disabling Haskell i686 packages - devel - Fedora Mailing-Lists

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

disabling Haskell i686 packages

SPDX Statistics - Rust edition

OpenColorIO builds failing due to...

Jens-Ulrik Petersen

Thursday, 18 May 2023 Thu, 18 May '23

3:08 a.m.

We have been building i386 packages for ghc and Haskell for a long time, but I think it is time to stop for Fedora 39. So as part of F39 Haskell updates, I plan to disable 32bit ix86 builds across all the packages: I doubt anyone is still using them anyway. Jens

Attachments:

attachment.html (text/html — 478 bytes)

Reply

Show replies by date

Florian Weimer

Thursday, 18 May Thu, 18 May

6:12 a.m.

* Jens-Ulrik Petersen:

We have been building i386 packages for ghc and Haskell for a long time, but I think it is time to stop for Fedora 39. So as part of F39 Haskell updates, I plan to disable 32bit ix86 builds across all the packages: I doubt anyone is still using them anyway.

I think this will make quite a few packages FTBFS on i686 because they require pandoc at build time. It's not that many—I count ~75 that build-depend on something pandoc-ish without also build-depending on something ghc-ish. The actual number will be lower once we stop building noarch packages on i686 (not yet implemented as far as I know). And there are probably some packages which build-depend indirectly on ghc. Thanks, Florian

Reply

Jens-Ulrik Petersen

Sunday, 21 May Sun, 21 May

10:53 p.m.

Thanks, Florian, very good point. On Thu, May 18, 2023 at 7:12 PM Florian Weimer <fweimer(a)redhat.com> wrote:

I think this will make quite a few packages FTBFS on i686 because they require pandoc at build time. It's not that many—I count ~75 that build-depend on something pandoc-ish without also build-depending on something ghc-ish. The actual number will be lower once we stop building noarch packages on i686 (not yet implemented as far as I know). And there are probably some packages which build-depend indirectly on ghc.

Then I guess it will have to wait then. It is not that big a deal yet: though a few i686 issues are slowly starting to creep in for a few Haskell packages. Jens

Reply

Matthew Miller

Friday, 26 May Fri, 26 May

3:09 p.m.

On Thu, May 18, 2023 at 01:12:00PM +0200, Florian Weimer wrote:

I think this will make quite a few packages FTBFS on i686 because they require pandoc at build time. It's not that many—I count ~75 that

Is there any reason to build docs for i686 packages? (With or without some cleverness that also removes doc-build dependencies without package changes?) -- Matthew Miller <mattdm(a)fedoraproject.org> Fedora Project Leader

Reply

Fabio Valentini

Saturday, 27 May Sat, 27 May

6:11 a.m.

On Fri, May 26, 2023 at 10:10 PM Matthew Miller <mattdm(a)fedoraproject.org> wrote:

On Thu, May 18, 2023 at 01:12:00PM +0200, Florian Weimer wrote: > I think this will make quite a few packages FTBFS on i686 because they > require pandoc at build time. It's not that many—I count ~75 that Is there any reason to build docs for i686 packages? (With or without some cleverness that also removes doc-build dependencies without package changes?)

Turning off docs in i686 builds only would require changes to all affected spec files ... And this doesn't take into account packages that are noarch. Though there might be movement in getting koji to just never schedule noarch builds to run on i686 builders, which would make the problem at least a bit easier to solve. Fabio

Reply

328

days inactive

337

days old

devel@lists.fedoraproject.org

Manage subscription

4 comments

4 participants

tags (0)

participants (4)

Fabio Valentini
Florian Weimer
Jens-Ulrik Petersen
Matthew Miller