[rkward-devel] Request for feedback on plot history

[Prasenjit Kapat]

Hi,

I was playing around with the plot history feature of Windows (windows
()) and Mac OS X's (qaurtz ()), something that I shuold've done much
earlier. They are both very minimalistic, although windows' is
slightly better. I realized that we were shooting a "near ideal"
solution, which is going to be next-to-impossible w/o breakage-prone
hacks.

On Thu, Sep 2, 2010 at 2:55 AM, Thomas Friedrichsmeier
<thomas.friedrichsmeier at ruhr-uni-bochum.de> wrote:
>
> 2) *Always* save to the history on plot.new, prev/next, etc. without the whole
> newPlotExists mechanism. The advantage is again that it offers much less
> potential for surprises. It does seem pretty wasteful, since most of the time,
> the plots are not actually modified, but we still record them again and again
> while browsing the history. On the other hand, recording plots appears to be
> pretty damn fast for anything that I have tried so far, while replaying plots
> is inherently slow. So any overhead that we add at this point will probably
> not be noticeable.

I think this is the only viable solution for now. I'll work on this in
the next few days.

BTW, Thomas:
When / if time permits, can you download R-2.11-1.tar.gz and browse through

src/library/grDevices/src/devWindows.c

Of particular interest is the string "xd->needsave" (and also the
"displayList" member)

Can you figure out the logic as to when they are setting needsave to
TRUE (~2139)? Or the other way: the logic when they are setting
needsave to FALSE (~1135,1177,1882,2137)

Is there anything helpful for us? This is again not for this release
but for the "bigger picture."

Regards,
-- 
Prasenjit