[Rspatial-devel] vignette that depends on PosgreSQL data base access

[Uwe Ligges]

On 15.07.2011 19:36, Edzer Pebesma wrote:
> Thanks Uwe! I decided to have the database sections with eval=FALSE, and
> upload a fully evaluated version off-CRAN; it's not a big deal after all.
>
> spacetime 0.5-1 is now on /incoming on CRAN.


Great, thanks,
uwe

> Best wishes,
>
> On 07/12/2011 09:54 PM, Uwe Ligges wrote:
>>
>>
>> On 12.07.2011 14:26, Edzer Pebesma wrote:
>>> I wrote a vignette for package spacetime that explains how massive
>>> spacetime data sets in a PosgreSQL data base can be proxied by R objects
>>> in package spacetime; see
>>> https://svn.52north.org/svn/geostatistics/main/spacetime/inst/doc/stpg.Rnw
>>>
>>>
>>> To compile it, it needs read/write access to a postgresql data base that
>>> has postgis extensions. I assume CRAN doesn't have these. How do I carry
>>> on with this? Can I tell CRAN to not try to compile it, or should I
>>> upload it somewhere else? Or simply upload&   ignore?
>>
>> Edzer,
>>
>> you can think about
>>
>> 1. if it is possible to include some objects derived from the data
>> rather than the whole data, but that contradicts the aim of the package,
>> as far as I understand.
>>
>> 2. just provide the pdf rather than the Rnw and calculate the vignette
>> for yourself only, but this way the users cannot run the code, if that
>> is relevant.
>>
>> 3. ask CRAN maintainers not to check the vignettes of your package. This
>> is not standard and I am a bit reluctant to suggest exceptions, but
>> since most (if not all) of us have code in place to easily omit vignette
>> checks for particular packages this might be the most appropriate
>> solution here.
>>
>>
>> Kurt is on vacations, so it may take a while to get his response on
>> this. CCing Brian Ripley who runs checks for CRAN but does not read
>> CRAN at ...
>>
>>
>> Best wishes,
>> Uwe
>