Re: nohup & SYSTEM::SET-STREAM-EXTERNAL-FORMAT

[Sam S. <sd...@gn...>]

> * Bruno Haible <oe...@py...t> [2004-09-08 18:29:49 +0200]:
>
>> on solaris 64 bit and on linux amd I get
>> 
>> $ nohup clisp foo.lisp
>> $ cat nohup.out
>
> I don't observe this: I just get
> WARNING: *FOREIGN-ENCODING*: reset to ASCII

I observe it only over ssh and telnet, not on local xterms.

>> 2. why does stream.d:make_terminal_io() create a terminal stream for
>>    input if output goes to a file? (and vice versa)
>
> On most systems (even on Unix, when using the readline library),
> accessing the terminal's input requires special routines; they are
> collected under the hood of the terminal-stream.
>
> In other words, if the user has redirected standard output, he is
> expecting the input to still come from the same source as previously.

1. this issue turned out to be separate and independent from the nohup
   problem: when I apply the appended patch, I get

*** - SYSTEM::SET-STREAM-EXTERNAL-FORMAT on
#<IO TWO-WAY-STREAM #<INPUT BUFFERED FILE-STREAM CHARACTER #P"/dev/fd/0">
  #<OUTPUT BUFFERED FILE-STREAM CHARACTER #P"/dev/fd/1">> is illegal

2. apparently I was not clear enough.
   what I am trying to say is that the only reason to create a
   #<IO TERMINAL-STREAM> (as opposed to #<INPUT BUFFERED FILE-STREAM
   CHARACTER #P"/dev/fd/0">) is to use readline.
   when output is redirected, the value of readline is dubious.
   when input is redirected, the value is readline is nil.
   so why not just create a regular buffered channel stream?
   i/o redirection indicates batch processing where the performance
   benefits of buffering outweigh UI benefits of readline.

>> why is SYSTEM::SET-STREAM-EXTERNAL-FORMAT called?
>
> set_terminalstream_external_format is only called from the
> *terminal-encoding* setter, You appear to set *terminal-encoding*
> after clisp has started up?!  Or what else are you doing?

I am not doing anything with *terminal-encoding*.

here is what I get right now (over ssh):

$ rm -fv nohup.out; nohup clisp -q -norc .clisprc ; cat nohup.out
removed `nohup.out'
nohup: appending output to `nohup.out'

*** - SYSTEM::SET-STREAM-EXTERNAL-FORMAT on
#<IO TWO-WAY-STREAM #<IO TERMINAL-STREAM>
  #<OUTPUT BUFFERED FILE-STREAM CHARACTER #P"/dev/fd/1">> is illegal
$ rm -fv nohup.out; nohup clisp -q -norc -i .clisprc -x '(1+ 2)' ; cat nohup.out
removed `nohup.out'
nohup: appending output to `nohup.out'
;; Loading file .clisprc ...
*** - SYSTEM::SET-STREAM-EXTERNAL-FORMAT on
#<IO TWO-WAY-STREAM #<IO TERMINAL-STREAM>
  #<OUTPUT BUFFERED FILE-STREAM CHARACTER #P"/dev/fd/1">> is illegal
$ rm -fv nohup.out; nohup clisp -q -norc -x '(1+ 2)' ; cat nohup.out
removed `nohup.out'
nohup: appending output to `nohup.out'
3
$

i.e., this appears to be somehow related to LOAD.



-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.honestreporting.com>
I want Tamagochi! -- What for?  Your pet hamster is still alive!



--- stream.d	06 Sep 2004 20:05:29 -0400	1.447
+++ stream.d	07 Sep 2004 12:03:41 -0400	
@@ -14451,13 +14451,9 @@
   var bool stdin_file = regular_handle_p(stdin_handle);
   var bool stdout_file = regular_handle_p(stdout_handle);
   if (stdin_file || stdout_file) {
-    var object stream = stdin_file /* input */
-      ? make_standard_input() : make_terminal_stream();
-    pushSTACK(stream);
-    stream = stdout_file        /* output */
-      ? make_standard_output() : make_terminal_stream();
+    pushSTACK(make_standard_input());
     /* Build a two-way-stream: */
-    return make_twoway_stream(popSTACK(),stream);
+    return make_twoway_stream(popSTACK(),make_standard_output());
   }
   return make_terminal_stream();
 }