[R-SIG-Mac] 'buggy' editor

[Robert Chatfield]

Here are some specific requests (since they were solicited)!

I find the Macintosh R editor rather useful; it lacks certain features
of the emacs environment, but it is more intuitive.

These requests progress from desired bug-fixes (or "feature"  
alternatives) to
more general requests about interaction of the editor and the R  
Console window.
"Delete last output" as in ess would be my greatest desire.

(I am running R 2.7.2 GUI 1.25 (5217) on a Dual-Core Xeon, under  
10.4.11  (and 10.5.x)

(1) Anomalous behavior after locating a cursor for edit.
place the editor cursor at the end of a line, hit "return" and start  
typing.
The editor continues at the end of the line.  To emulate the normal
behavior, hit "return" _twice_    This differs from every other  
editor I know of.

(2) Need to set all fonts to standard.
When copying in code lines from examples, other editors, etc, the  
font of
the original window is retained.  This may be a desirable feature for  
some,
but a "Preference" to adopt the existing font and size should be easy to
implement.  I have no use for alternate fonts.

(3) Line numbers not readable:
On long code scripts, numbers over 999 are not readable (in some cases
 >101 seem a bit hard to read).

(4) "Marks" would be useful
In light of (3), and especially when adding and removing many lines
of code, "marks" a, b, c ... (or whatever) would help a lot.

(5) "Stop on Fail" ... It would be wonderful if the editor could stop
feeding commands after the first error (or, preference-setable, at
first warning).

(6) Evaluate and execute:  It would be wonderful to step through code
easily, line by line.

(7777) "Delete Last Output"
Last but definitely not least (perhaps first!), it would be extremely  
helpful
to remove the last output from the console.  (Becoming frustrated about
details, I just typed
 > names(OH.NO2.byday.gam$smooth)
NULL
 > OH.NO2.byday.gam$smooth

to understand an object (list) produced by mgcv's gam function.
Well, 5500 lines of output later, I regretted my willingness to learn.
Of  course, simple mistakes can also mess up a console log.
Often I keep edited console logs to record free-exploration sessions ...
and free-exploration of alternatives, possibilities, and errors in  
assumptions
is what R is all about, right?

keep up the good work !
regards, Bob Chatfield