[Bioc-devel] Wishlist: on demand R CMD check

Laurent Gatto lg390 at cam.ac.uk
Tue Jun 2 14:39:26 CEST 2015


To what extend could the single package builder be used for such a
feature? This would not address Michael's point, but it is a way to get
access to all archs using existing software infrastructure.

Laurent

On  2 June 2015 13:18, Michael Lawrence wrote:

> Maybe the motion towards github and the work Gabor C has been doing on
> Travis integration might help with this?
>
> Note that a proper check should test all reverse dependencies of your
> package, and that would get expensive for some of the more core
> packages... so that might need to be deferred to the regular
> repository-level build process.
>
>
> On Tue, Jun 2, 2015 at 2:48 AM, Sean Davis <seandavi at gmail.com> wrote:
>> On Tue, Jun 2, 2015 at 5:32 AM, Vincent Carey <stvjc at channing.harvard.edu>
>> wrote:
>>
>>> I agree that this would be nice; not sure how feasible.  Could a specific
>>> approach with containers provide a solution?
>>>
>>
>> Unfortunately, containers are linux microkernels, so Windows and MacOS are
>> basically out.
>>
>>
>>> On Mon, Jun 1, 2015 at 11:09 PM, Kasper Daniel Hansen <
>>> kasperdanielhansen at gmail.com> wrote:
>>>
>>> > For a substantial number of issues with R CMD check, it is easy to
>>> > reproduce the errors/warnings/notes on my own system; simply update all
>>> > packages and run the test.
>>> >
>>> > For a small number of errors I have sometimes (over the years) had
>>> problems
>>> > reproducing them, perhaps because they are platform specific or because
>>> of
>>> > other issues.
>>> >
>>> > In those cases, especially for packages which I work on infrequently, the
>>> > current submission / check introduces a long delay and - more
>>> importantly -
>>> > substantial mental overhead for me.  This is because I submit a fix and
>>> > then I usually have to wait 36-48h to see the result of R CMD check.
>>> >
>>> > It would be great if we could get access to running R CMD check on all
>>> > three platforms by demand, with a results page which is semi-private (to
>>> > not confuse the package with a fix with the official devel version).  I
>>> > realize this could be misused or introduce a large computational overhead
>>> > on the project.  Perhaps only make this available for packages accepted
>>> > into the project.  But it would be nice.  For example, I have been
>>> hunting
>>> > a bug in minfi for like 10 days now, and the fact that I need to wait 48h
>>> > to get the result of a fix, really makes this painful.
>>> >
>>> > Best,
>>> > Kasper
>>> >
>>> >         [[alternative HTML version deleted]]
>>> >
>>> > _______________________________________________
>>> > Bioc-devel at r-project.org mailing list
>>> > https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>> >
>>>
>>>         [[alternative HTML version deleted]]
>>>
>>> _______________________________________________
>>> Bioc-devel at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>
>>
>>         [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> Bioc-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>
> _______________________________________________
> Bioc-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel

-- 
Laurent Gatto | @lgatt0
http://cpu.sysbiol.cam.ac.uk/



More information about the Bioc-devel mailing list