[Rd] library( ) and verbose=
Dirk Eddelbuettel
edd at debian.org
Sun Jul 11 20:09:41 CEST 2004
On Sun, Jul 11, 2004 at 04:23:10PM +0200, Simon Urbanek wrote:
> On Jul 11, 2004, at 1:07 PM, Frank E Harrell Jr wrote:
>
> >The Hmisc package has in its .First.lib function a verbose argument,
> >which I thought was the way to allow users to suppress certain
> >messages,
> >by issuing library(Hmisc, verbose=FALSE). But I see that library( )
> >does not pass verbose to .First.lib. The default for verbose in Hmisc
> >is TRUE because .First.lib prints some important information about
> >masked functions. What is the best way to fix this?
>
> AFAIR the official way to suppress unwanted messages is
> warn.conflicts=FALSE so the user probably wants to issue
> library(Hmisc, warn.conflicts=FALSE).
Not bad, but note that Hmisc has a very explicit
.First.lib <- function(lib, pkg, verbose=TRUE, ...) {
if(verbose)
cat("Hmisc library by Frank E Harrell Jr\n\n",
"Type library(help='Hmisc'), ?Overview, or ?Hmisc.Overview')\n",
[...]
so the additional option() Brian suggested way indeed be the best way.
> As of 'verbose' - it is not by default TRUE - it's actually determined
> by the current "verbose" option, so users can actually change the
> default as desired.
I tried that several different ways before I brought this to Frank's
attention. No method worked; I ended up altering the R code in the installed
package to set the verbose=FALSE in the snipped shown above.
> On the other hand .First.lib takes only two, fixed arguments, so AFAIK
> you cannot pass any additional arguments from outside (I believe this
> applies to .onLoad and friends, too).
> In fact I'm not quite sure what you want to "fix" here .. is it the
> wish in general to be able to pass arguments to packages while being
> loaded? In your specific case I'd say you don't need that since you
> have always access to the 'verbose' option ...
Quite so.
Dirk
--
Those are my principles, and if you don't like them... well, I have others.
-- Groucho Marx
More information about the R-devel
mailing list