PGTK+ warning

PGTK+ warning

[Alessandro Bertulli]

Hi all!

I managed to compile the latest version of Emacs. I enabled the PGTK+
feature, and now when I start Emacs I got the warning (in a GTK warning
dialog box) that says:

"You are trying to run Emacs configured with the "pure-GTK" interface
under the X Window System. That configuration is unsupported and will
lead to sporadic crashes during transfer of large selection data. It
will also lead to various problems with keyboard input."

Note that I am on Wayland, but for some reason (that I'll investigate)
XWayland seem to take precedence, so effectively I'm on X, I suppose.
For now, I don't have any issues of the one described, but I wondered:

1) Why do I see this warning? Isn't PGTK supported? Is Xorg the problem?

2) Is there a way to disable this pop up? Right now it steals focus from
Emacs, and it's a bit cumbersome to manually close it every time.

Thanks!
-- 
Alessandro Bertulli

Re: PGTK+ warning

[Po Lu]

Alessandro Bertulli <alessandro.bertulli96@gmail.com> writes:

> I managed to compile the latest version of Emacs. I enabled the PGTK+
> feature, and now when I start Emacs I got the warning (in a GTK warning
> dialog box) that says:
>
> "You are trying to run Emacs configured with the "pure-GTK" interface
> under the X Window System. That configuration is unsupported and will
> lead to sporadic crashes during transfer of large selection data. It
> will also lead to various problems with keyboard input."
>
> Note that I am on Wayland, but for some reason (that I'll investigate)
> XWayland seem to take precedence, so effectively I'm on X, I suppose.

Is WAYLAND_DISPLAY set?

> For now, I don't have any issues of the one described, but I wondered:
>
> 1) Why do I see this warning? Isn't PGTK supported? Is Xorg the problem?

PGTK isn't supported under the X Window system.  The reason is a
combination of GTK being a very lousy toolkit and us not wanting to
maintain code that only runs under the GDK X11 backend when there is
working code in the regular X builds.

> 2) Is there a way to disable this pop up?

We'd rather not.  You really shouldn't be running the PGTK build under
an X server.

Re: PGTK+ warning

[Stefan Monnier via Users list for the GNU Emacs text editor]

> 2) Is there a way to disable this pop up? Right now it steals focus from
> Emacs, and it's a bit cumbersome to manually close it every time.

Since you should likely (re)start Emacs every month or so, it shouldn't
be a problem.  If you constantly start new Emacs sessions, then maybe
that's the problem you should solve first.


        Stefan

Re: PGTK+ warning

[Alessandro Bertulli]

Po Lu <luangruo@yahoo.com> writes:

> Is WAYLAND_DISPLAY set?

You mean the environment variable? Then no. I'm going to troubleshoot my
Wayland system as soon as I can.

> PGTK isn't supported under the X Window system.  The reason is a
> combination of GTK being a very lousy toolkit and us not wanting to
> maintain code that only runs under the GDK X11 backend when there is
> working code in the regular X builds.

Ah, I see, thank you! I'll keep you updated in case I need further
clarification, but I think that with Wayland fixed it will all work.

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Emanuel Berg]

Stefan Monnier via Users list for the GNU Emacs text editor wrote:

>> Is there a way to disable this pop up? Right now it steals
>> focus from Emacs, and it's a bit cumbersome to manually
>> close it every time.
>
> Since you should likely (re)start Emacs every month or so,
> it shouldn't be a problem. If you constantly start new Emacs
> sessions, then maybe that's the problem you should
> solve first.

Agreed, but that can be tricky if the problem lies with the
computer itself and not Emacs ...
  
The fans collective dB(A) is computed to

  (db 18.9 14.7 15.9 15.9) ; 22.7 dB [Elisp last]

The PSU is semi-passive

  Fractal Ion+ 560W Platinum, semi-passive [1]

But there is still some always-on fan that annoys me enough to
shut the whole thing of every time I don't use it

Maybe the GPU which I don't need anyway LOL

  msi Nvidia Geforce GT 710, 2GB DDR3, PCI-E2.0, HDMI+DL-DVI-D

?

What happens if I just remove it, can the CPU manage, does it
fall back to that automatically?

  AMD4 x86_64
  $ lscpu | grep name
  Model name: AMD Ryzen 3 3200G with Radeon Vega Graphics

Hibernation is a possibility (actually I think I installed
Debian with too little swap space) but hibernation breaks the
network link so it isn't optimal either ...

(That Elisp to do the dB is cool actually.)

[1] https://dataswamp.org/~incal/ebchw/COMPUTER

;;; -*- lexical-binding: t -*-
;;
;; this file:
;;   https://dataswamp.org/~incal/emacs-init/audio.el

(require 'cl-lib)

(defun db (d &rest ds)
  (let ((all (cons d ds))
        (sum 0) )
    (cl-loop
      for a in all
      do (cl-incf sum (expt 10 (/ a 10.0))) )
    (* 10 (/ (log sum) (log 10))) ))

;; (db 15)          ; 15
;; (db 15 15)       ; 18
;; (db 15 15 15)    ; 19.7
;; (db 15 15 15 15) ; 21

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[Alessandro Bertulli]

Stefan Monnier via Users list for the GNU Emacs text editor <help-gnu-emacs@gnu.org> writes:

> Since you should likely (re)start Emacs every month or so, it shouldn't
> be a problem.  If you constantly start new Emacs sessions, then maybe
> that's the problem you should solve first.

Actually, I do start Emacs every time, mostly because:

1) I'm still experimenting with my init file, and I want changes to be
effective. I *could* just re-evaluate my init file, but I'd then have to
deal with side-effects (like removing functions from hooks), and
restarting Emacs is just way faster.

M-x restart-emacs would be great for this precise purpose, but somehow
it close Emacs, but doesn't restart it. Maybe is it an issue of Xorg?
I'll try that as well.

2) I power off my PC every evening :-)

IIUC, are there resons for which I should keep both PC and Emacs
running?

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Emanuel Berg]

Alessandro Bertulli wrote:

> deal with side-effects (like removing functions from hooks)
> way faster

;; (setq text-mode-hook nil)

(defun text-mode-hook-f ()
  (abbrev-mode)
  (define-key text-mode-map "\C-c\C-c" #'compile) )
  
(add-hook 'text-mode-hook #'text-mode-hook-f)

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[Stefan Monnier]

> 2) I power off my PC every evening :-)

My OS and all its running applications survive just fine when I power
off my machine (via hibernation).  :-)


        Stefan

Re: PGTK+ warning

[Alessandro Bertulli]

Stefan Monnier <monnier@iro.umontreal.ca> writes:
> My OS and all its running applications survive just fine when I power
> off my machine (via hibernation).  :-)

You're right, of course! And that's what I did for quite a while. But
now I did the wonderful transition to an SSD, which means:
- I want to save writes to the disk
- Boot process is ridicously fast
so I really don't mind to boot both the OS and Emacs. That's a matter of
taste, I suppose. Anyway thanks!

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Alessandro Bertulli]

Emanuel Berg <incal@dataswamp.org> writes:
> ;; (setq text-mode-hook nil)
>
> (defun text-mode-hook-f ()
>   (abbrev-mode)
>   (define-key text-mode-map "\C-c\C-c" #'compile) )
>   
> (add-hook 'text-mode-hook #'text-mode-hook-f)

Correct, but that doesn't account for functions possibly added by other
packages. Anyway, now you make me think that this should never happen,
so my concern is probably futile. Thanks for the suggestion!

Moreover, correct me, but uncommenting also the first line makes the
init file idempotent, right?

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Eli Zaretskii]

> From: Alessandro Bertulli <alessandro.bertulli96@gmail.com>
> Cc: help-gnu-emacs@gnu.org
> Date: Tue, 15 Nov 2022 19:02:23 +0100
> 
> now I did the wonderful transition to an SSD, which means:
> - I want to save writes to the disk

This is a myth, FWIW.  Your machine will die before your SSD will run
out of its write resources.

Re: PGTK+ warning

[Emanuel Berg]

Alessandro Bertulli wrote:

>> ;; (setq text-mode-hook nil)
>>
>> (defun text-mode-hook-f ()
>>   (abbrev-mode)
>>   (define-key text-mode-map "\C-c\C-c" #'compile) )
>>   
>> (add-hook 'text-mode-hook #'text-mode-hook-f)
>
> Correct, but that doesn't account for functions possibly
> added by other packages. Anyway, now you make me think that
> this should never happen, so my concern is probably futile.
> Thanks for the suggestion!
>
> Moreover, correct me, but uncommenting also the first line
> makes the init file idempotent, right?

Interesting question ...

It removes whatever stuff might be there from the start, so
look what is there first.

But even so it should be avoided because people might put
stuff in their hooks in future versions.

So that should perhaps be changed into something that removes
the mode-hook-f function from the hook, or just sets it (the
function) to `ignore' (#'ignore) perhaps ...

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[Alessandro Bertulli]

Emanuel Berg <incal@dataswamp.org> writes:
> But even so it should be avoided because people might put
> stuff in their hooks in future versions.
>
> So that should perhaps be changed into something that removes
> the mode-hook-f function from the hook, or just sets it (the
> function) to `ignore' (#'ignore) perhaps ...

Good ideas! Might be worth reasoning about it, thanks.

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Alessandro Bertulli]

Eli Zaretskii <eliz@gnu.org> writes:

> This is a myth, FWIW.  Your machine will die before your SSD will run
> out of its write resources.

I was based on what I learned at my university, but to be fair that
was a course about industrial-use of external memory. You're way more
expert than me, so I'll gladly listen to your help. Thanks!

P.S. I'm writing from a 2015 ASUS machine, which still works great after
I switched to SSD + Linux. Long live SSD!

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Emanuel Berg]

Alessandro Bertulli wrote:

>> This is a myth, FWIW. Your machine will die before your SSD
>> will run out of its write resources.
>
> I was based on what I learned at my university, but to be
> fair that was a course about industrial-use of external
> memory

:O

What course was that?

> I'm writing from a 2015 ASUS machine, which still works
> great after I switched to SSD + Linux.

Yes, why ever not?

> Long live SSD!

Agreed, ain't it cool stuff ...

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[Emanuel Berg]

Alessandro Bertulli wrote:

>> But even so it should be avoided because people might put
>> stuff in their hooks in future versions.
>>
>> So that should perhaps be changed into something that
>> removes the mode-hook-f function from the hook, or just
>> sets it (the function) to `ignore' (#'ignore) perhaps ...
>
> Good ideas! Might be worth reasoning about it, thanks.

I realized you can simply use `remove-hook' to remove your
addition to it :)

Clear the whole thing:

  (setq conf-space-mode-hook nil)

Make your addition do nothing: (useful)

  (setq conf-space-mode-hook-f #'ignore)

Remove your addition: (useful to do after you do it)

  (remove-hook 'conf-space-mode-hook #'conf-space-mode-hook-f)

Your addition can look like this for, for example, the
`conf-space-mode-hook'

  (defun conf-space-mode-hook-f ()
    ;; do stuff
    )

Then add it, this won't remove any stuff that's already there
or will appear in the future: (pretty safe)

  (add-hook 'conf-space-mode-hook #'conf-space-mode-hook-f)

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[Eli Zaretskii]

> From: Alessandro Bertulli <alessandro.bertulli96@gmail.com>
> Cc: help-gnu-emacs@gnu.org
> Date: Wed, 16 Nov 2022 23:42:03 +0100
> 
> P.S. I'm writing from a 2015 ASUS machine, which still works great after
> I switched to SSD + Linux. Long live SSD!

The system where I'm typing this uses an SSD drive as a system volume,
which means, among other things, that the system's paging file is on
it, since 2014.  I have yet to see any problem with it.

Re: PGTK+ warning

[Alessandro Bertulli]

Emanuel Berg <incal@dataswamp.org> writes:

> I realized you can simply use `remove-hook' to remove your
> addition to it :)

Yap, this together with the one-umbrella-function approach should do the
trick, thanks! I wonder:

- Is this the same approach used by use-package?

- Are there any shortcomings in having a single function in the hook?
maybe for debugging purposes. I'd say no, since actually you can
instrument a single function to debug, it becomes even simpler, and
there aren't really other reasons, are they?

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Alessandro Bertulli]

Emanuel Berg <incal@dataswamp.org> writes:

>> I was based on what I learned at my university, but to be
>> fair that was a course about industrial-use of external
>> memory
>
> :O
>
> What course was that?

"Computing infrastructures". But I want to stress: they were referring
to data-center-usage-level of disks, in which case HDDs do live longer
than SSDs, and the performance gap can be shortened (if not bridged)
using RAID. However, with a personal workstation usage, this is not the
case, Eli is right.

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Emanuel Berg]

Alessandro Bertulli wrote:

>> What course was that?
>
> "Computing infrastructures". But I want to stress: they were
> referring to data-center-usage-level of disks, in which case
> HDDs do live longer than SSDs, and the performance gap can
> be shortened (if not bridged) using RAID. However, with
> a personal workstation usage, this is not the case, Eli
> is right.

Workstation? :)

From what century are you? :)

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[Alessandro Bertulli]

Emanuel Berg <incal@dataswamp.org> writes:

> Workstation? :)
>
> From what century are you? :)

From a very old time, apparently :-)

-- 
Alessandro Bertulli

Re: PGTK+ warning

[Emanuel Berg]

Alessandro Bertulli wrote:

>> Workstation? :)
>>
>> From what century are you? :)
>
> From a very old time, apparently :-)

AI lessons from teaching computers human languages by means of
generalization:

- A train station is a place where trains stop.

- So a workstation is a place where work ... err, whatever :)

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[Emanuel Berg]

>>> Workstation? :)
>>>
>>> From what century are you? :)
>>
>> From a very old time, apparently :-)
>
> AI lessons from teaching computers human languages by means of
> generalization:
>
> - A train station is a place where trains stop.
>
> - So a workstation is a place where work ... err, whatever :)

Why, it's true for my computer and IRC/Gnus at least :)

-- 
underground experts united
https://dataswamp.org/~incal

Re: PGTK+ warning

[David Masterson]

Emanuel Berg <incal@dataswamp.org> writes:

>>>> Workstation? :)
>>>>
>>>> From what century are you? :)
>>>
>>> From a very old time, apparently :-)
>>
>> AI lessons from teaching computers human languages by means of
>> generalization:
>>
>> - A train station is a place where trains stop.
>>
>> - So a workstation is a place where work ... err, whatever :)
>
> Why, it's true for my computer and IRC/Gnus at least :)

Strangely, stationary is what you are when you stop at a station,

but stationery is where your pen stops from time to time.

:)

-- 
David Masterson