[Bioc-devel] BiocCheck warnings and errors

[Hervé Pagès]

Let's please keep this conversation on the bioc-devel list where it started.

On 13/09/2022 08:53, Christian Arnold wrote:
> Hi Hervé!
>
> On 13/09/2022 17:47, Hervé Pagès wrote:
>> Hi Christian,
>>
>> On 13/09/2022 08:03, Christian Arnold wrote:
>>
>>> I have a few questions regarding BiocCheck errors and warnings that I
>>> receive for a new package version that I feel are not warranted:
>>>
>>>
>>> 1. I want to use the testthat framework in combination with
>>> devtools::test, and from my understanding, a subdirectory /testthat/
>>> within tests has to be used for this and is created when using
>>> devtools::test(). I still get an ERROR when using BiocCheck because I
>>> have no .R files in tests/ but within tests/testthat:
>>>
>>> * Checking unit tests... * ERROR: Add a .R or .Rin file in tests/
>>> directory or unit tests will not be run by R CMD check. See
>>> http://bioconductor.org/developers/how-to/unitTesting-guidelines/
>>
>> The error message is pretty clear about **why** you must have a .R or
>> .Rin file in tests/ It's fine that with your current setup you can run
>> the tests with devtools::test() but that's not enough: the tests
>> should also be run by R CMD check.
>
>
> This is already clarified, thanks to another reply, I didnt have the
> required testthat.R in the main test directory that starts the whole
> testing and which is recognized by R CMD check as well.
>
>
>>
>>> 2. My package has many packages listed in Suggests, with a lot of effort
>>> I tried to move as many packages there as possible to reduce the
>>> installation burden. From my understanding, BiocManager does NOT install
>>> the packages listed under /Suggests/. Thus, since I have 10+ packages in
>>> /Suggests/ that actually enhance user functionality and may be needed, I
>>> wrote a helper function that automatically installs these if the user
>>> wants, as a convenience so users dont have to install them manually on
>>> their own. However, no I get this error:
>>>
>>> ERROR: Remove install() calls (found 2 times)
>>
>> Having BiocManager::install() calls in the package should be ok as
>> long as those calls are **not** executed by R CMD build or R CMD
>> check. Running R CMD build or R CMD check should never install
>> packages on the user machine.
>
> They are not executed no, but the Bioc error is still raised. I guess I
> can then just ignore it, the package is already on Bioconductor.
>
>>
>>>
>>> 3. Minor but still: Can I use Biochceck also for the devel branch? I get
>>> the following warning because I am checking the devel version and not
>>> the release version:
>>>
>>> WARNING: y of x.y.z version should be even in release
>>
>> If you are going to submit your package to Bioconductor, you must
>> check your package against BioC devel, not against the current release.
>
>
> See above, I used my devel branch for BiocCheck, which has an "uneven"
> version number as required for the devel branch but BiocCheck seems to
> check against the "release" branch and therefore throws the warning.

Yes, you're checking the devel branch of your package, I get that. But I 
suspect you're using BioC release (i.e. BioC 3.15). What I'm saying is 
that you must use BioC devel. Loading BiocManager will tell you the 
version of Bioconductor that you are using.

Best,

H.

>
>
>>
>> Hope this helps,
>>
>> H.
>>
>>
>>>
>>> ***Thanks for your input and thoughts!****Christian*
>>>
>>>     [[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