Skip to content

Comments in the DESCRIPTION file

6 messages · Simon Urbanek, Christophe Genolini

#
Hi all,

Since R 3.0.0, the '#' does no longuer works for comments.
But as noticed above, comments can be introduce by 'anyWord:' or
'myComment;' or 'toto:'.

Since '#' can be used to name a field, we can also used '#:' or '###:' (or
even '%:' for LaTeX's fans).

So '#:' is a new possible way for adding comments in DESCRIPTION file.

Christophe



--
View this message in context: http://r.789695.n4.nabble.com/Comments-in-the-DESCRIPTION-file-tp4648678p4665892.html
Sent from the R devel mailing list archive at Nabble.com.
#
On May 1, 2013, at 6:01 AM, cgenolin wrote:

            
No, it's not, they are not permitted -- please read the DCF specs:

"Field names must not begin with the comment character, #."

Cheers,
Simon
#
Hi,

I am not that familiar with the DCF... But R seems to accept # quite easely.
More precisely:

Before posting my message, I try it on a small package: R CMD check or R CMD
INSTALL did not make any warning or error (whereas they do if I use '#'
alone). 

After reading your response, I dig a bit more. The source of my DESCRIPTION
file was:

--- 8< ------------------
#############################:
###: This is some kind of 'section 1'
###:
Package: packBasic1
Title: Very simple package
Version: 0.9.2
License: GPL (>=2.0)
Description: A package
#:
#:
#############################:
###: This is 'section 2'
###:
Author: Christophe Genolini
Maintainer: Christophe Genolini <genolini at u-paris10.fr>
--- 8< ------------------

The DESCRIPTION file after installation was: 

--- 8< ------------------
#############################:
###:
Package: packBasic1
Title: Very simple package
Version: 0.9.2
License: GPL (>=2.0)
Description: A package
#:
Author: Christophe Genolini
Maintainer: Christophe Genolini <genolini at u-paris10.fr>
Built: R 3.0.0; ; 2013-05-02 06:50:57 UTC; windows
--- 8< ------------------

So I guess there was a problem.
But if I number the comments line, then it works:

--- 8< ---------------------
#############################1:
###2: This is some kind of 'section 1'
###3:
Package: packBasic1
Title: Very simple package
Version: 0.9.2
License: GPL (>=2.0)
Description: A package
#4:
#5:
#############################6:
###7: This is 'section 2'
###8:
Author: Christophe Genolini
Maintainer: Christophe Genolini <genolini at u-paris10.fr>
Built: R 3.0.0; ; 2013-05-02 06:49:27 UTC; windows
--- 8< ---------------------

Christophe



--
View this message in context: http://r.789695.n4.nabble.com/Comments-in-the-DESCRIPTION-file-tp4648678p4666017.html
Sent from the R devel mailing list archive at Nabble.com.
#
On May 2, 2013, at 2:58 AM, cgenolin wrote:

            
The fact that R currently accepts invalid DCF files is not a guarantee that it won't be following the standard more closely in the future. What you are doing is illegal in the DCF standard so don't do it. If you do so, don't be surprised that it will fail tomorrow, you have been warned.

Cheers,
S
#
So it IS curently accepted, but may NOT be in the futur. Thanks for your
answer.

So I guess that using 
--- 8< -------------
%%%%%%%%%1: 
%%%2: Section 1
%%%3
...
--- 8< -------------

is correct, isn't it?

Christophe



--
View this message in context: http://r.789695.n4.nabble.com/Comments-in-the-DESCRIPTION-file-tp4648678p4666058.html
Sent from the R devel mailing list archive at Nabble.com.
#
On May 2, 2013, at 10:17 AM, cgenolin wrote:

            
Yes, that is legal (assuming a colon after the 3).

FWIW this is legal as well:

A######:
B###: section 1

The two relevant requirements are:

"A paragraph must not contain more than one instance of a particular field name." and "Field names must not begin with the comment character, #."

see 5.1 Syntax of control files:
http://www.debian.org/doc/debian-policy/ch-controlfields.html

Cheers,
Simon