From robin at reportlab.com  Fri Aug 10 06:18:48 2018
From: robin at reportlab.com (Robin Becker)
Date: Fri, 10 Aug 2018 11:18:48 +0100
Subject: [Wheel-builders] multibuild manylinux builds started to fail
Message-ID: <104b5cda-aa41-45d0-8688-5f5f7f94ee8e@chamonix.reportlab.co.uk>

I use travis to create manylinux and osx builds for the reportlab toolkit. After a minor change to the source I find all of the 
linux builds have failed, the OSX builds seem to have worked. I'm not sure why this should be, but the travis logs seem to 
indicate problems during the setup phase (before rportlab is involved

unicode width=16 python 2.7 etc etc
> make  install-data-hook
> make[3]: Entering directory `/io/libpng-1.6.32'
> + cd /usr/local/include
> + for f in pnglibconf.h png.h pngconf.h
> + rm -f pnglibconf.h
> + ln -s libpng16/pnglibconf.h pnglibconf.h
> + for f in pnglibconf.h png.h pngconf.h
> + rm -f png.h
> + ln -s libpng16/png.h png.h
> + for f in pnglibconf.h png.h pngconf.h
> + rm -f pngconf.h
> + ln -s libpng16/pngconf.h pngconf.h
> + cd /usr/local/lib/pkgconfig
> + rm -f libpng.pc
> + ln -s libpng16.pc libpng.pc
> make[3]: Leaving directory `/io/libpng-1.6.32'
> make[2]: Leaving directory `/io/libpng-1.6.32'
> make[1]: Leaving directory `/io/libpng-1.6.32'
>   % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
>                                  Dload  Upload   Total   Spent    Left  Speed
> 100   271  100   271    0     0    808      0 --:--:-- --:--:-- --:--:--   808
> gzip: stdin: not in gzip format
> tar: Child returned status 1
> tar: Error exit delayed from previous errors


there might be some issue with travis so I tried the build locally and see the same kind of error. The multibuild repository is here

https://github.com/MrBitBucket/wheel-builder

not being an expert I was wondering if there's something wrong with my setup. The config.sh is the bit where libpng is built and 
differs for OSX it is here

https://github.com/MrBitBucket/wheel-builder/blob/master/config.sh


Pointers would be really useful.
--
Robin Becker

From matthew.brett at gmail.com  Fri Aug 10 06:29:09 2018
From: matthew.brett at gmail.com (Matthew Brett)
Date: Fri, 10 Aug 2018 11:29:09 +0100
Subject: [Wheel-builders] multibuild manylinux builds started to fail
In-Reply-To: <104b5cda-aa41-45d0-8688-5f5f7f94ee8e@chamonix.reportlab.co.uk>
References: <104b5cda-aa41-45d0-8688-5f5f7f94ee8e@chamonix.reportlab.co.uk>
Message-ID: <CAH6Pt5rt+eafkiEw1oYcx2KFhmp45PBwcLC0vHy-L98CUh+ajA@mail.gmail.com>

Hi,

On Fri, Aug 10, 2018 at 11:18 AM, Robin Becker <robin at reportlab.com> wrote:
> I use travis to create manylinux and osx builds for the reportlab toolkit.
> After a minor change to the source I find all of the linux builds have
> failed, the OSX builds seem to have worked. I'm not sure why this should be,
> but the travis logs seem to indicate problems during the setup phase (before
> rportlab is involved
>
> unicode width=16 python 2.7 etc etc
>>
>> make  install-data-hook
>> make[3]: Entering directory `/io/libpng-1.6.32'
>> + cd /usr/local/include
>> + for f in pnglibconf.h png.h pngconf.h
>> + rm -f pnglibconf.h
>> + ln -s libpng16/pnglibconf.h pnglibconf.h
>> + for f in pnglibconf.h png.h pngconf.h
>> + rm -f png.h
>> + ln -s libpng16/png.h png.h
>> + for f in pnglibconf.h png.h pngconf.h
>> + rm -f pngconf.h
>> + ln -s libpng16/pngconf.h pngconf.h
>> + cd /usr/local/lib/pkgconfig
>> + rm -f libpng.pc
>> + ln -s libpng16.pc libpng.pc
>> make[3]: Leaving directory `/io/libpng-1.6.32'
>> make[2]: Leaving directory `/io/libpng-1.6.32'
>> make[1]: Leaving directory `/io/libpng-1.6.32'
>>   % Total    % Received % Xferd  Average Speed   Time    Time     Time
>> Current
>>                                  Dload  Upload   Total   Spent    Left
>> Speed
>> 100   271  100   271    0     0    808      0 --:--:-- --:--:-- --:--:--
>> 808
>> gzip: stdin: not in gzip format
>> tar: Child returned status 1
>> tar: Error exit delayed from previous errors

I wonder if you've hit the fact that the bzip.org website went down a
few days ago.

Can you try the latest `devel` branch of multibuild.  I think that is
fixed there (by using the archive.org copy).

Cheers,

Matthew

From robin at reportlab.com  Fri Aug 10 07:04:54 2018
From: robin at reportlab.com (Robin Becker)
Date: Fri, 10 Aug 2018 12:04:54 +0100
Subject: [Wheel-builders] multibuild manylinux builds started to fail
In-Reply-To: <CAH6Pt5rt+eafkiEw1oYcx2KFhmp45PBwcLC0vHy-L98CUh+ajA@mail.gmail.com>
References: <104b5cda-aa41-45d0-8688-5f5f7f94ee8e@chamonix.reportlab.co.uk>
 <CAH6Pt5rt+eafkiEw1oYcx2KFhmp45PBwcLC0vHy-L98CUh+ajA@mail.gmail.com>
Message-ID: <c2687ea3-4876-58f7-02c3-cd8c3c642be9@chamonix.reportlab.co.uk>

On 10/08/2018 11:29, Matthew Brett wrote:
> Hi,
..........
>>> tar: Child returned status 1
>>> tar: Error exit delayed from previous errors
> 
> I wonder if you've hit the fact that the bzip.org website went down a
> few days ago.
> 
> Can you try the latest `devel` branch of multibuild.  I think that is
> fixed there (by using the archive.org copy).
> 
> Cheers,
> 
> Matthew
> 
Hi Matthew, yes after looking at your site I have updated my multibuild submodule and some of the tests have succeeded. The 
failures seem to be as before so perhaps some caching is going on somewhere.

-- 
Robin Becker

From robin at reportlab.com  Fri Aug 10 07:12:45 2018
From: robin at reportlab.com (Robin Becker)
Date: Fri, 10 Aug 2018 12:12:45 +0100
Subject: [Wheel-builders] multibuild manylinux builds started to fail
In-Reply-To: <CAH6Pt5rt+eafkiEw1oYcx2KFhmp45PBwcLC0vHy-L98CUh+ajA@mail.gmail.com>
References: <104b5cda-aa41-45d0-8688-5f5f7f94ee8e@chamonix.reportlab.co.uk>
 <CAH6Pt5rt+eafkiEw1oYcx2KFhmp45PBwcLC0vHy-L98CUh+ajA@mail.gmail.com>
Message-ID: <9668de8c-2f7e-2daf-db60-6e8bebce178b@chamonix.reportlab.co.uk>

Hi Matthew, pretty sure this might not be related, but one of the errors I see on travis looks like this



> Starting instance
> ? selected image "travis-ci-garnet-trusty-1512502259-986baf0"
> ? rendered startup script
> ? inserted instance
> ? sleeping 5s before checking instance insert
> ? polling for instance insert completion...
> ? instance is ready (9.085s)
> step_upload_script
> Uploading script
............
> Pre-installed Ruby versions0m
> ruby-2.2.7
> ruby-2.3.4
> ruby-2.4.1
> Network availability confirmed.
> 3.5 is not installed; attempting download
> Downloading archive: https://s3.amazonaws.com/travis-python-archives/binaries/ubuntu/14.04/x86_64/python-3.5.tar.bz2
> $ curl -sSf -o python-3.5.tar.bz2 ${archive_url}
> curl: (56) SSL read: error:00000000:lib(0):func(0):reason(0), errn


and another like this
................
> ff350cd7a61c: Pull complete
> a84c63cd7072: Pull complete
> Digest: sha256:8111345b6641541ccaa68597fbaf46eab5f6fc461b9a739db1bb952ec13d78a4
> Status: Downloaded newer image for quay.io/pypa/manylinux1_x86_64:latest
> Building libraries...
> https://mirrors.rit.edu/fedora/archive/epel/5/x86_64/repodata/repomd.xml: [Errno 4] IOError: <urlopen error (99, 'Cannot assign requested address')>
> Trying other mirror.
> Still building...
> http://people.centos.org/tru/devtools-2/5/x86_64/RPMS/repodata/repomd.xml: [Errno 12] Timeout: <urlopen error timed out>
> Trying other mirror.
> Error: Cannot retrieve repository metadata (repomd.xml) for repository: testing-devtools-2-centos-5. Please verify its path and try again

so perhaps travis is a bit flaky right now
-- 
Robin Becker

From robin at reportlab.com  Fri Aug 10 08:41:27 2018
From: robin at reportlab.com (Robin Becker)
Date: Fri, 10 Aug 2018 13:41:27 +0100
Subject: [Wheel-builders] multibuild manylinux builds started to fail
In-Reply-To: <9668de8c-2f7e-2daf-db60-6e8bebce178b@chamonix.reportlab.co.uk>
References: <104b5cda-aa41-45d0-8688-5f5f7f94ee8e@chamonix.reportlab.co.uk>
 <CAH6Pt5rt+eafkiEw1oYcx2KFhmp45PBwcLC0vHy-L98CUh+ajA@mail.gmail.com>
 <9668de8c-2f7e-2daf-db60-6e8bebce178b@chamonix.reportlab.co.uk>
Message-ID: <2e091d2f-0f4e-b2a7-363b-efd44008df16@chamonix.reportlab.co.uk>

On the other hand perhaps the web archive is feeling the strain

> make[1]: Leaving directory `/io/libpng-1.6.32'
>   % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
>                                  Dload  Upload   Total   Spent    Left  Speed
>   0     0    0     0    0     0      0      0 --:--:--  0:00:42 --:--:--     0Still building...
>   0     0    0     0    0     0      0      0 --:--:--  0:01:42 --:--:--     0Still building...
>   0     0    0     0    0     0      0      0 --:--:--  0:02:06 --:--:--     0curl: (7) Failed to connect to web.archive.org port 443: Connection timed out



-- 
Robin Becker

From matthew.brett at gmail.com  Fri Aug 10 08:57:18 2018
From: matthew.brett at gmail.com (Matthew Brett)
Date: Fri, 10 Aug 2018 13:57:18 +0100
Subject: [Wheel-builders] multibuild manylinux builds started to fail
In-Reply-To: <2e091d2f-0f4e-b2a7-363b-efd44008df16@chamonix.reportlab.co.uk>
References: <104b5cda-aa41-45d0-8688-5f5f7f94ee8e@chamonix.reportlab.co.uk>
 <CAH6Pt5rt+eafkiEw1oYcx2KFhmp45PBwcLC0vHy-L98CUh+ajA@mail.gmail.com>
 <9668de8c-2f7e-2daf-db60-6e8bebce178b@chamonix.reportlab.co.uk>
 <2e091d2f-0f4e-b2a7-363b-efd44008df16@chamonix.reportlab.co.uk>
Message-ID: <CAH6Pt5qXUWTOxXXZ1qtJq0qG1-9suK7N_G2xCRq_ff-KU-OEtw@mail.gmail.com>

Hi,

On Fri, Aug 10, 2018 at 1:41 PM, Robin Becker <robin at reportlab.com> wrote:
> On the other hand perhaps the web archive is feeling the strain
>
>> make[1]: Leaving directory `/io/libpng-1.6.32'
>>   % Total    % Received % Xferd  Average Speed   Time    Time     Time
>> Current
>>                                  Dload  Upload   Total   Spent    Left
>> Speed
>>   0     0    0     0    0     0      0      0 --:--:--  0:00:42 --:--:--
>> 0Still building...
>>   0     0    0     0    0     0      0      0 --:--:--  0:01:42 --:--:--
>> 0Still building...
>>   0     0    0     0    0     0      0      0 --:--:--  0:02:06 --:--:--
>> 0curl: (7) Failed to connect to web.archive.org port 443: Connection timed
>> out

Yes, I find that site is not very reliable in general.  I've committed
a copy of the bzip2 archive to multibuild now - try the latest latest
`devel` branch?

Cheers,

Matthew