Consolidates the 0.3.3.x development cycle.
check_description_cran() and
fix_description_cran(), along with the web-service-link
check (#23). A documentation generator should not lint DESCRIPTION
prose, and the checks produced false positives (e.g. flagging the
ordinary word “graphics”). The token-based code checker and example
checks (check_cran(), check_examples_cran())
remain, though check_cran()’s result list no longer carries
a description element.document() gains stale-Rd pruning: after regenerating
it removes man/*.Rd pages for renamed or deleted topics,
but only files tinyrox owns (first line is the tinyrox marker);
hand-written Rd is never touched. Pass prune_rd = FALSE to
disable (#22).@returns as a plural alias of
@return (#24).##' doc comments in addition to
#', matching roxygen2’s #+' (Dirk
Eddelbuettel, #27).document() for the whole package.utils::getParseData()) instead of raw source lines,
eliminating false positives from comments, strings, and look-alikes such
as torch.cat() (#20).#'
lines: an orphaned block’s @export can no longer bleed into
the next function (#18), and document() warns instead of
silently dropping a hand-added NAMESPACE directive (#17).@section blocks in the Rd for ordinary functions
and @rdname groups (#10).@rdname group
(#12).utils:::.getHelpFile() call with
tools::parse_Rd() for CRAN compliance.@return to clean(),
@examples to parse_tags().@rdname grouping of multiple functions into
one Rd file.check_cran() and
fix_description_cran().parse_tags() for programmatic access to parsed
documentation.