Axel Thimm <Axel.Thimm@xxxxxxxxxx> writes: >> >> >> When a package/daemon writes files and/or reads files which are >> >> >> protected by file permissions, it is a good candidate for fixed >> >> >> uids. > ... > Ok, let's bite. Please name a couple that would be candiates for doing > so. * The *milt* and defang users; they are using unix sockets shared between several vservers. * fnord (http server), twiki, tclhttpd sounds like a candidate I do not know the other applications; but I can not exclude that there exist a setup where they might share network resources. 'fedora-usermgmt' deals both with users who must have predictable uids, who need predictable uids under some circumstances and who never need predictable uids (although: say never "never"). Its flaws (causes lot of discussion, is proprietary, nobody else uses it) are of non-technical nature and negligible and I do not see why it should not be used for all users. > If there are *real* use cases for sharing data across machines the > packager should request a fixed uid/gid. I am really in doubt that the remaining free entries < 100 are enough. And when can a uid be reserved there? When there is at least 1 installation which needs a predictable uid, when there are 10, 100, 1000? Enrico
Attachment:
pgpIQQ7azjA1k.pgp
Description: PGP signature
-- Fedora-maintainers mailing list Fedora-maintainers@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers
-- Fedora-maintainers-readonly mailing list Fedora-maintainers-readonly@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers-readonly
