R devel - Nov 2013 - Blank lines in DESCRIPTION

Dear R-devel list,

I would like to enquire about the existence of blank lines in the
DESCRIPTION file.

Testing different possibilities with tools:::.read_description suggests
that starting or ending with blank lines is acceptable while blank lines
in the middle of the file get caught by

out <- tryCatch(read.dcf(dfile, keep.white = .keep_white_description_fields),
        error = function(e) stop(gettextf("file '%s' is not in
valid DCF format",
            dfile), domain = NA, call. = FALSE))
if (nrow(out) != 1) 
    stop("contains a blank line", call. = FALSE)

and produce the error. 

However, the following news item from [1] seems to suggest that it
should not fail in case of such incorrect blank lines.

R CMD check no longer fails with an error if a ?DESCRIPTION? file incorrectly
contains a blank line. (Reported by Bill Dunlap.)

This has come up here [2], when github-specific information is appended
to a DESCRIPTION file that ends with a blank line.

Thank you very much in advance.

Best wishes,

Laurent

[1] http://cran.r-project.org/bin/windows/base/old/2.15.3/NEWS.R-2.15.3.html
[2] https://github.com/hadley/devtools/issues/383

Laurent,

blank lines between records are not allowed in DESCRIPTION, because they
separate paragraphs in DCF and a package must consist of a single paragraph.
This properly is used by R for multi-package DCF files where each paragraph
defines one package. Blank lines at the end may be tolerated, because there is
no second paragraph in that case. If you look at the commit you quote (r60495),
it specifically adds an error if there is more than one paragraph.

Cheers,
Simon


On Nov 26, 2013, at 6:03 PM, Laurent Gatto <lg390 at cam.ac.uk> wrote:
> Dear R-devel list,
> 
> I would like to enquire about the existence of blank lines in the
> DESCRIPTION file.
> 
> Testing different possibilities with tools:::.read_description suggests
> that starting or ending with blank lines is acceptable while blank lines
> in the middle of the file get caught by
> 
> out <- tryCatch(read.dcf(dfile, keep.white =
.keep_white_description_fields),
>        error = function(e) stop(gettextf("file '%s' is not in
valid DCF format",
>            dfile), domain = NA, call. = FALSE))
> if (nrow(out) != 1) 
>    stop("contains a blank line", call. = FALSE)
> 
> and produce the error. 
> 
> However, the following news item from [1] seems to suggest that it
> should not fail in case of such incorrect blank lines.
> 
> R CMD check no longer fails with an error if a ?DESCRIPTION? file
incorrectly contains a blank line. (Reported by Bill Dunlap.)
> 
> This has come up here [2], when github-specific information is appended
> to a DESCRIPTION file that ends with a blank line.
> 
> Thank you very much in advance.
> 
> Best wishes,
> 
> Laurent
> 
> [1]
http://cran.r-project.org/bin/windows/base/old/2.15.3/NEWS.R-2.15.3.html
> [2] https://github.com/hadley/devtools/issues/383
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel