[slime-devel] Re: REPL results and output

Madhu enometh at meer.net
Fri Feb 2 04:28:38 UTC 2007


Thanks for checking. I looked at the code for this and noticed that
this bug occurs only when :*USE-DEDICATED-OUTPUT-STREAM* is set to
T. I believe this bug is related to the following Changelog entries


* 2006-12-31  Matthias Koeppe  <mkoeppe at mail.math.uni-magdeburg.de>
|	* swank.lisp (connection): New slot repl-results (a stream).
|	(make-output-function-for-target): New.
|	(open-streams): Use it here to also create a stream for REPL results.

* 2006-12-29  Matthias Koeppe  <mkoeppe at mail.math.uni-magdeburg.de>
|	Simplify the REPL-results protocol.  The results are now printed
|	using special :WRITE-STRING events from the Lisp side. 

	
Now there are two issues I would like some elucidation on:

2. Doesnt the new design contradict using a dedicated output stream?
   The repl results stream is not the dedicated output stream (opened
   in OPEN-STREAMS). Assuming for a moment that it were, how can the
   new protocol for printing repl results talk to that?

1. What is the status of SWANK:*USE-DEDICATED-OUTPUT-STREAM* ? I
   remember there was a move to eliminate it. Am I the only person
   using it? (not unusual!)

Actually I have not understood the presentation protocol and barely
grasp the connections stuff.
--
Madhu


* Jeffrey Cunningham <20070201155908.GE8077 at achilles.olympus.net> :
| On Tue Jan 30, 2007 at 03:04:01PM +0530, Madhu wrote:
|> The  printing of repl results seems to have changed recently:
|> 
|> CL-USER> (format *standard-output* "foo")
|> NIL
|> CL-USER> foo
|> 
|> CL-USER> (format *standard-output* "foo~%")
|> NIL
|> CL-USER> foo
|> 
|> i.e. The result is printed and the prompt is printed before the
|> command's output.
|> 
|> Do other people get this same behaviour?
|> 
|> Is this backend/interaction mode specific? (I'm using cmucl).
|> --
|
|
| I'm not getting it with SBCL 1.0.1 + SLIME 2007-01-24
|
| CL-USER>  (format *standard-output* "foo")
| foo
| NIL
|



More information about the slime-devel mailing list