[R] Severe growing pains with new help system

Duncan Murdoch murdoch at stats.uwo.ca
Tue Oct 27 10:21:20 CET 2009


On 27/10/2009 5:08 AM, Petr PIKAL wrote:
> Hi
> 
> I have just installed R 2.10.0 version (I used dev version before) and 
> have similar problems with help files.
> starting plain version freshly installed R on WXP
> 
> R version 2.10.0 (2009-10-26)
> Copyright (C) 2009 The R Foundation for Statistical Computing
> ISBN 3-900051-07-0
> 
> ?plot
> 
> only text help appears

There was a bug in the Windows installer (not noticed by anybody during 
the alpha/beta/rc test period) so it doesn't set the default help to 
HTML even if requested to do so.  You can fix it by manually editing 
R_HOME/etc/Rprofile.site according to the comments in it.

> 
>> help.start()
> starting httpd help server ... done
> If nothing happens, you should open
> 'http://127.0.0.1:31757/doc/html/index.html' yourself
> 
> pops up browser window with
> 
> ERROR
> The requested URL could not be retrieved

That's a more serious problem.  That looks as though you can't connect 
to the server.  Some people have said that they need to restrict the 
range of IP ports that it uses (yours used port 31757, but that's a 
random choice in each session).

> While trying to retrieve the URL: 
> http://127.0.0.1:31757/doc/html/index.html
> The following error was encountered:
> Connection to 127.0.0.1 Failed
> The system returned:
>     (111) Connection refused
> The remote host or network may be down. Please try the request again.
> 
> If I try to start help from local /doc/html/packages.html I can start the 
> browser but when trying to access let say plot help I end with
> 
> Tyto webové stránky nebyly nalezeny. 
> # web pages was not found
> 
> Na webové adrese se nepodařilo nalézt žádnou webovou stránku: 
> file:///D:/programy/R-2.10.0/library/graphics/html/plot.html
> # There is no web page  
> file:///D:/programy/R-2.10.0/library/graphics/html/plot.html


Right, that file doesn't exist unless you make a special build to 
contain it (and you'll lose some of the functionality of the new help 
system if you do).

> 
> Více informací o této chybě
> Níže uvádíme původní chybovou zprávu
> # more information about this fault
> 
> Chyba 6(net::ERR_FILE_NOT_FOUND): Soubor nebo adresář se nepodařilo 
> nalézt.
> 
> One workaround could be to copy help files from old version of R but I 
> wonder if this is intended method.

No.

> I apologise if I overlooked some simple solution but I did not find a clue 
> in NEWS and/or CHANGES.
> It may be due to security setting of our network but before I will ask our 
> IT guys I would like to cope with it myself. I could bear with text help 
> but html help is much more convenient.
> 
> BTW. Does this kind of html help work with PC which is not connected to 
> web?

Yes, it should. Address 127.0.0.1 is defined to be the local loopback 
port, so you're connecting to yourself.

Duncan Murdoch

> 
> Best regards
> Petr Pikal
> 
> r-help-bounces at r-project.org napsal dne 27.10.2009 00:05:53:
> 
>> I don't see anything below that is new today.  Are you saying you didn't 
> 
>> test any of the pre-release versions?  You're basically stuck with the 
>> features that come out in .0 releases until the next one.  I'll fix the 
>> typo in 4, but the rest of it is unlikely to change much in the next six 
> 
>> months.
>>
>> Duncan Murdoch
>>
>> On 26/10/2009 6:46 PM, Kevin Wright wrote:
>>> (On Windows XP with R 2.10.0)
>>>
>>> 1. In the past, I have used a desktop search engine (X1) to pre-index 
> all of
>>> R's html help files.  Obviously, this no longer works with dynamic 
> help
>>> pages.  It also appears that the .Rd files are converted to a binary 
> format
>>> when installed and cannot be indexed by X1.
>>>
>>> 2. I like to use firefox to type 'xyplot' into the 'awesome bar' and 
> go to
>>> the appropriate help page.  This obviously no longer works.
>>>
>>> 3. I daily run several different R sessions to keep different projects
>>> separate.  The NEWS file says:
>>>
>>> - The HTML help can now locate cross-references of the form
>>> \link[pkg]{foo} and \link[pkg:foo]{bar} where 'foo' is an
>>> alias in the package, rather than the documented (basename
>>> of a) filename (since the documentation has been much ignored).
>>>
>>> This appears to only be true if the original R session is still 
> active.
>>> Close one R session and the links die in the browser.  Aahg!  What 
> madness!
>>> 4.  The NEWS file says:
>>>
>>> o A new dynamic HTML help system is used by default, and may be
>>> controlled using tools::startDynamicHelp().  With this enabled,
>>> HTML help pages will be generated on request, resolving links
>>> by searching through the current .libPaths().  The user
>>> may set option("help.ports") to control which IP port is used
>>> by the server.
>>>
>>> This should be 'options', of course.  The 'options' page does not 
> reference
>>> "help.ports", but from an email on r-help I see options(help.ports =
>>> sample(6800:6850)).  So I can fix the port at 6800, but the help page 
> for
>>> xyplot still disappears after closing an R session.
>>>
>>> 5. If I want to use --enable-prebuilt-html, I assume that I will have 
> to
>>> download the source and configure/build it myself instead of using the
>>> pre-compiled Windows binaries?  Same with packages?  (Sorry if this is
>>> obvious, I've never needed to do this before.)
>>>
>>>
>>> Looking for a new workflow and remembering why I don't like to use 
> ".0"
>>> releases...
>>>
>>> Kevin Wright
>>>
>>>    [[alternative HTML version deleted]]
>>>
>>> ______________________________________________
>>> R-help at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>> PLEASE do read the posting guide 
> http://www.R-project.org/posting-guide.html
>>> and provide commented, minimal, self-contained, reproducible code.
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide 
> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>




More information about the R-help mailing list