[R-pkg-devel] check error with r-devel-windows-ix86+x86_64 due to missing Hmisc package

Duncan Murdoch murdoch@dunc@n @end|ng |rom gm@||@com
Fri Jun 25 16:10:50 CEST 2021


I'm back in civilization for a few days, and have submitted a PR to 
Hmisc:  https://github.com/harrelfe/Hmisc/pull/143 .

Duncan Murdoch

On 23/06/2021 8:51 a.m., Gilbert Ritschard wrote:
> Thanks for this suggestion.
> 
> Pushing the investigation a bit further, I found that the error resulted in chain (curl -> httr -> plotly -> Hmisc -> TraMineR -> TraMineRextras) from an issue of the package curl with the ucrt tool. An update of curl with a fix has been released today. So seems that I just have to wait until curl is built for the ucrt experimental toolchain.
> 
> Gilbert
>    
> 
>> -----Original Message-----
>> From: Duncan Murdoch <murdoch.duncan using gmail.com>
>> Sent: mercredi 23 juin 2021 14:19
>> To: Gilbert Ritschard <Gilbert.Ritschard using unige.ch>; R-package-devel using r-
>> project.org
>> Subject: Re: [R-pkg-devel] check error with r-devel-windows-ix86+x86_64 due to
>> missing Hmisc package
>>
>> On 23/06/2021 7:41 a.m., Gilbert Ritschard wrote:
>>> Hi all,
>>>
>>> The submission of an update of TraMineRextras on the CRAN does not pass the
>> test due to an error with r-devel-windows-ix86+x86_64.
>>>
>>> The error says:
>>>
>>>> library('TraMineRextras')
>>>     Loading required package: TraMineR
>>>     Error: package or namespace load failed for 'TraMineR' in
>> find.package(package, lib.loc, verbose = verbose):
>>>      there is no package called 'Hmisc'
>>>
>>> How can I fix this this error? The package relies strongly on TraMineR, I cannot
>> remove the dependence on that package.
>>>
>>> The Check Results table for Hmisc shows an error for r-devel-x86_64-gcc10-
>> UCRT due to missing package plotly.
>>
>> That's a real error in Hmisc, but it should be easy to fix.  Hmisc suggests plotly,
>> then uses it unconditionally.  You could prepare and test a patch for Hmisc to fix
>> this issue (reducing the ERROR to a NOTE if that's the only issue), and then your
>> package wouldn't get this error.
>>
>> I'd do so myself, but I have very expensive Internet access most of the time for the
>> next few weeks.
>>
>> Duncan Murdoch



More information about the R-package-devel mailing list