On Tue, Jun 08, 1999 at 08:55:17AM -0700, Brent Fulgham wrote: > I could be very, very wrong, but I am going to go out on a limb and > say that the hurd has absolutely nothing to do with those linux-kernel > sources. They are most likely links from potato that are mirrored in > our hurd directories because they are "*-all.deb" style packages that > are supposed to work for all architectures. This worked fine when > ports meant Linux to -alpha, -i386, -mips, etc., but is not working > out so well for us with our "hurd-i386" architecture.
Right. > The only proviso here is that we are using some Linux kernel source > for drivers in the hurd.... But I would advise against downloading > these sources. Exactly :) So they are useful to some people, although not many. > Marcus: Do we have a way to block out some of these types of > packages? NO! That's indeed a serious problem, but NOT for linux kernel sources, but for harmful packages like base/makedev!!! There has been a "naive" proposal by me, but Guy Maor wants more detailed information for this before implemnting it. Then a discussion happened, which circled around more sophisticated solutions. Those would need probably more time to implement, but they are really much better. I am afraid, with the speed busy people like Guy and Ian J. can process feature wishlist items, we are not going to have a proper solution before DPKG v2, which should have a proper architecture understanding implemented. Maybe we can do some adjustements to the arechive (seperation of binary-all for hurd and linux and common) before. As long as only very few packages "appear" to be affected, and as long only a few people realize there actually is a problem, and as long as everyone is so busy, we will have to live with it. Of course, I spoke with the makedev maintainer, and he would be willing to modify the package. But making it binary-arch would be circumventing the problem without fixing it, and I don't want to do this. Instead, hurd does currently Conflict with makedev, which prevents damage to the system. Gordon was working on a proper proposal, too. I would like to have this proposal or some derived from it be implemented in DPKG v2. Thanks, Marcus -- `Rhubarb is no Egyptian god.' Debian http://www.debian.org finger brinkmd@ Marcus Brinkmann GNU http://www.gnu.org master.debian.org [EMAIL PROTECTED] for public PGP Key http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/ PGP Key ID 36E7CD09
