[Rd] Problem with sink() in 1.20 on Windows (PR#779)

[Jens Oehlschlägel]

Dear Prof. Ripley,

You wrote
> Yes, but what is the bug exactly?  

Just testing connections, I think the bug is that the help for sink says that:

  If `file' is a connection if will be opened if necessary.
     Switching to another file or connection closes the existing
     connection unless it is the console.

and after switching to another file or connection the existing connection is not closed.

> showConnections()
     class description mode text isopen can read can write
> sink("tempfile")
> sink()
> showConnections()
  class      description mode text   isopen   can read can write
3 "tempfile" "file"      "w"  "text" "opened" "no"     "yes"    
> # now we have silently an open connection left

I tend to think that a sequence of

 sink(WhateverConnectionInvolved)
 :
 :
 sink()

should leave WhateverConnectionInvolved in the same state it was before the sequence.
i.e. if WhateverConnectionInvolved was non.existent it should be non.existent
i.e. if WhateverConnectionInvolved was existent but not opened it should be existent but not opened 
i.e. if WhateverConnectionInvolved was existent and opened it should be existent and opened 

BTW: this may be considered as a special case of a more general thing: 
undefined rights to modify/access connections owned by other parts of the software.


Regards

Jens Oehlschlägel


--please do not edit the information below--

Version:
 platform = i386-pc-mingw32
 arch = x86
 os = Win32
 system = x86, Win32
 status = 
 major = 1
 minor = 2.0
 year = 2000
 month = 12
 day = 15
 language = R

Windows NT 4.0 (build 1381) Service Pack 6

Search Path:
 .GlobalEnv, package:ctest, Autoloads, package:base

_______________________________________________________________________________
Alles unter einem Dach: Informationen, Fun, E-Mails. Bei WEB.DE: http://web.de
Die große Welt der Kommunikation: E-Mail, Fax, SMS, WAP: http://freemail.web.de

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._