new chkDots warnings from splom() in R-devel

[bastistician]

Modified from example(splom):

lattice::splom(~iris[1:4], groups = Species, data = iris, cex = 0.5)

In R >= r88881, this produces

Warning messages:
1: In pretty.default(x = base::quote(c(4.3, 7.9)), i = base::quote(1), 
    j = base::quote(1), cex = base::quote(0.5), n = base::quote(5)) :
 extra arguments ‘i’, ‘j’, ‘cex’ will be disregarded 
2: In pretty.default(x = base::quote(c(2, 4.4)), i = base::quote(2), 
    j = base::quote(2), cex = base::quote(0.5), n = base::quote(5)) :
 extra arguments ‘i’, ‘j’, ‘cex’ will be disregarded 
3: In pretty.default(x = base::quote(c(1, 6.9)), i = base::quote(3), 
    j = base::quote(3), cex = base::quote(0.5), n = base::quote(5)) :
 extra arguments ‘i’, ‘j’, ‘cex’ will be disregarded 
4: In pretty.default(x = base::quote(c(0.1, 2.5)), i = base::quote(4), 
    j = base::quote(4), cex = base::quote(0.5), n = base::quote(5)) :
 extra arguments ‘i’, ‘j’, ‘cex’ will be disregarded 

because pretty.default now does chkDots(...) -- intended to catch typos, see @mmaechler's comment in https://bugs.r-project.org/show_bug.cgi?id=18521#c4 -- so no longer simply ignores extra arguments.

[mmaechler]

aarghh... and indeed, in r88882, I've made an amendment, IIRC, just to satisfy another such lattice case ...
But this one did not show in R's make check-all (which does run all the lattice checks and examples, as lattice is formally recommended).

It would really be good and nice if this could be fixed in lattice code
(also the warning messages suggest that this part of the code may a bit "unusual", somehow applying all the base::quote(*) wrappers ??)

[mmaechler]

Of course, this happens in plot.trellis() which is called from the print() method:

> grb <- lattice::splom(~iris[,1:3], groups = Species, data = iris, cex = 0.5)
> lattice:::plot.trellis(grb)
Warning messages:
1: In pretty.default(x = base::quote(c(4.3, 7.9)), i = base::quote(1), 
    j = base::quote(1), cex = base::quote(0.5), n = base::quote(5)) :
 extra arguments ‘i’, ‘j’, ‘cex’ will be disregarded 
2: In pretty.default(x = base::quote(c(2, 4.4)), i = base::quote(2), 
    j = base::quote(2), cex = base::quote(0.5), n = base::quote(5)) :
 extra arguments ‘i’, ‘j’, ‘cex’ will be disregarded 
3: In pretty.default(x = base::quote(c(1, 6.9)), i = base::quote(3), 
    j = base::quote(3), cex = base::quote(0.5), n = base::quote(5)) :
 extra arguments ‘i’, ‘j’, ‘cex’ will be disregarded 
> 

[deepayan]

Yes, the "nint" case we discussed earlier is a little difficult to work around. But this one should be easy. ?diag.panel.splom says it will ignore ... --- but it actually didn't. Fixed now.

[mmaechler]

Thanks a lot! As these warnings happen in my R CMD check <pkg> checking when using R-devel, building vignettes for package copula https://CRAN.R-project.org/package=copula ; (even though not when the package is checked on CRAN, I don't know why), I'd be grateful if you could plan a CRAN release of lattice 0.22-8 somewhat soonish..