increase coverage of standard mathematical constants increase coverage of special mathematical functions random number generation, something like ‘sample’, ‘runif’ work around missing ‘arf’, ‘acf’ functionality formatting of floating-point types suppressing trailing zeros formatting of ‘arb’ with two ‘digits’ for midpoint, radius ‘names’ slot for the virtual class flintSize for getting the size of an allocated array flintIdentical testing for identical-modulo-address export, document flintIdentical abandon ‘nflint’ and subclasses methods for !, is.na, is.nan, is.finite, is.infinite export, document Num, Den, Mid, Rad, Real, Imag methods for as.character using format initialization with character vectors coercion to the virtual class methods for unique, duplicated, anyDuplicated using mtfrm x[i] <- value should conditionally promote ‘x’ to class(value) methods for Summary, mean not ignoring ‘…’ methods for as.POSIXlt, as.POSIXct, as.Date methods for cut, findInterval methods for seq, sequence methods for summary, quantile methods for as.data.frame register an S3 method format.flint for use by format.data.frame document generic functions whose default methods seem to work for us as.character, as.symbol, as.pairlist, as.list, as.expression,
diff, labels, rev, seq.int, sort, sort_by, split, split<-, subset,
toString
document details of operations on real, complex balls check for memory bugs: Valgrind, ASan, UBSan, rchk ideally after increasing test coverage
methods for ‘mag’ consulting rounding mode, conditionally giving lower bound export, document class generator functions configure tests for many missing acf_*, mag_* abandon 64-bit lengths on 32-bit builds support ‘flint’ subscripts of ‘flint’ vectors support ‘flint’ subscripts of traditional vectors fix all.equal(check.attributes=TRUE) fix coercions to character, symbol, pairlist, list, expression