Erlang busted on i386 on rawhide... workaround for erlang-based packages?

[Martin Langhoff]

Hi folks,

trying to upload a new release of elixir, I am butting into something that
looks and smells a lot like BZ#1240487 -- the (noarch) rpm builds nicely on
x86_64 but the erlang runtime segfaults every time on i386.

This FTBS repros for me on koji and under mockbuild locally (x66_64 f23).

 - What's best strategy to bypass this and get it built on x86_64? Setting
an arch on it seems ugly...

 - Anything we can do to diag?

cheers,


m
-- 
 martin.langhoff(a)gmail.com
 -  ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 ~ http://docs.moodle.org/en/User:Martin_Langhoff

[Martin Langhoff]

Here's a koji build showing FTBFS on i686
http://koji.fedoraproject.org/koji/taskinfo?taskID=12918486

I get the exact same result on my F23/x86_64 laptop with

  fedpkg --dist fedora-rawhide mockbuild --root
/etc/mock/fedora-rawhide-i386.cfg

failure happens on the same compile step.



m

On Tue, Feb 9, 2016 at 7:10 PM, Martin Langhoff <martin.langhoff(a)gmail.com&gt;
wrote:

>  Hi folks,
>
>  trying to upload a new release of elixir, I am butting into something that
>  looks and smells a lot like BZ#1240487 -- the (noarch) rpm builds nicely on
>  x86_64 but the erlang runtime segfaults every time on i386.
>
>  This FTBS repros for me on koji and under mockbuild locally (x66_64 f23).
>
>   - What's best strategy to bypass this and get it built on x86_64? Setting
>  an arch on it seems ugly...
>
>   - Anything we can do to diag?
>
>  cheers,
>
>
>  m
>  --
>   martin.langhoff(a)gmail.com
>   -  ask interesting questions
>   - don't get distracted with shiny stuff  - working code first
>   ~ http://docs.moodle.org/en/User:Martin_Langhoff
>  


-- 
 martin.langhoff(a)gmail.com
 -  ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 ~ http://docs.moodle.org/en/User:Martin_Langhoff

[Peter Robinson]

On Wed, Feb 10, 2016 at 12:10 AM, Martin Langhoff
<martin.langhoff(a)gmail.com&gt; wrote:
>  Hi folks,
>
>  trying to upload a new release of elixir, I am butting into something that
>  looks and smells a lot like BZ#1240487 -- the (noarch) rpm builds nicely on
>  x86_64 but the erlang runtime segfaults every time on i386.
>
>  This FTBS repros for me on koji and under mockbuild locally (x66_64 f23).
>
>   - What's best strategy to bypass this and get it built on x86_64? Setting
>  an arch on it seems ugly... 
The only strategy is to fix the issue that's causing the problem on i686.

Peter

[Peter Lemenkov]

Hello All!
Yes, that's a known issue I'm trying to address now. I've got a
possible fix from upstream developers, so things are getting better.


2016-02-10 1:10 GMT+01:00 Martin Langhoff <martin.langhoff(a)gmail.com&gt;:
>  Hi folks,
>
>  trying to upload a new release of elixir, I am butting into something that
>  looks and smells a lot like BZ#1240487 -- the (noarch) rpm builds nicely on
>  x86_64 but the erlang runtime segfaults every time on i386.
>
>  This FTBS repros for me on koji and under mockbuild locally (x66_64 f23).
>
>   - What's best strategy to bypass this and get it built on x86_64? Setting
>  an arch on it seems ugly...
>
>   - Anything we can do to diag?
>
>  cheers,
>
>
>  m
>  --
>   martin.langhoff(a)gmail.com
>   -  ask interesting questions
>   - don't get distracted with shiny stuff  - working code first
>   ~ http://docs.moodle.org/en/User:Martin_Langhoff 


-- 
With best regards, Peter Lemenkov.

[Martin Langhoff]

Hi Peter(s),

On Wed, Feb 10, 2016 at 5:37 AM, Peter Lemenkov <lemenkov(a)gmail.com&gt; wrote:
>  Yes, that's a known issue I'm trying to address now. I've
> got a
>  possible fix from upstream developers, so things are getting better. 
That's great to hear. Am I tracking the right BZ# with  BZ#1240487?

cheers,



m
-- 
 martin.langhoff(a)gmail.com
 -  ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 ~ http://docs.moodle.org/en/User:Martin_Langhoff

[Peter Lemenkov]

Hello All!

2016-02-10 13:08 GMT+01:00 Martin Langhoff <martin.langhoff(a)gmail.com&gt;:
>  Hi Peter(s),
>
>  On Wed, Feb 10, 2016 at 5:37 AM, Peter Lemenkov <lemenkov(a)gmail.com&gt; wrote:
> > Yes, that's a known issue I'm trying to address now. I've got a
> > possible fix from upstream developers, so things are getting better.
>
>  That's great to hear. Am I tracking the right BZ# with  BZ#1240487? 
Yes, that't the corresponding ticket.

https://bugzilla.redhat.com/1240487

-- 
With best regards, Peter Lemenkov.