[R-pkg-devel] R3.6.0 suppressing S3 overwrite messages

Duncan Murdoch murdoch@dunc@n @end|ng |rom gm@||@com
Sat May 18 22:23:35 CEST 2019


On 18/05/2019 4:03 p.m., Duncan Murdoch wrote:
> On 18/05/2019 3:42 p.m., Roy Mendelssohn - NOAA Federal via
> R-package-devel wrote:
>>
>> My memory is that this was discussed here,  but i can't find the emails, including in the archives ( i did find a similar query from Yihui on r-dev.).  Probably there,  I am just not seeing it.  In the package I am working on for submission,  when I load the package i get:
> 
> 
> There was a thread about this on R-help, started by Rolf Turner on April
> 29, subject "Message produced under R 3.6.0.".
> 
> Users can suppress those messages, but I don't think package writers
> can:  they indicate true bugs, and the packages mentioned (ggplot2,
> crul, hoardr in your case) need to be fixed to get rid of them.

Sorry, that claim isn't correct.  For example, hoardr has a conflict 
with httr, but it's really a weakness of S3 that is causing this:  they 
both define things with class "cache_info", and the warning you're 
seeing is because both have print.cache_info methods.  However, the 
classes are unrelated.  I think the problem is that R can only have one 
print.cache_info method active, and you somehow got both hoardr and httr 
loaded at the same time.

Either one of them could work around this S3 limitation by changing the 
name of the class (e.g. to "hoardr_cache_info" or "httr_cache_info"), 
but I don't think it was fair of me to suggest some flaw in either of 
those packages for the collision.

Duncan Murdoch



More information about the R-package-devel mailing list