Re: Ruby 3.0

[Vít Ondruch]

Dne 17. 12. 20 v 3:16 Pavel Valena napsal(a):

----- Original Message -----

From: "Jaroslav Prokop" <jar.pro...@volny.cz>
To: ruby-sig@lists.fedoraproject.org
Sent: Wednesday, December 16, 2020 7:05:03 PM
Subject: Re: Ruby 3.0

Hi,

WEBrick is default in some applications, IIRC - e.g Sinatra [0] or
Jekyll - as it was a server in standard lib that came with Ruby.

I am not sure 100% there, as it could just be rack default to search for
WEBrick, either way, it'll be better to watch out for
that.

Hello,

I think there's no need for anything else than simple declaration of dependency 
in Gemfile / gemspec. Which should be there in first place. Webrick is simply 
stand-alone gem now, right?

Well my point is that previously, when WEBRick was part of Ruby, it was natural to used it. When it is not part of Ruby, upstreams will have to deal with it and in case of updating Gemfile / gemspec, why not use different server, such as Puma? This should be probably super easy for Rack based projects. There will certainly be different projects, which might be using WEBRick directly. But also in this case, upstreams might decide to use different web server when they have to provide some fixes anyway. But I guess the only way to know what fails due to removed WEBRick is to update the Copr and rebuild everything.

Quickly scrolling through the code of Sinatra and Jekyll it seems like
WEBrick is used in tests as well, so it's a question if it will be just
a matter of
adding require (maybe Suggests?) to gemspec or if BuildRequires will
need appending too.

Yes, there'll be more of fixing like that for Ruby, and gems upstreams (f.e. 
`rexml` require). If you want to jump in on the train, here're some failed 
pre-builds:
   https://copr.fedorainfracloud.org/coprs/pvalena/rubygems-testing/builds/

Please create a PR to both gem upstream and Fedora package if you're able to 
fix some issue.

To answer your question, adjustments to packages might be needed and
packaging WEBrick might be worth it, but better inspection is needed
from maintainers of packages that package said software.

Jarek


[0]
https://github.com/sinatra/sinatra/blob/a9649b4f18a9059de3161906c9c6e95ec06fdef9/lib/sinatra/base.rb#L1827


On 16/12/2020 18:39, Vít Ondruch wrote:

Hi,

Another snapshot is available in private-ruby-3.0 branch and the build
is running here:

FYI it's not synced with PR#70.

https://github.com/ruby/ruby/pull/3907


Thx


Vít

If this still applies to the Copr context, could you please update your Copr and test how does it look with eventmachine? It seems the problematic patch was reverted upstream:

https://koji.fedoraproject.org/koji/taskinfo?taskID=57585626

I'll run COPR build in my COPR as well.

Most notable change is removal of WEBRick from Ruby. I am not
completely sure how much disruption this could cause in Fedora. I
wonder if it is worth of packaging it as a separate package. Thoughts?

Vít


P.S. @jaruga + @pvalena thx for handling the GCC11 issues.

Regards,

OpenPGP_0x0CE09EE79917B87C.asc
Description: application/pgp-keys

OpenPGP_signature
Description: OpenPGP digital signature

_______________________________________________
ruby-sig mailing list -- ruby-sig@lists.fedoraproject.org
To unsubscribe send an email to ruby-sig-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/ruby-sig@lists.fedoraproject.org