[Rpm-maint] [rpm-software-management/rpm] Don't insert rpmlib() dependencies for dynamic buildrequires (#878)

Panu Matilainen notifications at github.com
Fri Oct 4 11:15:53 UTC 2019


> These problems have no solution on SRPM level. Unless the spec file is preprocessed by rpm, no-one knows whether the package will depend on dynamic buildrequires feature or not.

Yes, like I said this is no different from any other data in an src.rpm - it is only valid for the particular environment it was built in. That doesn't mean the data there is at all *useless* (our whole buildsystem bases its operation around it!), one just needs to be careful about interpreting it.

Again the issue here is build versus install. Just like there can be arch-specific buildrequires in the src.rpm headers that still don't prevent the src.rpm to be rebuilt on another arch, it'd be equally correct to have a recorded dependency on rpm *build* capability in the src.rpm. It's just that rpmlib() dependencies are not appropriate for this task as they track installability (for example, if the source rpm used a strange compression in the payload you could not install/unpack the source rpm *at all*), not buildability.

t would be perfectly fine for an src.rpm to carry a dependency indicating that in the environment it was built in, it requires an rpm that can handle dynamic buildrequires, 

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/878#issuecomment-538355914
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rpm.org/pipermail/rpm-maint/attachments/20191004/579cd248/attachment.html>


More information about the Rpm-maint mailing list