https://bugzilla.redhat.com/show_bug.cgi?id=2468312



--- Comment #4 from Cristian Le <[email protected]> ---
> > - Whenever possible, please use `%autorelease`/`%autochangelog` [1], it 
> > helps other contributors, particularly with rebasing PRs
> I don't think the upstream git repo is compatible with that.

The upstream git repo does not play a role into this. The only fields that are
affected are `Release` and `%changelog` which are downstream only. The git
connection is with the dist-git, and even that one is not necessary.

> But we surely don't need to include license files from every package that we 
> link to?

Yes, that is not done in other packages either. Technically you need to add if
any of the other symbols are re-compiled from the header file, e.g. C++
templates, but even that case is hard to track and is often overlooked.

> I need access to the packager group to run scratch builds to see which of 
> those are already included in the standard build environment and can be 
> removed.

Firstly you can use copr [1] to do the builds which are 90% the same as koji
builds. `mock` would also do that. Secondly that is not what I was referring
to. If it is a dependency defined from cmake requirement then yes explicitly
define it. The one that made me raise my eyebrow was having both `make` and
`ninja-build`.

In the current list, `ninja-build` and `rpmdevtools` shold be dropped.

> This package Obsoletes every version of that old package.

You still MUST specify the specific version to obsolete even if that is the
case. You can use a version higher to not have to sync with the release.

> Consider an end-user who wants to read the doc files. Hardly any end-users 
> are familiar with either rtf or markdown.

I beg a differ. Using the original markdown and rtf is common standard [2] and
markdown is made to be human readable. In contrast ooffice conversion is not
used in other packages [3]. The man page should be the recommended user-facing
documentation.

This is not a blocking issue for me though, so do as you wish.

> If we use `%cmake` rather than `%cmake -DNO_GTEST=ON`, it does a git clone to 
> download googletest source code

Please highlight this because there are standard workarounds. Please request
upstream to use `FetchContent` instead of the manual build snippet in [4]. This
is the modern CMake approach that is packager friendly and will work with
`gtest-devel` out-of-the-box. You may ping me on the upstream issue (@LecrisUT)
or point them to a reference like this one [5]

[1]: https://copr.fedorainfracloud.org/coprs/
[2]:
https://sourcegraph.com/search?q=context:global+repo:src.fedoraproject.org+README.md&patternType=keyword&sm=0
[3]:
https://sourcegraph.com/search?q=context:global+repo:src.fedoraproject.org+ooffice&patternType=keyword&sm=0
[4]:
https://github.com/pwsafe/pwsafe/blob/a2f591079e7d889a487c59d00e497549401ea88d/CMakeLists.txt#L225-L251
[5]:
https://github.com/spglib/spglib/blob/12355c77fb7c505a55f52cae36341d73b781a065/test/CMakeLists.txt#L87-L102


-- 
You are receiving this mail because:
You are always notified about changes to this product and component
You are on the CC list for the bug.
https://bugzilla.redhat.com/show_bug.cgi?id=2468312

Report this comment as SPAM: 
https://bugzilla.redhat.com/enter_bug.cgi?product=Bugzilla&format=report-spam&short_desc=Report%20of%20Bug%202468312%23c4

-- 
_______________________________________________
package-review mailing list -- [email protected]
To unsubscribe send an email to [email protected]
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/[email protected]
Do not reply to spam, report it: 
https://forge.fedoraproject.org/infra/tickets/issues/new

Reply via email to