emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Don.Bashford]

I like to do my builds of almost anything from source under emacs Shell
mode.  But emacs Shell mode sets the environment variable EMACS to
"t".  This causes the output of autoconf's AC_CHECK_PROG(EMACS, ...)
to set EMACS="t", with the result that in any Makefile generated from
EMACS = @EMACS@ in Makefile.in end up with the makefile variable EMACS
set to "t".

One example where this is a problem is the current (Sept 10) CVS
version of gnuplot.

It's not clear to me who should change: autoconf, emacs, developers
using emacs as a built tool, or users like me.  If it's one of the
latter two, a warning would be nice.  Perhaps the simplest solution
would be for emacs Shell to either not set EMACS or set it to "emacs".

emacs version: 21.3.1
autoconf version: 2.60
m4 version: 1.4.6
system: fedora core 1 gnu/linux


-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
gnuplot-bugs mailing list
gnuplot-bugs@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/gnuplot-bugs

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Paul Eggert]

<Don.Bashford@stjude.org> writes:

> I like to do my builds of almost anything from source under emacs Shell
> mode.  But emacs Shell mode sets the environment variable EMACS to
> "t".  This causes the output of autoconf's AC_CHECK_PROG(EMACS, ...)
> to set EMACS="t", with the result that in any Makefile generated from
> EMACS = @EMACS@ in Makefile.in end up with the makefile variable EMACS
> set to "t".
>
> One example where this is a problem is the current (Sept 10) CVS
> version of gnuplot.
>
> It's not clear to me who should change: autoconf, emacs, developers
> using emacs as a built tool, or users like me.  If it's one of the
> latter two, a warning would be nice.  Perhaps the simplest solution
> would be for emacs Shell to either not set EMACS or set it to "emacs".
>
> emacs version: 21.3.1
> autoconf version: 2.60
> m4 version: 1.4.6
> system: fedora core 1 gnu/linux

Thanks for your bug report.  Other people have run into this problem
as well, but I don't know of any fix for it yet.  I like the simple
solution you suggested, but I think it'd be slightly better if EMACS
is set to Emacs's absolute file name, as that is less likely to cause
confusion later.  Here is a proposed patch to CVS Emacs to implement
this.

2006-09-11  Paul Eggert  <eggert@cs.ucla.edu>

	* make-dist (EMACS): Exit and fail if the EMACS environment
	variable is set to something other than an absolute file name.
	* etc/NEWS: In terminal-oriented subshells, the EMACS environment
	variable now defaults to Emacs's absolute file name, instead of
	to "t".
	* etc/PROBLEMS: Adjust tcsh advice for this.
	* lisp/comint.el (comint-exec-1): Set EMACS to the full name of Emacs,
	not to "t".
	* lisp/progmodes/compile.el (compilation-start): Likewise.
	* man/faq.texi (Escape sequences in shell output): EMACS is now set
	to Emacs's absolute file name, not to "t".
	(^M in the shell buffer): Likewise.
	* man/misc.texi (Interactive Shell): Likewise.

Index: make-dist
===================================================================
RCS file: /cvsroot/emacs/emacs/make-dist,v
retrieving revision 1.221
diff -p -c -b -w -r1.221 make-dist
*** make-dist	5 Jul 2006 06:42:27 -0000	1.221
--- make-dist	11 Sep 2006 17:33:22 -0000
*************** if [ ! -d src -o ! -f src/lisp.h -o ! -d
*** 120,127 ****
  fi
  
  ### Find where to run Emacs.
! ### (We don't accept EMACS=t as an answer, since that probably only means
! ### that the shell is running in an Emacs window.)
  if [ $update = yes ];
  then
    unset EMACS_UNIBYTE
--- 120,126 ----
  fi
  
  ### Find where to run Emacs.
! ### (Accept only absolute file names.)
  if [ $update = yes ];
  then
    unset EMACS_UNIBYTE
*************** then
*** 129,139 ****
    then
      EMACS=`pwd`/src/emacs
    else
!     if [ "x$EMACS" = "x" -o "x$EMACS" = "xt" ];
!     then
!       echo You must specify the EMACS environment variable 2>&1
        exit 1
!     fi
    fi
  fi
  
--- 128,142 ----
    then
      EMACS=`pwd`/src/emacs
    else
!     case $EMACS in
!       /*) ;;
!       *)
! 	if [ ! -f "$EMACS" ]; then
! 	  echo "$0: You must specify the EMACS environment variable " \
! 	       "to an absolute file name." 2>&1
  	  exit 1
! 	fi;;
!     esac
    fi
  fi
  
Index: etc/NEWS
===================================================================
RCS file: /cvsroot/emacs/emacs/etc/NEWS,v
retrieving revision 1.1393
diff -p -c -b -w -r1.1393 NEWS
*** etc/NEWS	10 Sep 2006 13:48:42 -0000	1.1393
--- etc/NEWS	11 Sep 2006 17:33:22 -0000
*************** otherwise behaves quite similarly to the
*** 1434,1439 ****
--- 1434,1443 ----
  `comint-use-prompt-regexp'.  The old name has been kept as an alias,
  but declared obsolete.
  
+ +++
+ *** The EMACS environment variable now defaults to Emacs's absolute
+ file name, instead of to "t".
+ 
  ** M-x Compile changes:
  
  ---
*************** it doesn't scroll the compilation output
*** 1490,1495 ****
--- 1494,1503 ----
  no arrow is displayed and a value of nil means display the message at the top
  of the window.
  
+ +++
+ *** The EMACS environment variable now defaults to Emacs's absolute
+ file name, instead of to "t".
+ 
  ** Occur mode changes:
  
  +++
Index: etc/PROBLEMS
===================================================================
RCS file: /cvsroot/emacs/emacs/etc/PROBLEMS,v
retrieving revision 1.196
diff -p -c -b -w -r1.196 PROBLEMS
*** etc/PROBLEMS	10 Sep 2006 12:28:28 -0000	1.196
--- etc/PROBLEMS	11 Sep 2006 17:33:23 -0000
*************** on the flag to output ^M at the end of e
*** 540,546 ****
  problem by adding this to your .cshrc file:
  
      if ($?EMACS) then
!         if ($EMACS == "t") then
              unset edit
              stty  -icrnl -onlcr -echo susp ^Z
          endif
--- 540,546 ----
  problem by adding this to your .cshrc file:
  
      if ($?EMACS) then
!         if ("$EMACS" =~ /*) then
              unset edit
              stty  -icrnl -onlcr -echo susp ^Z
          endif
Index: lisp/comint.el
===================================================================
RCS file: /cvsroot/emacs/emacs/lisp/comint.el,v
retrieving revision 1.346
diff -p -c -b -w -r1.346 comint.el
*** lisp/comint.el	2 Jun 2006 02:04:15 -0000	1.346
--- lisp/comint.el	11 Sep 2006 17:33:24 -0000
*************** buffer.  The hook `comint-exec-hook' is 
*** 765,771 ****
  		    (format "COLUMNS=%d" (window-width)))
  	    (list "TERM=emacs"
  		  (format "TERMCAP=emacs:co#%d:tc=unknown:" (window-width))))
! 	  (if (getenv "EMACS") nil (list "EMACS=t"))
  	  process-environment))
  	(default-directory
  	  (if (file-accessible-directory-p default-directory)
--- 765,772 ----
  		    (format "COLUMNS=%d" (window-width)))
  	    (list "TERM=emacs"
  		  (format "TERMCAP=emacs:co#%d:tc=unknown:" (window-width))))
! 	  (unless (getenv "EMACS")
! 	    (list (concat "EMACS=" invocation-directory invocation-name)))
  	  process-environment))
  	(default-directory
  	  (if (file-accessible-directory-p default-directory)
Index: lisp/progmodes/compile.el
===================================================================
RCS file: /cvsroot/emacs/emacs/lisp/progmodes/compile.el,v
retrieving revision 1.405
diff -p -c -b -w -r1.405 compile.el
*** lisp/progmodes/compile.el	5 Sep 2006 14:53:24 -0000	1.405
--- lisp/progmodes/compile.el	11 Sep 2006 17:33:24 -0000
*************** Returns the compilation buffer created."
*** 1068,1074 ****
  			      (window-width))))
  	      ;; Set the EMACS variable, but
  	      ;; don't override users' setting of $EMACS.
! 	      (unless (getenv "EMACS") '("EMACS=t"))
  	      (copy-sequence process-environment))))
  	(set (make-local-variable 'compilation-arguments)
  	     (list command mode name-function highlight-regexp))
--- 1068,1075 ----
  			      (window-width))))
  	      ;; Set the EMACS variable, but
  	      ;; don't override users' setting of $EMACS.
! 	      (unless (getenv "EMACS")
! 		(list (concat "EMACS=" invocation-directory invocation-name)))
  	      (copy-sequence process-environment))))
  	(set (make-local-variable 'compilation-arguments)
  	     (list command mode name-function highlight-regexp))
Index: man/faq.texi
===================================================================
RCS file: /cvsroot/emacs/emacs/man/faq.texi,v
retrieving revision 1.97
diff -p -c -r1.97 faq.texi
*** man/faq.texi	12 Aug 2006 14:13:31 -0000	1.97
--- man/faq.texi	11 Sep 2006 17:40:32 -0000
*************** shell init file.  You have two alternati
*** 2683,2689 ****
  @item
  Make the alias conditioned on the @code{EMACS} variable in the
  environment.  When Emacs runs a subsidiary shell, it exports the
! @code{EMACS} variable with the value @code{t} to that shell.  You can
  unalias @code{ls} when that happens, thus limiting the alias to your
  interactive sessions.
  
--- 2683,2690 ----
  @item
  Make the alias conditioned on the @code{EMACS} variable in the
  environment.  When Emacs runs a subsidiary shell, it exports the
! @code{EMACS} variable to that shell, with value equal to the absolute
! file name of Emacs.  You can
  unalias @code{ls} when that happens, thus limiting the alias to your
  interactive sessions.
  
*************** file:
*** 2753,2759 ****
  
  @example
  if ($?EMACS) then
!     if ("$EMACS" == t) then
          if ($?tcsh) unset edit
          stty nl
      endif
--- 2754,2760 ----
  
  @example
  if ($?EMACS) then
!     if ("$EMACS" =~ /*) then
          if ($?tcsh) unset edit
          stty nl
      endif
Index: man/misc.texi
===================================================================
RCS file: /cvsroot/emacs/emacs/man/misc.texi,v
retrieving revision 1.92
diff -p -c -b -w -r1.92 misc.texi
*** man/misc.texi	16 Aug 2006 05:14:31 -0000	1.92
--- man/misc.texi	11 Sep 2006 17:33:25 -0000
*************** Coding}.
*** 488,494 ****
  
  @cindex @env{EMACS} environment variable
    Unless the environment variable @env{EMACS} is already defined,
! Emacs defines it in the subshell, with value @code{t}.  A shell script
  can check this variable to determine whether it has been run from an
  Emacs subshell.
  
--- 488,495 ----
  
  @cindex @env{EMACS} environment variable
    Unless the environment variable @env{EMACS} is already defined,
! Emacs defines it in the subshell, with value equal to Emacs's absolute
! file name.  A shell script
  can check this variable to determine whether it has been run from an
  Emacs subshell.
  

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Richard Stallman]

Your fix is very clever, and I think it is a feature that running configure
scripts in a shell under Emacs uses the same Emacs.  So please install
your change, and we can consider the problem fully solved.

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Richard Stallman]

    I like to do my builds of almost anything from source under emacs Shell
    mode.  But emacs Shell mode sets the environment variable EMACS to
    "t".

      This causes the output of autoconf's AC_CHECK_PROG(EMACS, ...)
    to set EMACS="t", with the result that in any Makefile generated from
    EMACS = @EMACS@ in Makefile.in end up with the makefile variable EMACS
    set to "t".

    One example where this is a problem is the current (Sept 10) CVS
    version of gnuplot.

For a short-term correction, I think the easiest way is to change
Autoconf to ignore a value of `t' for this environment variable.
Autoconf maintainers, what do you think?

For longer term, perhaps the best thing is to change Emacs to use a
different environment variable.  Alas, I don't remember what software
checks that environment variable, or the motive for binding it.

Is it Bash?  I have cc'd the Bash maintainer so he can answer.

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Chet Ramey]

> For a short-term correction, I think the easiest way is to change
> Autoconf to ignore a value of `t' for this environment variable.
> Autoconf maintainers, what do you think?
> 
> For longer term, perhaps the best thing is to change Emacs to use a
> different environment variable.  Alas, I don't remember what software
> checks that environment variable, or the motive for binding it.
> 
> Is it Bash?  I have cc'd the Bash maintainer so he can answer.

Bash does inspect the EMACS environment variable as part of checking
whether or not to turn off line editing even when the shell is
interactive.  It also checks for TERM==emacs.  If the shell discovers
it's running in an emacs buffer, it turns off readline. 

Chet

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
				Live Strong.
Chet Ramey, ITS, CWRU    chet@case.edu    http://tiswww.tis.case.edu/~chet/

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Don.Bashford]

>>>>> "Chet" == Chet Ramey <chet.ramey@case.edu> writes:

    Chet> Bash does inspect the EMACS environment variable as part of
    Chet> checking whether or not to turn off line editing even when
    Chet> the shell is interactive.  It also checks for TERM==emacs.
    Chet> If the shell discovers it's running in an emacs buffer, it
    Chet> turns off readline.

Maybe emacs should be in charge of turning readline off directly
instead expecting the shell to check envars to deal with special
cases.  For example, emacs could send the shell something like "shopt
-u -o emacs; shopt -u -o vi".  Then emacs could leave EMACS alone.

It might also be a good idea for AC_CHECK_COMMAND to at least give a
warning if shell variables have crazy-seeming values like "t", "0" or
"1".  Better yet, when envars try to override command names it could
check whether the the command really exists.  Or,

  PROG=enval AC_CHECK_COMMAND(PROG, foo bar ...) 
         --> AC_CHECK_COMMAND(PROG, enval foo bar ...) 

-Don

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Don.Bashford]

>>>>> "rms" == Richard Stallman <rms@gnu.org> writes:


    rms> For a short-term correction, I think the easiest way is to
    rms> change Autoconf to ignore a value of `t' for this environment
    rms> variable.  Autoconf maintainers, what do you think?

    rms> For longer term, perhaps the best thing is to change Emacs to
    rms> use a different environment variable.  Alas, I don't remember
    rms> what software checks that environment variable, or the motive
    rms> for binding it.

    rms> Is it Bash?  I have cc'd the Bash maintainer so he can
    rms> answer.

Yes, I think that's it.  From the bash info pages:

`EMACS'
     If Bash finds this variable in the environment when the shell
     starts with value `t', it assumes that the shell is running in an
     emacs shell buffer and disables line editing.

-Don

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Miles Bader]

<Don.Bashford@stjude.org> writes:
>For example, emacs could send the shell something like "shopt
> -u -o emacs; shopt -u -o vi".  Then emacs could leave EMACS alone.

Stuffing input to a subshell like that is in general not something you
can do reliably.

-miles

-- 
We are all lying in the gutter, but some of us are looking at the stars.
-Oscar Wilde

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Chet Ramey]

> <Don.Bashford@stjude.org> writes:
> >For example, emacs could send the shell something like "shopt
> > -u -o emacs; shopt -u -o vi".  Then emacs could leave EMACS alone.
> 
> Stuffing input to a subshell like that is in general not something you
> can do reliably.

Yeah, that's dicy, but you don't have to do it.  Just invoke the shell
with the `--noediting' option.

Chet

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
				Live Strong.
Chet Ramey, ITS, CWRU    chet@case.edu    http://tiswww.tis.case.edu/~chet/

Re: emacs Shell mode vs AC_CHECK_PROG(EMACS, ...)

[Kim F. Storm]

Chet Ramey <chet.ramey@case.edu> writes:

>> <Don.Bashford@stjude.org> writes:
>> >For example, emacs could send the shell something like "shopt
>> > -u -o emacs; shopt -u -o vi".  Then emacs could leave EMACS alone.
>> 
>> Stuffing input to a subshell like that is in general not something you
>> can do reliably.
>
> Yeah, that's dicy, but you don't have to do it.  Just invoke the shell
> with the `--noediting' option.

CVS emacs already uses --noediting -i for bash (see variable explicit-bash-args).

-- 
Kim F. Storm  http://www.cua.dk