[ESS] iESS: point jumps to EOB on process output

[Sam Steingold]

I would like to reiterate that this behavior - jumping to eob all the
time - is a major inconvenience for me.
Is there really any reason to deviate from the default comint behavior?

> * Sam Steingold <fqf at tah.bet> [2012-09-05 09:46:47 -0400]:
>
>> * Vitalie Spinu <fcvahivg at tznvy.pbz> [2012-09-05 10:20:04 +0200]:
>>
>>   >> Sam Steingold <sds at gnu.org>
>>   >> on Tue, 4 Sep 2012 17:30:21 -0400 wrote:
>>
>>   > When I have an R function executing, whenever it prints something, the
>>   > point jumps to the end of the ESS buffer.
>>   > This makes is very difficult to get the previously printed data (since
>>   > the point jumps to the end of buffer all the time).
>>   > It would be nice if the point jumped to the EOB only when the function
>>   > finished (i.e., when the prompt is printed) or not at all.
>>
>> First check the values of 
>>
>>   comint-scroll-to-bottom-on-output
>>   comint-scroll-show-maximum-output
>>
>> if they are t this is comint thing. 
>
> comint-scroll-to-bottom-on-output is a variable defined in `comint.el'.
> Its value is nil
>
>   This variable is an alias for `comint-move-point-for-output'.
>
> comint-scroll-show-maximum-output ==> t
> but it only affects the behavior _after_ the command completes.
>
>> Second, ESS has two evaluation mechanisms visible and not, depending
>> on the value of ess-eval-visibly-p (defaults to t). If evaluation is
>> visible, that means you want to see the input text also in your output
>> buffer.
>
> what is "output buffer"?
>
> also, ess-eval-visibly-p does not comply with the emacs coding standards
> (variables should not have "-p", that's for functions) :-)
>
> also, I do want my commands and their output to show up in *R*,
> but I do NOT want that to affect the point in *R*.
>
>> I see, that in this case, ESS ignores
>> comint-scroll-to-bottom-on-output.
>>
>> We can in principle make ess-eval-linewise respect
>> comint-scroll-to-bottom-on-output. What do others think?
>
> In general, I think the interactive buffer behavior should be the
> vanilla comint unless there is a clear overriding need otherwise.
> The fact that I face different behavior in shell and R is quite
> inconvenient; it means that I have to customize twice essentially the
> same modes.
>
>> Third, what do you mean that it makes "very difficult to get
>> previously printed data"?
>
> Imagine I did something in R, so that the *R* buffer contains, say, 40
> lines which I want to examine (e.g., copy into a different buffer).
> Now, R is running a function which prints a line every 5 seconds.
> I scroll up to find my 40 important lines, but every 5 seconds my point
> jumps to the end of the buffer, so that I have no time to select the 40
> lines and copy them out.

-- 
Sam Steingold (http://sds.podval.org/) on Ubuntu 12.04 (precise) X 11.0.11103000
http://www.childpsy.net/ http://openvotingconsortium.org http://camera.org
http://www.PetitionOnline.com/tap12009/ http://honestreporting.com
If you try to fail, and succeed, which have you done?