PATCH: Add fields argument to installed.packages and available.packages
Rather than a plethora of fields, perhaps the Depends field could indicate what depends on the object: For example, if we use file extensions to indicate what is dependent then one might write this to indicate that some .Rd (i.e. examples) and .Rnw (i.e. vignette) files depend on lattice and the entire package depends on zoo and the package is related to but not dependent on tseries: Depends: lattice (.Rd, .Rnw), grid (.Rnw), zoo, tseries (0) Then there could be rules for each such suffix when processing the package. This has the advantage that its meaning is more obvious than a bunch of keywords (Depends, Suggests, CanUse, Related).
On 8/29/06, Martin Maechler <maechler at stat.math.ethz.ch> wrote:
"Seth" == Seth Falcon <sfalcon at fhcrc.org>
on Mon, 28 Aug 2006 17:42:39 -0700 writes:
Seth> Hi all, The write_PACKAGES function has a 'fields'
Seth> argument that allows a user generating a PACKAGES file
Seth> to specify additional fields to include. For
Seth> symmetry, it would be nice for the available.packages
Seth> function to be able to read those extra fields when
Seth> specified.
Seth> Similarly, it would be useful for installed.packages
Seth> to have a 'fields' argument. This would allow a user
Seth> to query the set of installed packages for other
Seth> fields in the DESCRIPTION file.
Seth> One use for this would be for repository hosters to
Seth> include the License field in their PACKAGES file. In
Seth> this way, end users could query a repository and only
Seth> download packages that they have a license to use.
Seth> Below is a patch against svn 38996 that attempts to
Seth> implement this.
I like the idea and will look into applying the patch
(note there's at least one typo which makes "make check" fail:
/priotiry/)
A propos:
A while back (in last summer?), we (some of R-core) have
discussed about a new field to be added to DESCRIPTION,
and AFAIR, the only problem we had, is to find a name we
all liked.
Or was there more then the name alone, and some where convinced
that it is superfluous and hence too complicated.
The idea was a field related to but weaker than 'Suggests' :
Something like
'canMakeUseOf: <pkg1> [, <pkg2>, ... ]
which is *not* used in any QA/QC checking, but is purely
informative: If <pkg1> is require()able, then some examples may
look nicer, a function may provide another feature, etc, etc.
Alternatives to 'canMakeUseOf' would have been
'isHappilyCoworkingWith' ....
What do you (R-devel listeners) think about the idea?
Martin
______________________________________________ R-devel at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel