without knowing for sure whether you build in > an older context for any reason... > > Thanks! > > _______________________________________________ > Foresight-devel mailing list > [email protected] > https://lists.foresightlinux.org/mailman/listinfo/foresight-devel > > _______________________________________________ Foresight-devel mailing list [email protected] https://lists.foresightlinux.org/mailman/listinfo/foresight-devel
