[rkward-devel] dev.off () bug

[Prasenjit Kapat]

Hi,

When you find some time, here is some more information:
this bug seems to be related to KDE 4.4.x series; see below.

On Mon, Jun 21, 2010 at 4:17 PM, Prasenjit Kapat <kapatp at gmail.com> wrote:
> On Mon, Jun 21, 2010 at 12:09 PM, Thomas Friedrichsmeier
> <thomas.friedrichsmeier at ruhr-uni-bochum.de> wrote:
>> On Monday 21 June 2010, Prasenjit Kapat wrote:
>>> Calling dev.off() creates a new cairo device instead of closing the
>>>
>>> existing one:
>>> > plot (1,1)
>>> > dev.off()
>>
>> I don't see this, here. Do you get this every time or only sometimes?
>
> Everytime.

Case 1: Kubuntu 9.10 in a virtualbox; compiled rkward trunk against R
2.9.2 as well as 2.11.1 (from cran). In both cases dev.off () worked
fine. Note here that Kubuntu 9.10 has KDE 4.3.2

Case 2: Debian/Sid moved to KDE 4.4.3 very recently; so even near the
end of April / beginning of May my system had KDE 4.3.y and I hadn't
noticed this bug

Case 3: Kubuntu 10.04 in a virtualbox; ran stock rkward-0.5.2 against
stock R 2.10.2. KDE is 4.4.2 and the bug shows its ugly head ;)

On Mon, Jun 21, 2010 at 6:24 PM, Thomas Friedrichsmeier
<thomas.friedrichsmeier at ruhr-uni-bochum.de> wrote:
> On Monday 21 June 2010, Prasenjit Kapat wrote:
>> I do have a wrapper around dev.off: testable as soon as the bug is
>> taken care of.
>
> Ok, then here's an idea for a hack around the bug: Directly before the "real"
> dev.off() add
>   .rk.lock.device <- TRUE
> and
>   .rk.lock.device <- FALSE
> directly after. In rk.screen.device add
>   if (exists (.rk.lock.device) && .rk.lock.device) return
> at the start.

Well, this didn't help :( Issue persists.

Let me know if you need any more info (when you find time).
Regards
-- 
Prasenjit