On Thu, 6 Oct 2022 14:38:38 +0200 > This is good point. I have already forget the details. So if there > was embedded just the right amount of information in the main data, > we would not need the full list (and lazy loading). Currently, the > data contains e.g. /usr/bin/*, which is useful for installing > `/usr/bin/foo`. But we know that in the repository, there is RPM with > `Requires: /some/strange/path`. Therefore during creating the > repository metadata, we could look for RPM providing this path and > include it into the main metadata. This would blow the metadata a > bit, but it would allow to not care about the full file list at all. > Of course that would mean `dnf install /some/random/path` won't work > universally, but 1) I don't think this is widely used for random > stuff 2) it is easy to detect and download the full file list instead > if needed. > > Should I open request against createrepo_c? Fixing it in the repo metadata would only work if the dependency and its provider are in the same repository. So for instance if there was a dependency on /some/random/path in the main Fedora repo and the package containing that file was updated, the createrepo run for the updates repository wouldn't know about the dependency and wouldn't add the provide. That's before even considering third-party repos. Paul. _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
