On Wed, May 05, 2004 at 06:17:11PM -0400, Michael Jennings wrote: > On Wednesday, 05 May 2004, at 15:13:00 (-0700), > Steve Beattie wrote: > > > just pointing out that the tool makes it possible for this to > > happen. > > No, you were saying that the tool shouldn't have those options > *because* it's possible for that to happen. Big difference. If you wish to be pedantic, I argue that the -bs and -bb options don't need to exist because they make it easier for a package distributer to ship a broken srpm *and* they add very little utility -- essentially an optimization over -ba; i.e. it's a bad tradeoff. The options --no-deps and --force, while dangererous to an unknowledgable user, serve important purposes and thus add utility. The rest of the -b options, combined with --short-circuit, make the write-test-debug loop go more quickly for packagers, again adding utility. Complexity that gives power is good. Needless complexity is not. Also note that rpm already has slightly rounded edges in this area, otherwise -bb --short-circuit (or something similar) would be possible. Obviously, you feel that -bb and -bs add worthwhile utility to rpm. I disagree. However, I don't feel all that strongly about it, it was a minor niggle to me. I don't really expect the options to get deprecated and will let the issue drop here. On Wed, May 05, 2004 at 06:15:35PM -0400, Michael Jennings wrote: > On Wednesday, 05 May 2004, at 15:13:00 (-0700), > Steve Beattie wrote: > > > Except that I've seen *Red Hat* ship srpms on multiple occasions > > that were broken in this way -- there was no way the supposed source > > rpm would build, and yet they had a binary rpm of the same > > version. It's not the user I'm concerned about, it's the > > distributer. > > I don't recall excluding any particular entity from my previous > statement. Fairy nuff. But consider that the entity that should know how to use the tool the best (whether it actually does or not) still has managed to shoot itself in the foot with these particular options on multiple occasions. -- Steve Beattie Don't trust programmers? <steve@xxxxxxxxx> Complete StackGuard distro at http://NxNW.org/~steve/ immunix.org http://www.sardonix.org -- Audit code, earn respect.
Attachment:
pgp00064.pgp
Description: PGP signature
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
