IE5 html help search failure (PR#205)

[guido@hal.stat.unipd.it]

On Thu, Jun 03, 1999 at 03:56:15PM +0200, fears@roycastle.liv.ac.uk wrote:
> I hesitate to report this one because it is very likely system
> dependent, and I have very little faith in my browser. Still, it may
> point to something useful for someone to know.

  You did the right thing. It was one of my (many?) faults.
  I overlooked the effect of one of the last changes done before
  preparing the release.

> 
> On NT4, after help.start() brings up the web interface just fine
> (despite the pessimistic warning on the terminal; is that still
> necessary?), I can do apparaently anything in the help system, except
> search.

  help.start() uses standard Windows association between file type and
  application,i.e., if html files are not associate to a browser 
  it fails. This is the reason of the warning. I have not looked
  to the source but I suppose that we can modify actual behaviour
  to give the warning only when actual implementation fails.  
                      ~~~~

> 
> If I'm on the top search page page
> (E:\programs\R\rw0641\doc\html\search\SearchEngine.html) and click on
> any topic, say 'aplot', I go to 'address'
> javascript:searchInIndex%20('%20aplot') and Explorer says 'Done', but
> nothing happens. Click on 'Go back' and I get 'No matches for aplot have
> been found'.
> 

  The distribution misses a file used by the search engine. 
  However, at least on my system, it can be generated on the fly executing 
  > link.html.help()
  from a running R process (or from Rgui or from Rterm)
  Indeed, to reproduce the behaviour you reported, I had to
  do a fresh installation. On my "real' installation, since I
  added some packages and runned "link.html.help()", all went well. 


> I've tried several others, and also typing easy terms such as 'plot'
> into the dialog box, all to no avail.
> 
> Now here's the bit where you might get cross with me. I'm using MS
> Internet Explorer 5, and I can't readily confirm the same thing happens
> with a standard release such as 4, since our local administrators
> decided in their wisdom to remove 4 when they installed 5. And of course
> it may be a problem with our local java installation too (or as well).
> I'm not in charge of that, but if anyone  can guess as to what might be
> set up wrong at least perhaps I can point the IT helpdesk in the right
> direction.
> 

  Since we are on the topic (search and version of IE), I noted
  that using IE4 on a Win98 machine, when I click on 'aplot' or
  friends I get two html pages, the first with the rigth informations,
  the second blank. So, I have to click on the "previous page"
  button to look to the results. On the same machine and on the same 
  R installation, if I use Netscape 4.04 as the default browser
  all works as expected (i.e, no 'second' page). So, my
  conclusion in this moment is that the problem is due to the
  IE4 Java VM not to R. I haven't test IE5 but I would be glad to
  know what happens.

  guido










-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._