Implict rows/cols>0 assumption in arr_parts_combine

[MichaelChirico]

repr::repr_text(matrix(nrow = 0, ncol = 100))
# Warning message:
# In cbind(parts$left, ellip_h, parts$right, deparse.level = 0L) :
#   number of rows of result is not a multiple of vector length (arg 2)

Ditto

repr::repr_text(matrix(ncol = 0, nrow = 100))
# Warning message:
# In rbind(parts$upper, ellip_v, parts$lower, deparse.level = 0L) :
#   number of columns of result is not a multiple of vector length (arg 2)

The diagnosis is simple enough; here:

repr/R/repr_matrix_df.r

Lines 84 to 85 in 7f27d6a

	rows = rbind(parts$upper, ellip_v, parts$lower, deparse.level = 0L),
	cols = cbind(parts$left, ellip_h, parts$right, deparse.level = 0L),

rbind/cbind is combining 0-length input with length-1 input (ellip_v/ellip_h).

I'm not 100% sure the correct behavior/best fix, however.

[MichaelChirico]

Bumping here, it's still a pretty common issue to wind up with dim-0 results and get spurious warnings.

Any thoughts on the right solution @flying-sheep? Happy to file a PR with a solution