R devel - May 2015 - CRAN check for package on Sparc Solaris

Dear All,

The metafor package currently fails CRAN checks on Sparc Solaris:

http://cran.r-project.org/web/checks/check_results_metafor.html

The problem is probably due to an unintended (= stupid) use of identical() in a
couple tests. I have changed that to more appropriate tests using all.equal().
However, before I resubmit the package to CRAN, I would really like to make sure
that the updated package passes all checks also on Sparc Solaris.

The issue/question of how to test packages on Sparc Solaris has come up before:

https://stat.ethz.ch/pipermail/r-devel/2010-September/058538.html
https://stat.ethz.ch/pipermail/r-devel/2011-November/062430.html

Unless anything has changed in the meantime, it seems like the help of kind
volunteers is still the only option to conduct such tests. So, is there anybody
out there with access to a Sparc machine running Solaris willing to run 'R
CMD check --as-cran' on the updated tarball? Would be very much appreciated!

Best,
Wolfgang

Hi Wolfgang, I can test your package, please send it to me in private.

Btw. I only have an x86 machine, and no sparc, so if your problem is
sparc-specific, then my test is probably useless.

Gabor

On Wed, May 13, 2015 at 3:48 PM, Viechtbauer Wolfgang (STAT) <
wolfgang.viechtbauer at maastrichtuniversity.nl> wrote:
> Dear All,
>
> The metafor package currently fails CRAN checks on Sparc Solaris:
>
> http://cran.r-project.org/web/checks/check_results_metafor.html
>
> The problem is probably due to an unintended (= stupid) use of identical()
> in a couple tests. I have changed that to more appropriate tests using
> all.equal(). However, before I resubmit the package to CRAN, I would really
> like to make sure that the updated package passes all checks also on Sparc
> Solaris.
>
> The issue/question of how to test packages on Sparc Solaris has come up
> before:
>
> https://stat.ethz.ch/pipermail/r-devel/2010-September/058538.html
> https://stat.ethz.ch/pipermail/r-devel/2011-November/062430.html
>
> Unless anything has changed in the meantime, it seems like the help of
> kind volunteers is still the only option to conduct such tests. So, is
> there anybody out there with access to a Sparc machine running Solaris
> willing to run 'R CMD check --as-cran' on the updated tarball?
Would be
> very much appreciated!
>
> Best,
> Wolfgang
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
	[[alternative HTML version deleted]]

Thank you for the kind offer! Unfortunately, the problem is really Sparc
Solaris:

http://cran.r-project.org/web/checks/check_results_metafor.html

On Solaris x86 and all other flavors, it passes without errors.

Best,
Wolfgang

From: G?bor Cs?rdi [csardi.gabor at gmail.com]
Sent: Wednesday, May 13, 2015 9:53 PM
To: Viechtbauer Wolfgang (STAT)
Cc: r-devel at r-project.org
Subject: Re: [Rd] CRAN check for package on Sparc Solaris

Hi Wolfgang, I can test your package, please send it to me in private.  

Btw. I only have an x86 machine, and no sparc, so if your problem is
sparc-specific, then my test is probably useless.


Gabor


On Wed, May 13, 2015 at 3:48 PM, Viechtbauer Wolfgang (STAT)
<wolfgang.viechtbauer at maastrichtuniversity.nl> wrote:

Dear All,

The metafor package currently fails CRAN checks on Sparc Solaris:

http://cran.r-project.org/web/checks/check_results_metafor.html

The problem is probably due to an unintended (= stupid) use of identical() in a
couple tests. I have changed that to more appropriate tests using all.equal().
However, before I resubmit the package to CRAN, I would really like to make sure
that the updated package passes all checks also on Sparc Solaris.

The issue/question of how to test packages on Sparc Solaris has come up before:

https://stat.ethz.ch/pipermail/r-devel/2010-September/058538.html
https://stat.ethz.ch/pipermail/r-devel/2011-November/062430.html

Unless anything has changed in the meantime, it seems like the help of kind
volunteers is still the only option to conduct such tests. So, is there anybody
out there with access to a Sparc machine running Solaris willing to run 'R
CMD check --as-cran' on the updated tarball? Would be very much appreciated!

Best,
Wolfgang
______________________________________________
R-devel at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

This issue is discussed in 
http://cran.r-project.org/doc/manuals/r-release/R-exts.html#Writing-portable-packages,
to which he was referred by the CRAN report.  It fails in the same way 
with --disable-long-double on an x86_64 Linux box (as he was told).

On 13/05/2015 20:48, Viechtbauer Wolfgang (STAT) wrote:
> Dear All,
>
> The metafor package currently fails CRAN checks on Sparc Solaris:
>
> http://cran.r-project.org/web/checks/check_results_metafor.html
>
> The problem is probably due to an unintended (= stupid) use of identical()
in a couple tests. I have changed that to more appropriate tests using
all.equal(). However, before I resubmit the package to CRAN, I would really like
to make sure that the updated package passes all checks also on Sparc Solaris.
>
> The issue/question of how to test packages on Sparc Solaris has come up
before:
>
> https://stat.ethz.ch/pipermail/r-devel/2010-September/058538.html
> https://stat.ethz.ch/pipermail/r-devel/2011-November/062430.html
>
> Unless anything has changed in the meantime, it seems like the help of kind
volunteers is still the only option to conduct such tests. So, is there anybody
out there with access to a Sparc machine running Solaris willing to run 'R
CMD check --as-cran' on the updated tarball? Would be very much appreciated!
>
> Best,
> Wolfgang
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Emeritus Professor of Applied Statistics, University of Oxford
1 South Parks Road, Oxford OX1 3TG, UK

Yes, I did read that part but it doesn't say anything about who to contact
in case one doesn't have access to a Sparc Solaris machine, but one still
wants to run the tests on that architecture. As I said, I think I got the issue
resolved, but before I resubmit, I would prefer to check this -- I am ultimately
just trying to avoid potentially wasting people's time here.

As for Linux x86_64 without long doubles. Is this something that is part of the
checks here?

http://cran.r-project.org/web/checks/check_results_metafor.html

Doesn't seem that way. But is this still part of the CRAN checks?

I can of course simply assume/hope that the issue, which is now hopefully
resolved on Sparc Solaris, is the same that caused problems for Linux x86_64
without long doubles and that is then also hopefully resolved. Or I could hope
that another kind soul with a compiled version of R on Linux x86_64 with
--disable-long-double is willing to give the new tarball a whirl.
> -----Original Message-----
> From: R-devel [mailto:r-devel-bounces at r-project.org] On Behalf Of Prof
> Brian Ripley
> Sent: Thursday, May 14, 2015 08:06
> To: r-devel at r-project.org
> Subject: Re: [Rd] CRAN check for package on Sparc Solaris
> 
> This issue is discussed in
> http://cran.r-project.org/doc/manuals/r-release/R-exts.html#Writing-
> portable-packages,
> to which he was referred by the CRAN report.  It fails in the same way
> with --disable-long-double on an x86_64 Linux box (as he was told).
> 
> On 13/05/2015 20:48, Viechtbauer Wolfgang (STAT) wrote:
> > Dear All,
> >
> > The metafor package currently fails CRAN checks on Sparc Solaris:
> >
> > http://cran.r-project.org/web/checks/check_results_metafor.html
> >
> > The problem is probably due to an unintended (= stupid) use of
> identical() in a couple tests. I have changed that to more appropriate
> tests using all.equal(). However, before I resubmit the package to CRAN,
> I would really like to make sure that the updated package passes all
> checks also on Sparc Solaris.
> >
> > The issue/question of how to test packages on Sparc Solaris has come
up
> before:
> >
> > https://stat.ethz.ch/pipermail/r-devel/2010-September/058538.html
> > https://stat.ethz.ch/pipermail/r-devel/2011-November/062430.html
> >
> > Unless anything has changed in the meantime, it seems like the help of
> kind volunteers is still the only option to conduct such tests. So, is
> there anybody out there with access to a Sparc machine running Solaris
> willing to run 'R CMD check --as-cran' on the updated tarball?
Would be
> very much appreciated!
> >
> > Best,
> > Wolfgang
> > ______________________________________________
> > R-devel at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-devel
> >
> 
> 
> --
> Brian D. Ripley,                  ripley at stats.ox.ac.uk
> Emeritus Professor of Applied Statistics, University of Oxford
> 1 South Parks Road, Oxford OX1 3TG, UK
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel