On Mon, Jul 15, 2024 at 08:55:03AM -0400, Stephen Gallagher wrote:
> This seems like overkill. Wouldn't the simplest valid installability
> test just be to test whether each subpackage *individually* could be
> installed?

That's a really nice idea.

> If we have 20 subpackages, just launch 20 separate minimal containers
> and see if `dnf install subpackageN` succeeds. Then it doesn't matter
> if there are conflicts; we know that at least installing that package
> directly will work. (Dependency resolution may pull in other
> subpackages of course, which is proving that it works properly.)

I'm not sure if we actually want a container. Because if it's a container,
then we need _some_ packages inside. But that creates a problem for
some packages like cannot be just installed, but need to be swapped
with other packages. (systemd-standalone-*, coreutils-single, etc.)
I think it'd be more reliable to do something like
  dnf install --enablerepo=/path/to/repo/with/updates 
--sysroot=/var/tmp/inst-package1 /path/to/repo/with/updates/package1.rpm
  dnf install --enablerepo=/path/to/repo/with/updates 
--sysroot=/var/tmp/inst-package1 /path/to/repo/with/updates/package2.rpm
  ...

And to make this work reliably, the invocation of dnf should be
wrapped in 'bwrap' to set up /dev, /proc for the invocation.

This should be quick and more reliable than the current tests,
even with no config.

Zbyszek
-- 
_______________________________________________
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org
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@lists.fedoraproject.org
Do not reply to spam, report it: 
https://pagure.io/fedora-infrastructure/new_issue

Reply via email to