[ESS] Interactive SAS mode hangs

Erin Jonaitis jon@|t|@ @end|ng |rom |@tr|-w|@org
Tue Sep 1 16:49:26 CEST 2015


Ah, never mind,  I figured out what you meant: that setq command goes in the lisp scratch window, followed by another lisp command in the same window to load ESS manually.

I did as you suggested, and one thing I learned is that the same crashing pattern happens (S mode works, but SAS mode crashes emacs when I try to submit code). That's useful because it tells me it's not due to something else bad in the init file. Unfortunately, I can't seem to find a file containing backtrace information. Does the underlying Linux system need to be set up in a certain way to produce such a file? Should I expect it to in ~/.xsession-errors, or could it have another name or location?

Erin

----- Original Message -----
From: "Erin Jonaitis" <jonaitis using fstrf-wi.org>
To: "ess-help" <ess-help using stat.math.ethz.ch>
Sent: Tuesday, September 1, 2015 8:55:39 AM
Subject: Re: Interactive SAS mode hangs

Hi Rodney,

Thanks for your help. This is a newbie question -- I figured out what you meant with the -q flag, but I'm not sure how to follow your (setq debug-on-error t) suggestion. Is this something I type at the command line while calling emacs, or something I enter in the minibuffer, or something that goes in the init file? (If the latter, I'm confused because it looks like -q tells it not to load the init file; I'm definitely missing something here.)

Erin

----- Original Message -----
From: "Sparapani, Rodney" <rsparapa using mcw.edu>
To: jonaitis using fstrf-wi.org
Cc: "ess-help" <ess-help using stat.math.ethz.ch>
Sent: Tuesday, September 1, 2015 7:48:26 AM
Subject: Re: Interactive SAS mode hangs

> 
> Hi all,
> 
> I'm working on a Linux machine (not sure what distribution) using emacs 23.1-28.el6.x86_64 and ESS
> version 15.03.
> 

Although, we support 23.x, I suspect you might have better luck with 24.3+

> I'm used to working in SAS interactively in ESS (i.e. by sending M-x SAS and then executing code chunks one at
> a time). I'd been working on a much older version of ESS, but my system is in the process of being upgraded and
> I'm testing ESS on the new machine. Right now R seems to work fine, but when I try to execute (much) SAS code
> inside ESS, it crashes emacs badly and can only be killed via the command line. Sometimes the crash
> involves very strange cycling of the cursor image through various possible cursor images (an arrow, two
> balls, arrows specific to window resizing).
> 
> Questions: Any idea what could be the problem? What other information would you need to be able to help
> diagnose it?

Well, obviously, using ESS should not crash emacs.  You could try the usual
diagnostics, i.e. starting emacs with -q, (setq debug-on-error t), then 
loading ESS and see if a *backtrace* is generated.
 
> Details of the crashing:
> - Trying to define a new data step crashes.
> - libname statements do not crash.
> - Trying to read data from a set inside a successfully defined library causes a crash.
> - Using proc contents on such a dataset also crashes.
> - Macro %put statements do not crash.
> 
> Sample code:
> 
> %put I read the first line; *this works -- prints to the log;
> libname temp '/my/path/here'; *this works too -- gives me a "successfully defined" message;
> 
> data temp2; *causes a hard crash with weird cursor behavior;
>     set temp.temp1;
> run;
> 
> data temp3; *causes a hard crash;
>     infile datalines dsd;
>     input id name $;
>     datalines;
>     1,Fred
>     2,Joni
>     3,Cara
>     4,Mick
> ;
> run;
> 
> proc contents data=temp.temp1 order=varnum; *crashes but without the weird cursor stuff;
> run;
> 
> Erin Jonaitis  |  Biostatistician

I just tried this example and it did not crash emacs for me.  Instead of
M-x SAS, you can try M-x ess-sas-interactive; it is not a replacement for
M-x SAS, but it might help you diagnose the problem you are having.




More information about the ESS-help mailing list