Skip to content

[R-pkg-devel] warnings by 3rd party headers

5 messages · Guido Kraemer, Dirk Eddelbuettel, Iñaki Ucar

#
I need to distribute some 3rd party headers with my package. The headers 
make R CMD check produce warnings:

    File which contains pragma(s) suppressing important diagnostics:
    ...
    Files which contain pragma(s) suppressing diagnostics:
    ...

There are some additional NOTEs on the use of `std::cout` which is being 
used in some of the headers.

If I want to submit the package to CRAN in the future, do I have to edit 
the headers?

Cheers,

Guido
#
On Thu, 16 Jan 2020 at 16:58, Guido Kraemer <gkraemer at bgc-jena.mpg.de> wrote:
I think so. That's what, e.g., Dirk does for BH:
https://github.com/eddelbuettel/bh/commit/313da3cb04227d96daafa2d73668b30fa3fd10fc

I?aki
#
On 16 January 2020 at 17:06, I?aki Ucar wrote:
| On Thu, 16 Jan 2020 at 16:58, Guido Kraemer <gkraemer at bgc-jena.mpg.de> wrote:
| >
| > I need to distribute some 3rd party headers with my package. The headers
| > make R CMD check produce warnings:
| >
| >     File which contains pragma(s) suppressing important diagnostics:
| >     ...
| >     Files which contain pragma(s) suppressing diagnostics:
| >     ...
| >
| > There are some additional NOTEs on the use of `std::cout` which is being
| > used in some of the headers.
| >
| > If I want to submit the package to CRAN in the future, do I have to edit
| > the headers?
| 
| I think so. That's what, e.g., Dirk does for BH:
| https://github.com/eddelbuettel/bh/commit/313da3cb04227d96daafa2d73668b30fa3fd10fc

=:-)

Beat me to it. Was about to show the last diff:
https://github.com/eddelbuettel/bh/blob/master/local/patches/bh_1.72.0-0.diagnostics.patch

The policy is, if I may use salty language for once, idiotic^Ha really bad
call. But I don't get to make the policy so I just live by it.

So yes Guido, you have to.  And again on each upstream upgrade you cover.

Dirk
1 day later
#
Thanks, just to drive this a bit further:

What if the headers then start producing some ugly compile warnings? Do 
I have to fix these, too?
On 1/16/20 5:12 PM, Dirk Eddelbuettel wrote:
#
On Fri, 17 Jan 2020 at 20:42, Guido Kraemer <gkraemer at bgc-jena.mpg.de> wrote:
How ugly? You can ignore "variable set but not used"-kind of stuff,
but if things turn ugly enough, a good citizen should take a look into
it and maybe contribute a patch upstream. But of course that's not
CRAN policy, that's my opinion.

I?aki