On 29/12/2016 6:50 PM, Duncan Murdoch wrote:
On 29/12/2016 2:58 PM, Roy Mendelssohn - NOAA Federal wrote:
HI All:
If any cares, the offending line is in the ggplot2 file utilities.r where it
has:
#' A waiver object.
#'
#' A waiver is a "flag" object, similar to \code{NULL}, that indicates the
#' calling function should just use the default value. It is used in certain
#' functions to distinguish between displaying nothing (\code{NULL}) and
#' displaying a default value calculated elsewhere (\code{waiver()})
#'
#' @export
#' @keywords internal
waiver <- function() structure(NULL, class = "waiver")
is.waive <- function(x) inherits(x, "waiver")
If you then do a multi-file search on "waive" you find that it is used all
throughout the code base, so anytime waiver() is called, so is the NULL structure.
That usage will work: though waiver() tries to put a class on NULL,
structure() puts it on a length-zero list instead. R-devel's complaint
is unnecessary here, but could be avoided with the code change
waiver <- function() structure(list(), class = "waiver")
I've cc'd Hadley to let him know.
Hadley submitted a patch, which is on CRAN now (version 2.2.1). It will
take a little while to propagate to the mirrors, but things should be
fine in your package once you update ggplot2.
Duncan Murdoch
Duncan Murdoch
-Roy
On Dec 29, 2016, at 11:44 AM, Roy Mendelssohn - NOAA Federal
wrote:
Thanks. I can wait. I will also see if I can figure out a work around in the
meantime. Is the submission to winbuild automated? That is really the only
way I have to check, but I don't want to keep on checking if someone's time is
being wasted by that. If automated, the I don't mind making repeated
submissions.
Thanks again,
-Roy
On Dec 29, 2016, at 11:40 AM, Duncan Murdoch wrote:
On 29/12/2016 1:44 PM, Roy Mendelssohn - NOAA Federal wrote:
Thanks, but as I said, my next question is how best to proceed with CRAN. I
do not want to waste peoples' time with a submission that I know before hand
will be rejected. Can I submit with this warning?
I'd recommend waiting a few days. If you really want to submit soon, you can
figure out which is the offending command, and wrap it in something that
suppresses the warning (e.g. suppressWarnings(), assuming the conversion to an
error happens later).
I don't think your package would be accepted if the vignette won't build on
R-devel.
Duncan Murdoch
-Roy
On Dec 29, 2016, at 10:38 AM, Duncan Murdoch wrote:
On 29/12/2016 1:24 PM, Ben Bolker wrote:
Sorry, correction/clarification to my last post: it's *not* a bug in
ggplot2, rather apparently (?) it's something in base R that has broken
tests in both data.table and ggplot2. (Since your code calls ggplot,
though, it's presumably in there somewhere, and (?) not your problem.)
The NEWS item is here:
http://developer.r-project.org/blosxom.cgi/R-devel/2016/12/28#n2016-12-28
The issue is discussed in a bit more detail on R-devel (subject "[Rd] Unexpected I(NULL)
output") and .
Overall it seems like a good idea: some code (perhaps in ggplot2, I haven't
tried to track it down) appears to be trying to set attributes on NULL. This
will silently fail in versions of R prior to R-devel rev 71841, and will fail
with a warning in that version or later. (Since the change to R-devel is very
recent, it may change again.)
Duncan Murdoch
cheers
Ben
On 16-12-29 01:16 PM, Roy Mendelssohn - NOAA Federal wrote:
Hi All:
I am working on a new submission of my xtractomatic package (the
main change being the use of https). I develop on a Mac. When I run
on the Mac:
devtools::check()
I get no errors, notes, or warnings. However, when I run
devtools::build_win(), the response I get back isL
* checking for unstated dependencies in vignettes ... OK * checking
package vignettes in 'inst/doc' ... OK * checking re-building of
vignette outputs ... WARNING Error in re-building vignettes: ...
Quitting from lines 248-261 (Usingxtractomatic.Rmd) Error:
processing vignette 'Usingxtractomatic.Rmd' failed with
diagnostics: (converted from warning) Calling 'structure(NULL, *)'
is deprecated, as NULL cannot have attributes. Consider
'structure(list(), *)' instead. Execution halted
So this error is from my Vignette. The offending lines appear to
be:
```{r topotagPlot, fig.align = 'center', fig.width = 6, fig.height
= 4, warning = FALSE} require("ggplot2") alldata <- cbind(tagData,
topo) alldata$lon <- alldata$lon - 360 z <- ggplot(alldata, aes(x =
lon,y = lat)) + geom_point(aes(colour = mean), size = 2.) +
scale_shape_manual(values = c(19, 1)) z + geom_polygon(data = w,
aes(x = long, y = lat, group = group), fill = "grey80") +
theme_bw() + scale_colour_gradient("Depth") + coord_fixed(1.3, xlim
= xlim, ylim = ylim) + ggtitle("Bathymetry at marlin tag
locations")
```
Now the vignette builds fine on my machine, and I ca
