[R-pkg-devel] Package failing reverse dependency checks

[Lluís Revilla]

Hi David,

Dependency checks cannot be run on win /Mac builders as far as I know.
Some packages have more than 2000 packages depending on them.
It would be too much for the heavily used CRAN's machines.

Best,

Lluís

On Thu, 8 Feb 2024 at 13:52, David Hugh-Jones <davidhughjones using gmail.com>
wrote:

> Thanks for this tip. Out of interest, is there a way to make win/Mac
> builder run revdep checks?
>
> Writing: wyclif.substack.com
> Book: www.wyclifsdust.com
>
>
> On Thu, 8 Feb 2024 at 10:26, Lluís Revilla <lluis.revilla using gmail.com>
> wrote:
>
>> Hi David,
>>
>> If you didn't increase the version number it could happen that the old
>> version of the package was used (as CRAN might not be aware that there is a
>> new version).
>> The CRAN repository policy says: "Increasing the version number at each
>> submission reduces confusion so is preferred even when a previous
>> submission was not accepted".
>> In addition, it is easier to track how smooth the submission process is
>> for maintainers/developers.
>>
>> I would recommend submitting again, changing the version number of the
>> package.
>> Checking on  win builders and macos builders CRAN provides is the closest
>> one, other approaches include rhub2.
>>
>> Best,
>>
>> Lluís
>>
>>
>>
>> On Thu, 8 Feb 2024 at 10:24, David Hugh-Jones <davidhughjones using gmail.com>
>> wrote:
>>
>>> Hi all,
>>>
>>> I'm trying to upload a new version of my "huxtable" package. It is
>>> currently failing reverse dependency checks for two packages (homnormal
>>> and
>>> RSStest). The relevant failures are below.
>>>
>>> I got this failure one time, and fixed the problem, which relates to
>>> mistakenly relying on the Suggested: knitr package (see here for the
>>> commit: https://github.com/hughjonesd/huxtable/commit/5a3edc). After the
>>> commit, reverse dependency checks for homnormal and RSStest pass on my
>>> machine, when testing either with revdepcheck::revdep_check or
>>> tools::check_packages_in_dir, and even when knitr is not installed. But,
>>> after I uploaded the new package to CRAN, the same failure recurred.
>>>
>>> My new release candidate had the same version number as the previous one
>>> (which had failed the revdep check, and therefore not been published on
>>> CRAN). Is it possible that CRAN just tested the old version again?
>>>
>>> If not, then can anyone suggest the best way to debug a revdep check on
>>> as
>>> close a setup to the CRAN machines as possible?
>>>
>>> Cheers,
>>> David
>>>
>>> Git tag for the last CRAN submission:
>>> https://github.com/hughjonesd/huxtable/releases/tag/v5.5.4-rc3
>>>
>>> Info from the CRAN email:
>>> --
>>> Debian: <
>>>
>>> https://win-builder.r-project.org/incoming_pretest/huxtable_5.5.4_20240205_164815/reverseDependencies/summary.txt
>>> >
>>> RSStest, homnormal
>>>
>>> Log dir: <
>>>
>>> https://win-builder.r-project.org/incoming_pretest/huxtable_5.5.4_20240205_164815/
>>> >
>>> The files will be removed after roughly 7 days.
>>>
>>> Pretests:
>>> Windows: <
>>>
>>> https://win-builder.r-project.org/incoming_pretest/huxtable_5.5.4_20240205_164815/Windows/00check.log
>>> >
>>> Debian: <
>>>
>>> https://win-builder.r-project.org/incoming_pretest/huxtable_5.5.4_20240205_164815/Debian/00check.log
>>> >
>>> --
>>> Changes to worse in reverse depends:
>>>
>>> Package: homnormal
>>> Check: examples
>>> New result: ERROR
>>>   Running examples in ‘homnormal-Ex.R’ failed
>>>   The error most likely occurred in:
>>>
>>>   > base::assign(".ptime", proc.time(), pos = "CheckExEnv")
>>>   > ### Name: Brown_Forsythe
>>>   > ### Title: Brown-Forsythe Test for Homogeniety
>>>   > ### Aliases: Brown_Forsythe
>>>   >
>>>   > ### ** Examples
>>>   >
>>>   >     data(FH_data)
>>>   >    x1=FH_data$SurvivalTime
>>>   >    x2=FH_data$HospitalNo
>>>   >    Brown_Forsythe(x1,x2)
>>>   Error in loadNamespace(x) : there is no package called ‘knitr’
>>>   Calls: Brown_Forsythe ... loadNamespace -> withRestarts ->
>>> withOneRestart
>>> -> doWithOneRestart
>>>   Execution halted
>>>
>>> Package: RSStest
>>> Check: examples
>>> New result: ERROR
>>>   Running examples in ‘RSStest-Ex.R’ failed
>>>   The error most likely occurred in:
>>>
>>>   > base::assign(".ptime", proc.time(), pos = "CheckExEnv")
>>>   > ### Name: teststat_MRSS
>>>   > ### Title: Median Ranked Set Sampling Test
>>>   > ### Aliases: teststat_MRSS
>>>   >
>>>   > ### ** Examples
>>>   >
>>>   > x1=matrix(c(1,2.3, 3.4,4.5,5.6,4 ),nrow=3)
>>>   > x2=matrix(c(2,3.2, 4.2,6.5,4.6,6 ),nrow=3)
>>>   > teststat_MRSS(x1,x2,tn=1000)
>>>   Error in loadNamespace(x) : there is no package called ‘knitr’
>>>   Calls: teststat_MRSS ... loadNamespace -> withRestarts ->
>>> withOneRestart
>>> -> doWithOneRestart
>>>   Execution halted
>>>
>>>         [[alternative HTML version deleted]]
>>>
>>> ______________________________________________
>>> R-package-devel using r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-package-devel
>>>
>>

	[[alternative HTML version deleted]]