[R-SIG-Mac] Can't get help.start() to work --- update.

[Prof Brian Ripley]

On Thu, 29 Oct 2009, Duncan Murdoch wrote:

> On 29/10/2009 1:07 AM, Rolf Turner wrote:
>> After looking at the help for tools::startDynamicHelp and groping
>> around considerably, I had a look at the Firefox Preferences.
>> Going to Advanced --> Network --> Configure how Firefox connects
>> to the the Internet (Settings), I found stuff about ``Proxies''.
>> (Dunno what they are, but the word ``proxy'' came up in the error
>> message from Firefox.)
>> 
>> The existing setting was ``Manual proxy configuration''.  I changed
>> to ``No proxy'' (hammer and hope; no idea what I'm doing) and
>> help.start() took effect without any error messages appearing
>> from Firefox.
>> 
>> But I still got help in the command window.  Then I set
>>
>> 	options(help_type="html")
>> 
>> and it all worked.
>> 
>> I shall put options(help_type="html") into my .Rprofile, and
>> then help.start() should work automatically.  Said he optimistically.
>> 
>> What trouble am I leaving myself open to by setting ``No proxy'', but?
>
> You may find that you can't connect to websites outside of your local 
> network.  Proxies are used to control connections in and out of networks. 
> Without the proxy you might be blocked; with it, some sites will be blocked, 
> and others will be redirected through some server that lets them out.  It's 
> also possible there will be no trouble, because the same scheme is sometimes 
> used without the external blocking, but then usually the user of the system 
> can't turn it on and off.
>
> You should be able to program the proxy to be selective, and let Firefox 
> connect to R directly but still redirect external connections to other sites, 
> but you'll need to take some action to do it.

In my Mac Firefox, the default for 'Manual proxy configuration' is to 
have no-proxy exceptions for 'localhost, 127.0.0.1', which is exactly 
what you need -- these are set on the same screen.  (And I tested it 
with a proxy for HTTP, and help.start worked.)  So selectivity is the 
default.


-- 
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