bug#20202: EMACS=t Joy and Happiness

bug#20202: EMACS=t Joy and Happiness

[Stefan Monnier]

> I worry that "Bash 4.4 or later is common" is rather indeterminate; my

At least Debian's stable (stretch) is on bash-4.4 (but still at 4.3 in
oldstable (jessie)).

> desktop is now 4.4.19. but another machine I am using has bash
> 4.2.46. Given that Emacs-26 is now destined to come out with
> EMACS=26.1.0, I would like to propose that we now remove this on master,
> i.e. make it happen for Emacs-27 which should be in about 2020.

Sounds good to me: by the time Emacs-27 comes out, Debian's oldstable
will be stretch (hence at 4.4).


        Stefan

bug#20484: EMACS=t Joy and Happiness

[Paul Eggert]

Bash 4.4 came out only in September 2016, and it's too soon to assume 
it. For example, my students regularly use Emacs on a server running 
RHEL 7, which ships Bash 4.2.46, and as I understand it, this is the 
most recent RHEL available.

How about this idea: At Emacs build time, we check whether Bash is 4.4 
or later, and if so we use a term.el that assumes bash 4.4 or later. 
Otherwise, we use a term.el that interrogates the shell dynamically for 
whether it is shell and if so what its version number is, the first time 
that Emacs runs a process under a shell.

bug#20202: EMACS=t Joy and Happiness

[Paul Eggert]

Bash 4.4 came out only in September 2016, and it's too soon to assume 
it. For example, my students regularly use Emacs on a server running 
RHEL 7, which ships Bash 4.2.46, and as I understand it, this is the 
most recent RHEL available.

How about this idea: At Emacs build time, we check whether Bash is 4.4 
or later, and if so we use a term.el that assumes bash 4.4 or later. 
Otherwise, we use a term.el that interrogates the shell dynamically for 
whether it is shell and if so what its version number is, the first time 
that Emacs runs a process under a shell.

bug#20202: EMACS=t Joy and Happiness

[Van L]

> Paul Eggert writes:
> 
> this is the most recent RHEL available

You could have your sclerotic institution’s IT manager provide Enterprise RHEL for students studying topics as old as the hills and mountains (and they won’t change) and request them to “think" about the latest unstable Fedora or Debian or Ubuntu releases for research oriented experimental AI & CS “work” which are hot career topics and ought not to be held back by bureaucratically convenient RHEL IT hell. Ask them for what Tesla or Uber or Facebook or Google use to keep up with “innovation” before it is "too late" on RHEL.

bug#20484: EMACS=t Joy and Happiness

[Ricardo Wurmus]

Van L <van@scratch.space> writes:

>> Paul Eggert writes:
>>
>> this is the most recent RHEL available
>
> You could have your sclerotic institution’s IT manager provide
> Enterprise RHEL for students studying topics as old as the hills and
> mountains (and they won’t change) and request them to “think" about
> the latest unstable Fedora or Debian or Ubuntu releases for research
> oriented experimental AI & CS “work” which are hot career topics and
> ought not to be held back by bureaucratically convenient RHEL IT
> hell. Ask them for what Tesla or Uber or Facebook or Google use to
> keep up with “innovation” before it is "too late" on RHEL.

This is not constructive.

There are even RHEL 6 deployments out there, because it’s still
supported until 2020.  There are people who pay for the privilege of not
having to upgrade.

--
Ricardo

bug#20202: EMACS=t Joy and Happiness

[Ricardo Wurmus]

Van L <van@scratch.space> writes:

>> Paul Eggert writes:
>>
>> this is the most recent RHEL available
>
> You could have your sclerotic institution’s IT manager provide
> Enterprise RHEL for students studying topics as old as the hills and
> mountains (and they won’t change) and request them to “think" about
> the latest unstable Fedora or Debian or Ubuntu releases for research
> oriented experimental AI & CS “work” which are hot career topics and
> ought not to be held back by bureaucratically convenient RHEL IT
> hell. Ask them for what Tesla or Uber or Facebook or Google use to
> keep up with “innovation” before it is "too late" on RHEL.

This is not constructive.

There are even RHEL 6 deployments out there, because it’s still
supported until 2020.  There are people who pay for the privilege of not
having to upgrade.

--
Ricardo

bug#20202: EMACS=t Joy and Happiness

[Paul Eggert]

On 05/24/2018 07:25 PM, Van L wrote:
> Ask them for what Tesla or Uber or Facebook or Google use


I just checked the first company mentioned on your list (Tesla) and it 
is posting want-ads that ask for Red Hat expertise. Unfortunately I 
can't disclose what I privately know about Tesla, but suffice it to say 
that like most other successful companies they use a wide variety of 
systems in their work, and not all of these systems are the very latest 
version of everything.


> I can understand the excuse for long-term-support system contracts but not for the poster’s student environment.
>
> You refer to lazy privileged people


I'm afraid there's a misunderstanding here. At UCLA we have a wide 
variety of systems running many different OSes. Servers at UCLA run many 
operating systems; it's quite a gamut, as we have a lot of departments 
and research groups, most of which focus on areas other than computer 
science and who have a lot of software of varying quality that they 
don't necessarily have time or budget to upgrade or maintain 
extensively. The courses I teach are assigned to a batch of RHEL 7 
servers that support many different courses, not just mine. I do not 
specify or maintain the server OSes; that's a responsibility of the 
operations staff of the School of Engineering, and they're busy people 
who do not report to me.

To help keep my course material up-to-date I install copies of the 
latest released versions of Emacs (currently 25.3), Bash (currently 
4.4.19), and many other free software programs. (I don't overwrite 
/usr/bin of course; I install into a separate directory that students 
put into their PATH settings.) So it'll be little trouble to me or my 
students if Emacs requires Bash 4.4 or later. When I mentioned RHEL 7, I 
wasn't simply talking about my personal situation; I was observing that 
requiring Bash 4.4 would be a hassle for people installing newer Emacs 
versions onto Red Hat servers, as it would require these people to also 
install newer Bash versions. If it's important for Emacs to require Bash 
4.4 of course Emacs can do so; still, I expect that the overall hassle 
to Emacs users of requiring Bash 4.4 still outweighs the relatively 
minor technical benefit in question. Sorry, but that's how things often 
work in production software.

With all this in mind you might want to rethink your comments that among 
other things seem to imply that the staff at UCLA consists of "lazy 
privileged people".

bug#20484: EMACS=t Joy and Happiness

[Paul Eggert]

On 05/24/2018 07:25 PM, Van L wrote:
> Ask them for what Tesla or Uber or Facebook or Google use


I just checked the first company mentioned on your list (Tesla) and it 
is posting want-ads that ask for Red Hat expertise. Unfortunately I 
can't disclose what I privately know about Tesla, but suffice it to say 
that like most other successful companies they use a wide variety of 
systems in their work, and not all of these systems are the very latest 
version of everything.


> I can understand the excuse for long-term-support system contracts but not for the poster’s student environment.
>
> You refer to lazy privileged people


I'm afraid there's a misunderstanding here. At UCLA we have a wide 
variety of systems running many different OSes. Servers at UCLA run many 
operating systems; it's quite a gamut, as we have a lot of departments 
and research groups, most of which focus on areas other than computer 
science and who have a lot of software of varying quality that they 
don't necessarily have time or budget to upgrade or maintain 
extensively. The courses I teach are assigned to a batch of RHEL 7 
servers that support many different courses, not just mine. I do not 
specify or maintain the server OSes; that's a responsibility of the 
operations staff of the School of Engineering, and they're busy people 
who do not report to me.

To help keep my course material up-to-date I install copies of the 
latest released versions of Emacs (currently 25.3), Bash (currently 
4.4.19), and many other free software programs. (I don't overwrite 
/usr/bin of course; I install into a separate directory that students 
put into their PATH settings.) So it'll be little trouble to me or my 
students if Emacs requires Bash 4.4 or later. When I mentioned RHEL 7, I 
wasn't simply talking about my personal situation; I was observing that 
requiring Bash 4.4 would be a hassle for people installing newer Emacs 
versions onto Red Hat servers, as it would require these people to also 
install newer Bash versions. If it's important for Emacs to require Bash 
4.4 of course Emacs can do so; still, I expect that the overall hassle 
to Emacs users of requiring Bash 4.4 still outweighs the relatively 
minor technical benefit in question. Sorry, but that's how things often 
work in production software.

With all this in mind you might want to rethink your comments that among 
other things seem to imply that the staff at UCLA consists of "lazy 
privileged people".

bug#20202: EMACS=t Joy and Happiness

[Eli Zaretskii]

> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Thu, 24 May 2018 15:03:56 -0700
> Cc: 20202@debbugs.gnu.org, Stefan Monnier <monnier@iro.umontreal.ca>,
> 	20484@debbugs.gnu.org
> 
> Bash 4.4 came out only in September 2016, and it's too soon to assume 
> it.

I agree, 1.5 years is too soon.  I think we need 5 years at least.

> How about this idea: At Emacs build time, we check whether Bash is 4.4 
> or later, and if so we use a term.el that assumes bash 4.4 or later. 
> Otherwise, we use a term.el that interrogates the shell dynamically for 
> whether it is shell and if so what its version number is, the first time 
> that Emacs runs a process under a shell.

The first part will only work if Emacs is used on the same system
where it is built, right?  Is that a good assumption, given all the
binary distributions?

bug#20484: EMACS=t Joy and Happiness

[Eli Zaretskii]

> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Thu, 24 May 2018 15:03:56 -0700
> Cc: 20202@debbugs.gnu.org, Stefan Monnier <monnier@iro.umontreal.ca>,
> 	20484@debbugs.gnu.org
> 
> Bash 4.4 came out only in September 2016, and it's too soon to assume 
> it.

I agree, 1.5 years is too soon.  I think we need 5 years at least.

> How about this idea: At Emacs build time, we check whether Bash is 4.4 
> or later, and if so we use a term.el that assumes bash 4.4 or later. 
> Otherwise, we use a term.el that interrogates the shell dynamically for 
> whether it is shell and if so what its version number is, the first time 
> that Emacs runs a process under a shell.

The first part will only work if Emacs is used on the same system
where it is built, right?  Is that a good assumption, given all the
binary distributions?

bug#20484: EMACS=t Joy and Happiness

[Van L]

> Eli Zaretskii writes:
> 
>> Bash 4.4 came out only in September 2016, and it's too soon to assume 
>> it.
> 
> I agree, 1.5 years is too soon.  I think we need 5 years at least.

How about applying the 5-multiplier to 9-month cycles?

bug#20202: EMACS=t Joy and Happiness

[Van L]

> Eli Zaretskii writes:
> 
>> Bash 4.4 came out only in September 2016, and it's too soon to assume 
>> it.
> 
> I agree, 1.5 years is too soon.  I think we need 5 years at least.

How about applying the 5-multiplier to 9-month cycles?

bug#20202: EMACS=t Joy and Happiness

[Eli Zaretskii]

> From: Van L <van@scratch.space>
> Date: Fri, 25 May 2018 17:28:31 +1000
> Cc: Paul Eggert <eggert@cs.ucla.edu>, emacs-devel@gnu.org,
> 	20202@debbugs.gnu.org, 20484@debbugs.gnu.org,
> 	monnier@iro.umontreal.ca, phillip.lord@russet.org.uk
> 
> > I agree, 1.5 years is too soon.  I think we need 5 years at least.
> 
> How about applying the 5-multiplier to 9-month cycles?

I think you are confusing this with an entirely different domain.

bug#20484: EMACS=t Joy and Happiness

[Eli Zaretskii]

> From: Van L <van@scratch.space>
> Date: Fri, 25 May 2018 17:28:31 +1000
> Cc: Paul Eggert <eggert@cs.ucla.edu>, emacs-devel@gnu.org,
> 	20202@debbugs.gnu.org, 20484@debbugs.gnu.org,
> 	monnier@iro.umontreal.ca, phillip.lord@russet.org.uk
> 
> > I agree, 1.5 years is too soon.  I think we need 5 years at least.
> 
> How about applying the 5-multiplier to 9-month cycles?

I think you are confusing this with an entirely different domain.

bug#20484: EMACS=t Joy and Happiness

[Phillip Lord]

Paul Eggert <eggert@cs.ucla.edu> writes:

> Bash 4.4 came out only in September 2016, and it's too soon to assume
> it. For example, my students regularly use Emacs on a server running
> RHEL 7, which ships Bash 4.2.46, and as I understand it, this is the
> most recent RHEL available.


Well, Emacs-27 will come out in 2020, so that's 4 years lead time. And
the most recent of RHEL also ships with Emacs-24. So, by the time RHEL
ships Emacs-27, good chance it will ship Bash 4.4.


> How about this idea: At Emacs build time, we check whether Bash is 4.4
> or later, and if so we use a term.el that assumes bash 4.4 or
> later. Otherwise, we use a term.el that interrogates the shell
> dynamically for whether it is shell and if so what its version number
> is, the first time that Emacs runs a process under a shell.


Well, the latter half we could do (I don't think the build idea makes
sense). But I was hoping to make life simpler, not more complex.

Phil

bug#20484: EMACS=t Joy and Happiness

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 332 bytes --]

On 05/24/2018 01:46 PM, Phillip Lord wrote:
> I would like to propose that we now remove this on master,
>

Given the compatibility concerns expressed, how about the attached patch 
instead? The idea is to remove the need to setenv EMACS with newer 
Bashes, while still setting EMACS for older Bashes, for backward 
compatibility.


[-- Attachment #2: 0001-Don-t-set-EMACS-t-if-Bash-is-4.4-or-newer.txt --]
[-- Type: text/plain, Size: 3015 bytes --]

From f7efdc4d21c22a81398b06cdc58144ec2f9c7697 Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Fri, 25 May 2018 13:25:51 -0700
Subject: [PROPOSED] =?UTF-8?q?Don=E2=80=99t=20set=20EMACS=3Dt=20if=20Bash?=
 =?UTF-8?q?=20is=204.4=20or=20newer?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

* lisp/term.el (term--bash-needs-EMACS-status): New var.
(term--bash-needs-EMACSp): New function.
(term-exec-1): Use it instead of always setting EMACS.
---
 lisp/term.el | 31 +++++++++++++++++++++++++------
 1 file changed, 25 insertions(+), 6 deletions(-)

diff --git a/lisp/term.el b/lisp/term.el
index 017b0221ec..fa43774ae8 100644
--- a/lisp/term.el
+++ b/lisp/term.el
@@ -1489,6 +1489,26 @@ term-termcap-format
   ;; don't define :te=\\E[2J\\E[?47l\\E8:ti=\\E7\\E[?47h\
   "Termcap capabilities supported.")
 
+;; This is for backwards compatibility with Bash 4.3 and earlier.
+;; Remove this hack once Bash 4.4-or-later is reasonably universal, because
+;; it slows down execution slightly, just before the first subshell.
+(defvar term--bash-needs-EMACS-status nil
+  "43 if Bash is so old that it needs EMACS set.
+Some other integer if Bash is new or not in use.
+Nil if unknown.")
+(defun term--bash-needs-EMACSp ()
+  "t if Bash is old, nil if it is new or not in use."
+  (unless term--bash-needs-EMACS-status
+    (let ((process-environment (copy-sequence process-environment)))
+      (setenv "BASH_ENV")
+      (setq term--bash-needs-EMACS-status
+            (condition-case nil
+                (call-process
+                 "bash" nil nil nil "-c"
+                 "case $BASH_VERSION in [0123].*|4.[0123].*) exit 43;; esac")
+              (error 0)))))
+  (eq 43 term--bash-needs-EMACS-status))
+
 ;; This auxiliary function cranks up the process for term-exec in
 ;; the appropriate environment.
 
@@ -1506,12 +1526,6 @@ term-exec-1
 	   (format term-termcap-format "TERMCAP="
 		   term-term-name term-height term-width)
 
-	   ;; This is for backwards compatibility with Bash 4.3 and earlier.
-	   ;; Remove this hack once Bash 4.4-or-later is common, because
-	   ;; it breaks './configure' of some packages that expect it to
-	   ;; say where to find EMACS.
-	   (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
-
 	   (format "INSIDE_EMACS=%s,term:%s" emacs-version term-protocol-version)
 	   (format "LINES=%d" term-height)
 	   (format "COLUMNS=%d" term-width))
@@ -1523,6 +1537,11 @@ term-exec-1
 	;; escape codes, so we need to see the raw output.  We will have to
 	;; do the decoding by hand on the parts that are made of chars.
 	(coding-system-for-read 'binary))
+    (when (term--bash-needs-EMACSp)
+      (setq process-environment
+            (cons
+             (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
+             process-environment)))
     (apply 'start-process name buffer
 	   "/bin/sh" "-c"
 	   (format "stty -nl echo rows %d columns %d sane 2>/dev/null;\
-- 
2.17.0

bug#20202: EMACS=t Joy and Happiness

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 332 bytes --]

On 05/24/2018 01:46 PM, Phillip Lord wrote:
> I would like to propose that we now remove this on master,
>

Given the compatibility concerns expressed, how about the attached patch 
instead? The idea is to remove the need to setenv EMACS with newer 
Bashes, while still setting EMACS for older Bashes, for backward 
compatibility.


[-- Attachment #2: 0001-Don-t-set-EMACS-t-if-Bash-is-4.4-or-newer.txt --]
[-- Type: text/plain, Size: 3015 bytes --]

From f7efdc4d21c22a81398b06cdc58144ec2f9c7697 Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Fri, 25 May 2018 13:25:51 -0700
Subject: [PROPOSED] =?UTF-8?q?Don=E2=80=99t=20set=20EMACS=3Dt=20if=20Bash?=
 =?UTF-8?q?=20is=204.4=20or=20newer?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

* lisp/term.el (term--bash-needs-EMACS-status): New var.
(term--bash-needs-EMACSp): New function.
(term-exec-1): Use it instead of always setting EMACS.
---
 lisp/term.el | 31 +++++++++++++++++++++++++------
 1 file changed, 25 insertions(+), 6 deletions(-)

diff --git a/lisp/term.el b/lisp/term.el
index 017b0221ec..fa43774ae8 100644
--- a/lisp/term.el
+++ b/lisp/term.el
@@ -1489,6 +1489,26 @@ term-termcap-format
   ;; don't define :te=\\E[2J\\E[?47l\\E8:ti=\\E7\\E[?47h\
   "Termcap capabilities supported.")
 
+;; This is for backwards compatibility with Bash 4.3 and earlier.
+;; Remove this hack once Bash 4.4-or-later is reasonably universal, because
+;; it slows down execution slightly, just before the first subshell.
+(defvar term--bash-needs-EMACS-status nil
+  "43 if Bash is so old that it needs EMACS set.
+Some other integer if Bash is new or not in use.
+Nil if unknown.")
+(defun term--bash-needs-EMACSp ()
+  "t if Bash is old, nil if it is new or not in use."
+  (unless term--bash-needs-EMACS-status
+    (let ((process-environment (copy-sequence process-environment)))
+      (setenv "BASH_ENV")
+      (setq term--bash-needs-EMACS-status
+            (condition-case nil
+                (call-process
+                 "bash" nil nil nil "-c"
+                 "case $BASH_VERSION in [0123].*|4.[0123].*) exit 43;; esac")
+              (error 0)))))
+  (eq 43 term--bash-needs-EMACS-status))
+
 ;; This auxiliary function cranks up the process for term-exec in
 ;; the appropriate environment.
 
@@ -1506,12 +1526,6 @@ term-exec-1
 	   (format term-termcap-format "TERMCAP="
 		   term-term-name term-height term-width)
 
-	   ;; This is for backwards compatibility with Bash 4.3 and earlier.
-	   ;; Remove this hack once Bash 4.4-or-later is common, because
-	   ;; it breaks './configure' of some packages that expect it to
-	   ;; say where to find EMACS.
-	   (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
-
 	   (format "INSIDE_EMACS=%s,term:%s" emacs-version term-protocol-version)
 	   (format "LINES=%d" term-height)
 	   (format "COLUMNS=%d" term-width))
@@ -1523,6 +1537,11 @@ term-exec-1
 	;; escape codes, so we need to see the raw output.  We will have to
 	;; do the decoding by hand on the parts that are made of chars.
 	(coding-system-for-read 'binary))
+    (when (term--bash-needs-EMACSp)
+      (setq process-environment
+            (cons
+             (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
+             process-environment)))
     (apply 'start-process name buffer
 	   "/bin/sh" "-c"
 	   (format "stty -nl echo rows %d columns %d sane 2>/dev/null;\
-- 
2.17.0

bug#20202: EMACS=t Joy and Happiness

[Phillip Lord]

Paul Eggert <eggert@cs.ucla.edu> writes:

> On 05/24/2018 01:46 PM, Phillip Lord wrote:
>> I would like to propose that we now remove this on master,
>>
>
> Given the compatibility concerns expressed, how about the attached
> patch instead? The idea is to remove the need to setenv EMACS with
> newer Bashes, while still setting EMACS for older Bashes, for backward
> compatibility.
>
> From f7efdc4d21c22a81398b06cdc58144ec2f9c7697 Mon Sep 17 00:00:00 2001
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Fri, 25 May 2018 13:25:51 -0700
> Subject: [PROPOSED] =?UTF-8?q?Don=E2=80=99t=20set=20EMACS=3Dt=20if=20Bash?=
>  =?UTF-8?q?=20is=204.4=20or=20newer?=
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
>
> * lisp/term.el (term--bash-needs-EMACS-status): New var.
> (term--bash-needs-EMACSp): New function.
> (term-exec-1): Use it instead of always setting EMACS.
> ---
>  lisp/term.el | 31 +++++++++++++++++++++++++------
>  1 file changed, 25 insertions(+), 6 deletions(-)
>
> diff --git a/lisp/term.el b/lisp/term.el
> index 017b0221ec..fa43774ae8 100644
> --- a/lisp/term.el
> +++ b/lisp/term.el
> @@ -1489,6 +1489,26 @@ term-termcap-format
>    ;; don't define :te=\\E[2J\\E[?47l\\E8:ti=\\E7\\E[?47h\
>    "Termcap capabilities supported.")
>  
> +;; This is for backwards compatibility with Bash 4.3 and earlier.
> +;; Remove this hack once Bash 4.4-or-later is reasonably universal, because
> +;; it slows down execution slightly, just before the first subshell.
> +(defvar term--bash-needs-EMACS-status nil
> +  "43 if Bash is so old that it needs EMACS set.
> +Some other integer if Bash is new or not in use.
> +Nil if unknown.")
> +(defun term--bash-needs-EMACSp ()
> +  "t if Bash is old, nil if it is new or not in use."
> +  (unless term--bash-needs-EMACS-status
> +    (let ((process-environment (copy-sequence process-environment)))
> +      (setenv "BASH_ENV")
> +      (setq term--bash-needs-EMACS-status
> +            (condition-case nil
> +                (call-process
> +                 "bash" nil nil nil "-c"
> +                 "case $BASH_VERSION in [0123].*|4.[0123].*) exit 43;; esac")
> +              (error 0)))))
> +  (eq 43 term--bash-needs-EMACS-status))
> +
>  ;; This auxiliary function cranks up the process for term-exec in
>  ;; the appropriate environment.
>  
> @@ -1506,12 +1526,6 @@ term-exec-1
>  	   (format term-termcap-format "TERMCAP="
>  		   term-term-name term-height term-width)
>  
> -	   ;; This is for backwards compatibility with Bash 4.3 and earlier.
> -	   ;; Remove this hack once Bash 4.4-or-later is common, because
> -	   ;; it breaks './configure' of some packages that expect it to
> -	   ;; say where to find EMACS.
> -	   (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
> -
>  	   (format "INSIDE_EMACS=%s,term:%s" emacs-version term-protocol-version)
>  	   (format "LINES=%d" term-height)
>  	   (format "COLUMNS=%d" term-width))
> @@ -1523,6 +1537,11 @@ term-exec-1
>  	;; escape codes, so we need to see the raw output.  We will have to
>  	;; do the decoding by hand on the parts that are made of chars.
>  	(coding-system-for-read 'binary))
> +    (when (term--bash-needs-EMACSp)
> +      (setq process-environment
> +            (cons
> +             (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
> +             process-environment)))
>      (apply 'start-process name buffer
>  	   "/bin/sh" "-c"
>  	   (format "stty -nl echo rows %d columns %d sane 2>/dev/null;\



It's a solution, afaict. However, one which is more complex than we
already have. The problem with RHEL probably only effects weird people
like you (and me!) who use a stock bash on an old RHEL, then compile the
latest Emacs. And us weird people could probably circumvent the problem
with term anyway.

You, Stefan and Eli have far more experience here clearly. Happy with
what ever you decide.

Phil

bug#20484: EMACS=t Joy and Happiness

[Eli Zaretskii]

> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Fri, 25 May 2018 13:36:03 -0700
> Cc: 20202@debbugs.gnu.org, Stefan Monnier <monnier@iro.umontreal.ca>,
> 	20484@debbugs.gnu.org
> 
> Given the compatibility concerns expressed, how about the attached patch 
> instead?

I like it because it is future proof, and makes our burden of
remembering this issue easier.

Thanks.

bug#20484: EMACS=t Joy and Happiness

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 434 bytes --]

Eli Zaretskii wrote:

> I like it because it is future proof, and makes our burden of
> remembering this issue easier.
> 
> Thanks.

You're welcome. Stefan privately emailed me some helpful comments that improved 
the patch (thanks!), and I installed the attached into master. In the hope that 
this has laid Bug#20202 to rest (at least on modern GNU systems running Bash 4.4 
or later), I'm boldly closing the bug report.

[-- Attachment #2: 0001-Don-t-set-EMACS-t-if-Bash-is-4.4-or-newer.txt --]
[-- Type: text/plain, Size: 3257 bytes --]

From 6fcab83600317e94ea7b915da7730a8c7e50226d Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Sat, 26 May 2018 13:29:06 -0700
Subject: [PATCH] =?UTF-8?q?Don=E2=80=99t=20set=20EMACS=3Dt=20if=20Bash=20i?=
 =?UTF-8?q?s=204.4=20or=20newer?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

(Thanks to Stefan Monnier for improvements to this patch.)
* lisp/term.el (term--bash-needs-EMACS-status): New var.
(term--bash-needs-EMACSp): New function.
(term-exec-1): Use it instead of always setting EMACS.
---
 lisp/term.el | 34 ++++++++++++++++++++++++++++------
 1 file changed, 28 insertions(+), 6 deletions(-)

diff --git a/lisp/term.el b/lisp/term.el
index 017b022..19e68dd 100644
--- a/lisp/term.el
+++ b/lisp/term.el
@@ -1489,6 +1489,31 @@ term-termcap-format
   ;; don't define :te=\\E[2J\\E[?47l\\E8:ti=\\E7\\E[?47h\
   "Termcap capabilities supported.")
 
+;; This private hack is for backwards compatibility with Bash 4.3 and earlier.
+;; It can be useful even when running a program other than Bash, as the
+;; program might invoke Bash as an interactive subshell.  See this thread:
+;; https://lists.gnu.org/r/emacs-devel/2018-05/msg00670.html
+;; Remove this hack and its uses once Bash 4.4-or-later is reasonably
+;; universal, because it slows down execution slightly when
+;; term--bash-needs-EMACSp is first called.
+(defvar term--bash-needs-EMACS-status nil
+  "43 if Bash is so old that it needs EMACS set.
+Some other integer if Bash is new or not in use.
+Nil if unknown.")
+(defun term--bash-needs-EMACSp ()
+  "t if Bash is old, nil if it is new or not in use."
+  (eq 43
+      (or term--bash-needs-EMACS-status
+          (setf
+           term--bash-needs-EMACS-status
+           (let ((process-environment
+                  (cons "BASH_ENV" process-environment)))
+             (condition-case nil
+                 (call-process
+                  "bash" nil nil nil "-c"
+                  "case $BASH_VERSION in [0123].*|4.[0123].*) exit 43;; esac")
+               (error 0)))))))
+
 ;; This auxiliary function cranks up the process for term-exec in
 ;; the appropriate environment.
 
@@ -1506,12 +1531,6 @@ term-exec-1
 	   (format term-termcap-format "TERMCAP="
 		   term-term-name term-height term-width)
 
-	   ;; This is for backwards compatibility with Bash 4.3 and earlier.
-	   ;; Remove this hack once Bash 4.4-or-later is common, because
-	   ;; it breaks './configure' of some packages that expect it to
-	   ;; say where to find EMACS.
-	   (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
-
 	   (format "INSIDE_EMACS=%s,term:%s" emacs-version term-protocol-version)
 	   (format "LINES=%d" term-height)
 	   (format "COLUMNS=%d" term-width))
@@ -1523,6 +1542,9 @@ term-exec-1
 	;; escape codes, so we need to see the raw output.  We will have to
 	;; do the decoding by hand on the parts that are made of chars.
 	(coding-system-for-read 'binary))
+    (when (term--bash-needs-EMACSp)
+      (push (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
+            process-environment))
     (apply 'start-process name buffer
 	   "/bin/sh" "-c"
 	   (format "stty -nl echo rows %d columns %d sane 2>/dev/null;\
-- 
2.7.4

bug#20202: EMACS=t Joy and Happiness

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 434 bytes --]

Eli Zaretskii wrote:

> I like it because it is future proof, and makes our burden of
> remembering this issue easier.
> 
> Thanks.

You're welcome. Stefan privately emailed me some helpful comments that improved 
the patch (thanks!), and I installed the attached into master. In the hope that 
this has laid Bug#20202 to rest (at least on modern GNU systems running Bash 4.4 
or later), I'm boldly closing the bug report.

[-- Attachment #2: 0001-Don-t-set-EMACS-t-if-Bash-is-4.4-or-newer.txt --]
[-- Type: text/plain, Size: 3257 bytes --]

From 6fcab83600317e94ea7b915da7730a8c7e50226d Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Sat, 26 May 2018 13:29:06 -0700
Subject: [PATCH] =?UTF-8?q?Don=E2=80=99t=20set=20EMACS=3Dt=20if=20Bash=20i?=
 =?UTF-8?q?s=204.4=20or=20newer?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

(Thanks to Stefan Monnier for improvements to this patch.)
* lisp/term.el (term--bash-needs-EMACS-status): New var.
(term--bash-needs-EMACSp): New function.
(term-exec-1): Use it instead of always setting EMACS.
---
 lisp/term.el | 34 ++++++++++++++++++++++++++++------
 1 file changed, 28 insertions(+), 6 deletions(-)

diff --git a/lisp/term.el b/lisp/term.el
index 017b022..19e68dd 100644
--- a/lisp/term.el
+++ b/lisp/term.el
@@ -1489,6 +1489,31 @@ term-termcap-format
   ;; don't define :te=\\E[2J\\E[?47l\\E8:ti=\\E7\\E[?47h\
   "Termcap capabilities supported.")
 
+;; This private hack is for backwards compatibility with Bash 4.3 and earlier.
+;; It can be useful even when running a program other than Bash, as the
+;; program might invoke Bash as an interactive subshell.  See this thread:
+;; https://lists.gnu.org/r/emacs-devel/2018-05/msg00670.html
+;; Remove this hack and its uses once Bash 4.4-or-later is reasonably
+;; universal, because it slows down execution slightly when
+;; term--bash-needs-EMACSp is first called.
+(defvar term--bash-needs-EMACS-status nil
+  "43 if Bash is so old that it needs EMACS set.
+Some other integer if Bash is new or not in use.
+Nil if unknown.")
+(defun term--bash-needs-EMACSp ()
+  "t if Bash is old, nil if it is new or not in use."
+  (eq 43
+      (or term--bash-needs-EMACS-status
+          (setf
+           term--bash-needs-EMACS-status
+           (let ((process-environment
+                  (cons "BASH_ENV" process-environment)))
+             (condition-case nil
+                 (call-process
+                  "bash" nil nil nil "-c"
+                  "case $BASH_VERSION in [0123].*|4.[0123].*) exit 43;; esac")
+               (error 0)))))))
+
 ;; This auxiliary function cranks up the process for term-exec in
 ;; the appropriate environment.
 
@@ -1506,12 +1531,6 @@ term-exec-1
 	   (format term-termcap-format "TERMCAP="
 		   term-term-name term-height term-width)
 
-	   ;; This is for backwards compatibility with Bash 4.3 and earlier.
-	   ;; Remove this hack once Bash 4.4-or-later is common, because
-	   ;; it breaks './configure' of some packages that expect it to
-	   ;; say where to find EMACS.
-	   (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
-
 	   (format "INSIDE_EMACS=%s,term:%s" emacs-version term-protocol-version)
 	   (format "LINES=%d" term-height)
 	   (format "COLUMNS=%d" term-width))
@@ -1523,6 +1542,9 @@ term-exec-1
 	;; escape codes, so we need to see the raw output.  We will have to
 	;; do the decoding by hand on the parts that are made of chars.
 	(coding-system-for-read 'binary))
+    (when (term--bash-needs-EMACSp)
+      (push (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)
+            process-environment))
     (apply 'start-process name buffer
 	   "/bin/sh" "-c"
 	   (format "stty -nl echo rows %d columns %d sane 2>/dev/null;\
-- 
2.7.4

bug#20202: EMACS=t Joy and Happiness

[Phillip Lord]

Paul Eggert <eggert@cs.ucla.edu> writes:

> Eli Zaretskii wrote:
>
>> I like it because it is future proof, and makes our burden of
>> remembering this issue easier.
>>
>> Thanks.
>
> You're welcome. Stefan privately emailed me some helpful comments that
> improved the patch (thanks!), and I installed the attached into
> master. In the hope that this has laid Bug#20202 to rest (at least on
> modern GNU systems running Bash 4.4 or later), I'm boldly closing the
> bug report.


Not wishing to reopen the bug report that you have so boldly closed,
given that this would appear to be a conservative change rather than the
breaking one I was suggesting, any reason why this can't be installed
on emacs-26?

Phil

bug#20484: EMACS=t Joy and Happiness

[Paul Eggert]

On 05/31/2018 02:07 PM, Phillip Lord wrote:
> any reason why this can't be installed
> on emacs-26?

None that I know of. Will there be a 26.2, though?

bug#20202: EMACS=t Joy and Happiness

[Paul Eggert]

On 05/31/2018 02:07 PM, Phillip Lord wrote:
> any reason why this can't be installed
> on emacs-26?

None that I know of. Will there be a 26.2, though?

bug#20484: EMACS=t Joy and Happiness

[Stefan Monnier]

> None that I know of. Will there be a 26.2, though?

I'd be surprised if we don't need to make a 26.2 before 27.1.


        Stefan

bug#20202: EMACS=t Joy and Happiness

[Stefan Monnier]

> None that I know of. Will there be a 26.2, though?

I'd be surprised if we don't need to make a 26.2 before 27.1.


        Stefan

bug#20202: EMACS=t Joy and Happiness

[Eli Zaretskii]

> Cc: Eli Zaretskii <eliz@gnu.org>, 20484@debbugs.gnu.org, emacs-devel@gnu.org,
>  20202-done@debbugs.gnu.org, monnier@iro.umontreal.ca,
>  Eli Barzilay <eli@barzilay.org>
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Thu, 31 May 2018 16:45:59 -0700
> 
> On 05/31/2018 02:07 PM, Phillip Lord wrote:
> > any reason why this can't be installed
> > on emacs-26?
> 
> None that I know of. Will there be a 26.2, though?

Yes, I'm quite sure it will be.  So if we want this change backported
to Emacs 26.2, we can do that now.

bug#20484: EMACS=t Joy and Happiness

[Eli Zaretskii]

> Cc: Eli Zaretskii <eliz@gnu.org>, 20484@debbugs.gnu.org, emacs-devel@gnu.org,
>  20202-done@debbugs.gnu.org, monnier@iro.umontreal.ca,
>  Eli Barzilay <eli@barzilay.org>
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Thu, 31 May 2018 16:45:59 -0700
> 
> On 05/31/2018 02:07 PM, Phillip Lord wrote:
> > any reason why this can't be installed
> > on emacs-26?
> 
> None that I know of. Will there be a 26.2, though?

Yes, I'm quite sure it will be.  So if we want this change backported
to Emacs 26.2, we can do that now.

bug#20484: EMACS=t Joy and Happiness

[Paul Eggert]

Eli Zaretskii wrote:
> if we want this change backported to Emacs 26.2, we can do that now.

Phillip prefers that, and I do too. Any objections?

bug#20202: EMACS=t Joy and Happiness

[Paul Eggert]

Eli Zaretskii wrote:
> if we want this change backported to Emacs 26.2, we can do that now.

Phillip prefers that, and I do too. Any objections?

bug#20484: EMACS=t Joy and Happiness

[Eli Zaretskii]

> Cc: eli@barzilay.org, emacs-devel@gnu.org, 20202@debbugs.gnu.org,
>  phillip.lord@russet.org.uk, 20484@debbugs.gnu.org, monnier@iro.umontreal.ca
> From: Paul Eggert <eggert@cs.ucla.edu>
> Date: Fri, 1 Jun 2018 07:08:56 -0700
> 
> Eli Zaretskii wrote:
> > if we want this change backported to Emacs 26.2, we can do that now.
> 
> Phillip prefers that, and I do too. Any objections?

Not from me.

Thanks.

bug#20484: EMACS=t Joy and Happiness

[Paul Eggert]

On 06/01/2018 07:14 AM, Eli Zaretskii wrote:
>> Eli Zaretskii wrote:
>>> if we want this change backported to Emacs 26.2, we can do that now.
>> Phillip prefers that, and I do too. Any objections?
> Not from me.
>
> Thanks.

You're welcome. I installed that patch into the emacs-26 branch.

bug#20202: EMACS=t Joy and Happiness

[Paul Eggert]

On 06/01/2018 07:14 AM, Eli Zaretskii wrote:
>> Eli Zaretskii wrote:
>>> if we want this change backported to Emacs 26.2, we can do that now.
>> Phillip prefers that, and I do too. Any objections?
> Not from me.
>
> Thanks.

You're welcome. I installed that patch into the emacs-26 branch.

bug#20202: 24.3; Comint mode sets a bad $EMACS

[Eli Barzilay]

I was surprised to see that compiling some random Emacs code via a
Makefile fails when running inside Emacs with an obscure

  /bin/sh: t: command not found

I found this setting in comint.el:

	  (unless (getenv "EMACS")
	    (list "EMACS=t"))

And that would obviously break such Makefile uses (and IIUC, $EMACS is
a popular choice for specifying which emacs binary to use).

It looks like this was done in this commit:

  commit cfefbbf4404963cdf042fb794e0456503aa8b591
  Author: Chong Yidong <cyd@stupidchicken.com>
  Date:   2006-11-18 21:01:33 +0000

    (comint-exec-1): Set EMACS and INSIDE_EMACS to t.

Before this commit, EMACS was set with

    (concat "EMACS=" invocation-directory invocation-name)

which doesn't look like a good idea either (if you happen to use some
other Emacs version, you still expect the shell to be a plain shell),
but it at least didn't break it.  But this change was done shortly
after:

  commit 4b1aaa8b07cf2797b5a57e2a1fd88f3ec0aa41e2
  Author: Paul Eggert <eggert@twinsun.com>
  Date:   2006-09-12 16:43:25 +0000

    * etc/NEWS: In terminal-oriented subshells, the EMACS environment
    variable now defaults to Emacs's absolute file name, instead of
    to "t".

and before that (all the way to the initial comint version in git), it
was always "t".  It looks like this was intended as a way to tell if
you're running inside emacs, which was superseded by $INSIDE_EMACS --
misc.texi says

  (It also sets the @env{EMACS} environment variable to @code{t}, if
  that environment variable is not already defined.  However, this
  environment variable is deprecated; programs that use it should
  switch to using @env{INSIDE_EMACS} instead.)

and the changelog dates this to the same date as the first commit
above.

So, since it has been deprecated for almost 8 years, it looks fine to
remove it.  If not, then setting it back to the running Emacs would
work too, but better to not do such an unexpected change, so something
like "EMACS=emacs" is probably going to be unobtrusive.

Or if there's some motivation behind intentionally making it some
string that is not an executable, then "EMACS=some-descriptive-text"
would be better.


As a sidenote, misc/efaq.texi uses $EMACS still.  (But for tcsh, so
not that anyone should care...)

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!

bug#20202: EMACS=t Joy and Happiness

[Phillip Lord]

Well, it hardly seems like Mon, 11 Apr 2016 since last we discussed this
issue.

If you remember, we managed to blitz, EMACS=t everywhere, then Paul
restored it in term.el because it broke bash otherwise. master currently
looks like this:

  ;; This is for backwards compatibility with Bash 4.3 and earlier.
  ;; Remove this hack once Bash 4.4-or-later is common, because
  ;; it breaks './configure' of some packages that expect it to
  ;; say where to find EMACS.
  (format "EMACS=%s (term:%s)" emacs-version term-protocol-version)

I worry that "Bash 4.4 or later is common" is rather indeterminate; my
desktop is now 4.4.19. but another machine I am using has bash
4.2.46. Given that Emacs-26 is now destined to come out with
EMACS=26.1.0, I would like to propose that we now remove this on master,
i.e. make it happen for Emacs-27 which should be in about 2020.

Phil