[Rd] 'R CMD check' gets apparently confused by '%' (PR#8198)

Kurt Hornik Kurt.Hornik at wu-wien.ac.at
Mon Oct 10 21:01:01 CEST 2005


>>>>> lgautier  writes:

> Full_Name: Laurent
> Version: R-2.0.0-beta-from-a-week-ago
> OS: linux
> Submission from: (NULL) (202.156.6.77)


> When running 'R CMD check' for a package that contains the symbol '%'
> in the signature of a function (say the default parameter for a string
> argument is like "10% - 20%"), the 'codoc' part of the check throws a
> warning about mismatching doc and codoc.

> While the '%' has apparently to be escaped in the .Rnw (otherwise
> latex soon figures out that what is after on the same line is a
> comment), the warning remains whether or not '%' is escaped in the
> code.

> The package 'splicegear' in Bioconductor (in the section "devel") is
> what brought me to submit a bug report.

If we run your splicegear from BioC devel (1.1.0) through R CMD check
from a current version of R (as you are actually supposed to when filing
a bug report), we get

* checking Rd \usage sections ... WARNING
Bad \usage lines found in documentation object 'buildSpliceSites':
  queryPALSdb(query, disp = c("data", "browser"),
              field = c("keyword", "ug_id", "gb_id", "cluster_count"),
              species = c("human", "mouse"),
              e.value = "1e-1",
              ident.threshold = c("90
              verbose = FALSE)
  getPALSdbURL(query, disp = c("data", "browser"),
              field = c("keyword", "ug_id", "gb_id", "cluster_count"),
              species = c("human", "mouse"),
              e.value = "1e-1",
              ident.threshold = c("90
              verbose = FALSE)
Bad \usage lines found in documentation object 'grid.expand.gp':
  grid.numeric2npc(x, xlim=NULL, lower.blank=0, upper.blank=0
Bad \usage lines found in documentation object 'grid.plot.Probes':
  grid.plot.SpliceExprSet(function (x, probes.opt = list(), expr.opt = list(col = NA, lty = 1:6),
      fig.xratio = c(2, 1), fig.yratio = c(2, 1), probepos.yscale = NULL,
      ylim = NULL, ...)

as the usage entries are syntactically invalid.

This is because (see R-exts) % in Rd files *ALWAYS* indicates an Rd
comment:

        The "comment" character `%' and unpaired braces(1) _always_ need
	to be escaped by `\' ...

in your case, your

splicegear/man/buildSpliceSites.Rd:            ident.threshold = c("90% 50b", "95% 50b", "90% 45b"),
splicegear/man/buildSpliceSites.Rd:            ident.threshold = c("90% 50b", "95% 50b", "90% 45b"),

should use

  ident.threshold = c("90\% 50b", "95\% 50b", "90\% 45b"),

Hth
-k



More information about the R-devel mailing list