[Bioc-devel] Check report error on Windows platform

[Jiping Wang]

Hi Herv�,

   I searched similar warning or error message as in the check results of mac version of my package:


** using staged installation
** libs
gfortran -mmacosx-version-min=10.13 -fno-optimize-sibling-calls  -fPIC  -Wall -g -O2  -c  cvtbfb.f90 -o cvtbfb.o
gfortran -mmacosx-version-min=10.13 -fno-optimize-sibling-calls  -fPIC  -Wall -g -O2  -c  cvtbfbNL4.f90 -o cvtbfbNL4.o
cvtbfbNL4.f90:13:51:

   13 |   real*8    freqN4(4,4,4,4),tranN4(5:147,4,4,4,4,4),freqL4(4,4,4,4)
      |                                                   1
Warning: Array �trann4� at (1) is larger than limit set by �-fmax-stack-var-size=�, moved from stack to static storage. This makes the procedure unsafe when called recursively, or concurrently from multiple threads. Consider using �-frecursive�, or increase the �-fmax-stack-var-size=� limit, or change the code to use an ALLOCATABLE array. [-Wsurprising]
gfortran -mmacosx-version-min=10.13 -fno-optimize-sibling-calls  -fPIC  -Wall -g -O2  -c  vtbfb.f90 -o vtbfb.o
gfortran -mmacosx-version-min=10.13 -fno-optimize-sibling-calls  -fPIC  -Wall -g -O2  -c  vtbfbNL4.f90 -o vtbfbNL4.o
vtbfbNL4.f90:13:51:

   13 |   real*8    freqN4(4,4,4,4),tranN4(5:147,4,4,4,4,4),freqL4(4,4,4,4)
      |                                                   1
Warning: Array �trann4� at (1) is larger than limit set by �-fmax-stack-var-size=�, moved from stack to static storage. This makes the procedure unsafe when called recursively, or concurrently from multiple threads. Consider using �-frecursive�, or increase the �-fmax-stack-var-size=� limit, or change the code to use an ALLOCATABLE array. [-Wsurprising]

and I did found a few posts:

  1.  https://www.scivision.dev/gfortran-stack-to-static-warning/
  2.  https://stackoverflow.com/questions/65384213/using-very-large-arrays-in-fortran-gfortran
  3.  https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98411


#1 says this is true warning and allocable array should defined for large arrays (by gfortran 10). However #2 and #3 both suggested that the array size was not the issue here. The warning message was generated as false alarm.

The definition of array that has warning in line 13 has been in the package since inception. I updated gfortran8  from my local computer (mac osx) to gfortran12, and observe no such warnings. I still don�t have clues whether this warning caused the error in windows platform, as other two platforms didn�t have errors in it.

Thanks.

jiping





From: Herv� Pag�s <hpages.on.github using gmail.com>
Date: Monday, August 8, 2022 at 4:40 PM
To: Jiping Wang <jzwang using northwestern.edu>, bioc-devel using r-project.org <bioc-devel using r-project.org>
Subject: Re: [Bioc-devel] Check report error on Windows platform
Hi jiping,

I suspect R dies during the call to predNuPoP(). Could this be related
to the recent changes you made to the master branch of the package, in
particular to changes to the Fortran code? Note that the Fortran
compiler of the Mac builder emits a warning about the size of the
'trann4' array being unsafe:

https://urldefense.com/v3/__https://bioconductor.org/checkResults/3.16/bioc-LATEST/NuPoP/lconway-install.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWr5V4byK$<https://urldefense.com/v3/__https:/bioconductor.org/checkResults/3.16/bioc-LATEST/NuPoP/lconway-install.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWr5V4byK$>

We don't see this warning on the Mac builder in release so may it's an
issue that got introduced recently?

https://urldefense.com/v3/__https://bioconductor.org/checkResults/3.15/bioc-LATEST/NuPoP/merida1-install.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWsT7BmyF$<https://urldefense.com/v3/__https:/bioconductor.org/checkResults/3.15/bioc-LATEST/NuPoP/merida1-install.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWsT7BmyF$>

On a different note, could you please remove the files in the
NuPoP/inst/doc/ folder? These files don't belong to git (they get
automatically generated and added to the package source tarball by 'R
CMD build').

Thanks,

H.


On 08/08/2022 07:24, Jiping Wang wrote:
> Hi,
>
> Could anyone give some hints on the check errors of my package for windows system? It seems everything was fine for linux and macs. Please see checkResults page below. Thanks.
>
> Best,
> jiping
>
> https://urldefense.com/v3/__https://bioconductor.org/checkResults/devel/bioc-LATEST/NuPoP/palomino4-checksrc.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWu2zNr4O$<https://urldefense.com/v3/__https:/bioconductor.org/checkResults/devel/bioc-LATEST/NuPoP/palomino4-checksrc.html__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWu2zNr4O$>
>
>
>        [[alternative HTML version deleted]]
>
> _______________________________________________
> Bioc-devel using r-project.org mailing list
> https://urldefense.com/v3/__https://stat.ethz.ch/mailman/listinfo/bioc-devel__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWqxv8TUu$<https://urldefense.com/v3/__https:/stat.ethz.ch/mailman/listinfo/bioc-devel__;!!Dq0X2DkFhyF93HkjWTBQKhk!V_ZFRVHQ6u7_bW6tU1RvOAi5fAXp6AAn8EcIXajtP2_xjy2lQrcJB8oEu9Bsqi75MvDmzczGg65nBH0gACpiiu5ZWqxv8TUu$>

--
Herv� Pag�s

Bioconductor Core Team
hpages.on.github using gmail.com

	[[alternative HTML version deleted]]