Re: Packages that failed to build with Python 3.10 (and what to do)

2021-06-15 Thread Miro Hrončok
On 08. 06. 21 15:58, Tomas Hrnciar wrote: Chances are, you already got an automated F35FailsToInstall bugzilla from Miro, that your package fails to install. It would be really helpful if you could find the missing dependency and mark the bugzilla for your package dependingon the bugzilla for

Re: Draft of New Python Packaging Guidelines

2021-06-15 Thread Neal Gompa
On Tue, Jun 15, 2021 at 5:35 AM Miro Hrončok wrote: > > On 14. 06. 21 7:31, Zbigniew Jędrzejewski-Szmek wrote: > >> BuildRequires: %{py3_dist pytest} > > Does it make sense to recommend py3_dist here? python3dist(pytest) is > > not more complex but can be fed to 'dnf install' directly, so in the

Re: Draft of New Python Packaging Guidelines

2021-06-15 Thread Miro Hrončok
On 14. 06. 21 7:31, Zbigniew Jędrzejewski-Szmek wrote: BuildRequires: %{py3_dist pytest} Does it make sense to recommend py3_dist here? python3dist(pytest) is not more complex but can be fed to 'dnf install' directly, so in the end it's more flexible. I always find the macro more obfuscation

Re: Draft of New Python Packaging Guidelines

2021-06-15 Thread Miro Hrončok
On 14. 06. 21 7:31, Zbigniew Jędrzejewski-Szmek wrote: For example, if a user runs pip install requests[security] Please quote 'requests[security]' in the command. [] are special to the shell. Done. -- Miro Hrončok -- Phone: +420777974800 IRC: mhroncok

Re: Draft of New Python Packaging Guidelines

2021-06-15 Thread Miro Hrončok
On 14. 06. 21 7:57, Zbigniew Jędrzejewski-Szmek wrote: What about making this conditional: Requires: (pyproject-rpm-macros if rpm-build) You can't really*use* the macros without rpmbuild, and one less hard dependency is always good. With the conditional dep, pretty much any real system