[Bioc-devel] Windows specific error during CHECK

[Hervé Pagès]

Hi Felix,

Interestingly the unit tests error on 64-bit Windows is gone today but 
now the unit tests trigger a segfault on Linux:

 
https://bioconductor.org/checkResults/3.14/bioc-LATEST/mia/nebbiolo2-checksrc.html

The segfault seems to happen during a call to ape:::all.equal.phylo(). 
This kind of hard-to-reproduce segfault is usually caused by some 
problematic native code. The usual suspects are uninitialized C 
variables/arrays. Running the problematic test thru Valgrind might give 
you some hints.

Best,
H.


On 26/10/2021 13:09, Felix Ernst wrote:
> Thanks Herve!
> 
> -----Ursprüngliche Nachricht-----
> Von: Hervé Pagès <hpages.on.github using gmail.com>
> Gesendet: Dienstag, 26. Oktober 2021 21:53
> An: Felix Ernst <felix.gm.ernst using outlook.com>; bioc-devel using r-project.org
> Betreff: Re: [Bioc-devel] Windows specific error during CHECK
> 
> Hi Felix,
> 
> On 26/10/2021 12:10, Felix Ernst wrote:
>> Hi Bioc-Team,
>>
>> I have got a problem with an error on Windows, which I am not able to debug.
>>
>> http://bioconductor.org/checkResults/devel/bioc-LATEST/mia/riesling1-c
>> hecksrc.html
>>
>> The testthat.out.fail output doesn't indicate any problems and the testthat.out on other systems looks the same.
>>
>> Can someone provide any hint on what might be going on? Thanks in advance!
> 
> Hmm.. even more puzzling is that if I go on riesling1, start R (in 64-bit mode), and run the tests manually, I don't see any problem:
> 
>   > test_check("mia")
> 
>   
> |======================================================================|
> 100%
> 
> 
> 
> Time difference of 6.58 secs
> 
> 
> 
> Initializing error rates to maximum possible estimate.
> 
> selfConsist step 1 .
> 
>      selfConsist step 2
> 
>      selfConsist step 3
> 
>      selfConsist step 4
> 
> Convergence after  4  rounds.
> 
> Initializing error rates to maximum possible estimate.
> 
> selfConsist step 1 .
> 
>      selfConsist step 2
> 
>      selfConsist step 3
> 
>      selfConsist step 4
> 
> Convergence after  4  rounds.
> 
> initial  value 0.383462
> 
> iter   5 value 0.161655
> 
> iter  10 value 0.113278
> 
> final  value 0.003270
> 
> converged
> 
> initial  value 0.000000
> 
> final  value 0.000000
> 
> converged
> 
> initial  value 0.000000
> 
> final  value 0.000000
> 
> converged
> 
> initial  value 0.000000
> 
> final  value 0.000000
> 
> converged
> 
> [ FAIL 0 | WARN 2 | SKIP 0 | PASS 630 ]
> 
> 
> So this only seems to happen in the context of 'R CMD check'. I wonder
> if the fact that we run 'R CMD check' with the --force-multiarch option
> (so the examples and tests are run for the 2 Windows archs) could be
> somehow related to the issue.
> 
> Anyways, I won't be able to do more testing today because the builds on
> riesling1 are going to start in a few minutes. I'll try to get back to
> this tomorrow.
> 
> Cheers,
> H.
> 
> 
>>
>> Best regards,
>> Felix
>>
>> 	[[alternative HTML version deleted]]
>>
>> _______________________________________________
>> Bioc-devel using r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>
> 

-- 
Hervé Pagès

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