Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!samsung!spool.mu.edu!snorkelwacker.mit.edu!bloom-beacon!eru!hagbard!sunic!mcsun!corton!irisa!decouty
From: decouty@irisa.fr (Bertrand Decouty)
Newsgroups: comp.emacs
Subject: Re: uemacs 3.10m/3.11B porting questions
Message-ID: <1991Mar29.092148.6534@irisa.fr>
Date: 29 Mar 91 09:21:48 GMT
References: <1991Mar28.083424.7514@odin.corp.sgi.com>
Sender: news@irisa.fr
Organization: IRISA-INRIA, Campus de Beaulieu, 35042 Rennes Cedex, FRANCE
Lines: 73

In article <1991Mar28.083424.7514@odin.corp.sgi.com> bam@rudedog.asd.sgi.com (Brian McClendon) writes:
| 
| After my silly ftp situation (I was just not patient enough), take any 
| problems I have with a grain of salt.
| 
| I picked up the src for both 3.10m and 3.11B and, after switching back to
| the old (3.10e) makefile.unx, the first problem I had was with ctrl-S
| acting as XOFF instead of a normal ctrl character.  This was fixed by:
| 
| changing:
| 	curterm.c_iflag &= ~(INLCR|ICRNL|IGNCR);
| to:
| 	curterm.c_iflag &= ~(INLCR|ICRNL|IGNCR|IXON|IXOFF);
| 
| in unix.c.

For catching 8-bit characters, you must add | ISTRIP. Now i can type 8-bit
chars, but there is still a bug in search proc which prevent me from
searching these characters. I shall investigate, as i have done it in 3.10e.
BTW, when you do ^S/^R in incremental search, these ^S/^R are put in the
search string!! A big (and old ?) bug.

As we are speaking of 3.11Beta and bugs:
- in eproto.h, mlwrite(va_dcl) must be mlwrite(va_alist) (if you use
varargs)
- in unix.c, 
	return write(0, outbuf, len) != len;
should be (in order for me311 to be called from less pager)
        status = (fwrite( outbuf, 1, len, stdout) != len);
        fflush(stdout);
	return(status)

Also, in input.c, there is still no ~/ and $VARIABLE expansion. Hopefully,
older patch works but it should be included in the final release.

I have compiled me311B on sun3/sun4 with gcc, apart from a lot of warning,
it works (flags SUN and GCC, no BSD)!

| 
| Now that my ctrl-S gets me into incremental search, I find that a
| second ctrl-S gets me into a _new_ incremental search instead of 
| searching for the 2nd occurence of the current search string.  Is this
| one obvious?
It works for me, but i have made this patch:

*** isearch.c.orig	Wed Mar 27 12:27:34 1991
--- isearch.c	Wed Mar 27 12:30:46 1991
***************
*** 195,200 ****
--- 195,201 ----
  	            while (pat[cpos] != 0)
  	                col = echochar((int)pat[cpos++],col);
  
+ 		cmd_buff[--cmd_offset] = '\0';  /* Delete last char (^R/^S)  */
  	        status = scanmore(dir);
  	        continue;
              }
==============================
If anybody have other patches...
Bertrand Decouty

| 
| --
| ----------------------------------------------------------------------------
|  Brian McClendon bam@rudedog.SGI.COM ...!uunet!sgi!rudedog!bam 415-335-1110
| ----------------------------------------------------------------------------


-- 
[ Bertrand DECOUTY             | EMAIL : decouty@irisa.fr, decouty@irisa.uucp  ]
[ IRISA - INRIA (Atelier)      |         {uunet,mcsun,inria}!irisa!decouty     ]
[ Campus de Beaulieu           | PHONE : +33  99 36 20 00                      ]
[ F-35042 Rennes Cedex - FRANCE| FAX : +33  99 38 38 32 | TELEX: 950473 UNIRISA]