>>>>> "Phillip" == Phillip Susi <ps...@ubuntu.com> writes:
Phillip> Not having the .pc file and headers etc in the -dev package Phillip> would prevent the build of anything with a decent Phillip> pkg-config enabled build system, so that could work with a Phillip> tweak to the policy to allow it. That would prevent the Phillip> building of broken packages, though it's a bit of throwing Phillip> the baby out with the bath water. It would be preferable Phillip> to allow the build, but with a warning and a strict == Phillip> dependency instead of the >= one. I actually think that an = dependency between source packages is a really bad idea. (And I believe it's forbidden by policy.) It's less of an issue for Ubuntu, but for Debian it's going to block testing migrations of the depended package until the depending package updates. Especially if packages are maintained by different folks this tends to be a bad idea. It may be in the specific case of ntfsprogs the upstream changes infrequently enough that it's acceptable. In the general case though I'd be very concerned about that approach. More generally, this is all about stability. The question is whether an interface is stable enough to be used outside of a narrow scope. If a maintainer is willing to make an interface stable enough then including a -dev package is reasonable. If not, then it's better not to have the interface public beyond the scope of the source package. There's negotiations about stable enough. Clearly we're not going to say that nothing outside ntfsprogs gets to read ntfs filesystems. --Sam -- To UNSUBSCRIBE, email to debian-ctte-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/tslbobdiswz....@mit.edu