make.packages.html

It used to be that, whenever I added or updated a new package, the
file /usr/lib/R/doc/html/packages.html would be updated (on Linux).
Now I find that this does not happen anymore.

It does happen, but only if you install a package to the .Library 
directory.  That file is only intended as a listing of the main library, 
not other libraries (as which are in the library path is 
session-specific).  As NEWS for R 2.6.0 says

     o	install.packages() on a Unix-alike now updates
 	doc/html/packages.html only if packages are installed to
 	.Library (by that exact name).

So I found this function make.packages.html, which seems to do what I 
think should be done automatically.  But it puts its output in a file in 
/tmp/...,

It is a helper for help.start(), which sets up details of the current 
packages in the session temporary directory.  Please do consult its help 
file.

The function used to update doc/html/packages.html is 
utils:::link.html.help(), which is a wrapper for 
tools:::unix.packages.html.

which I then have to move, in an extra step, to where it should be.  I
don't understand why this happens.  I think that the old behavior was
better.

It caused a lot of needless updating.

I have two reasons for this.  First, my R page (finzi.psych.upenn.edu)
has a list of all packages.  As it happens, since this change has
occurred, it has been out of date, until today.  Second, for my own
use, I have a Firefox bookmark to
file:/usr/lib/R/doc/html/packages.html, which was also out of date.

Let's hope you have a vaiid file:// URL ....

I don't subscribe to r-devel, so, in case this gets posted, please use
group reply.

What we don't know is where you are installing your packages.  If it is to 
the main library, the behaviour should be unchanged (and I have just 
tested it and it works for me).  If it is to another library, the 
intention never was to list packages in other libraries.

[Note that all this applies to Unix-alikes: Windows does not have symbolic 
links (OK, some filesystems on some versions have a very limited form) and 
R does things differently there.]

Jon
-- 
Jonathan Baron, Professor of Psychology, University of Pennsylvania
Home page: http://www.sas.upenn.edu/~baron

Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595