Odd problem with window geometry

Odd problem with window geometry

[Al Eisner]

I have had a consistent problem in starting up emacs on a Rocky 9 Linux
platform.  (I have no control over the machine, but it appears to be
using emacs-27.2.)  If I specify window size (with either -g or --geometry)
on the command line, I first see a window of appropriate size (which
changes when I change the values), but then as initialization completes
it collapses to a window with only 3 or 4 lines; there is also a small
change in width.  This happens whether or not I have a .emacs file.  (I
have no other initializtion file.)  I have longed used emacs on both
Redhat=6 Linix and Linux OS-7, with no such problem, but I think the
emacs version there was 25.  In case it matters, I use XMing as my X
server, but that is unchanged from my usage on the other platforms.

Does this behavior ring a bell with anyone?  I must be missing something.
Thanks.

Re: Odd problem with window geometry

[tomas]

[-- Attachment #1: Type: text/plain, Size: 1089 bytes --]

On Sun, Dec 01, 2024 at 08:56:01PM -0800, Al Eisner wrote:
> I have had a consistent problem in starting up emacs on a Rocky 9 Linux
> platform.  (I have no control over the machine, but it appears to be
> using emacs-27.2.)  If I specify window size (with either -g or --geometry)
> on the command line, I first see a window of appropriate size (which
> changes when I change the values), but then as initialization completes
> it collapses to a window with only 3 or 4 lines; there is also a small
> change in width.  This happens whether or not I have a .emacs file.  (I
> have no other initializtion file.)  I have longed used emacs on both
> Redhat=6 Linix and Linux OS-7, with no such problem, but I think the
> emacs version there was 25.  In case it matters, I use XMing as my X
> server, but that is unchanged from my usage on the other platforms.

You might try invoking emacs with --no-site-file -q to find out whether
what you see is coming from a default resp. site initialization file
(Cf. the manual, "51.4 The Emacs Initialization File").

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Odd problem with window geometry

[Al Eisner]

Good idea.   Unfortunately, it didn't help.  But thanks.

Al

On Mon, 2 Dec 2024, tomas@tuxteam.de wrote:

> You might try invoking emacs with --no-site-file -q to \x01find out whether
> what you see is coming from a default resp. site initialization file
> (Cf. the manual, "51.4 The Emacs Initialization File").
-- 

Re: Odd problem with window geometry

[tomas]

[-- Attachment #1: Type: text/plain, Size: 215 bytes --]

On Mon, Dec 02, 2024 at 01:26:55PM -0800, Al Eisner wrote:
> Good idea.   Unfortunately, it didn't help.  But thanks.

Hm.

What's left? X resources? The window manager trying to be helpful?

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Odd problem with window geometry

[Joost Kremers]

On Sun, Dec 01 2024, Al Eisner wrote:
> I have had a consistent problem in starting up emacs on a Rocky 9 Linux
> platform.  (I have no control over the machine, but it appears to be
> using emacs-27.2.)  If I specify window size (with either -g or --geometry)
> on the command line, I first see a window of appropriate size (which
> changes when I change the values), but then as initialization completes
> it collapses to a window with only 3 or 4 lines; there is also a small
> change in width.  This happens whether or not I have a .emacs file.  (I
> have no other initializtion file.)  I have longed used emacs on both
> Redhat=6 Linix and Linux OS-7, with no such problem, but I think the
> emacs version there was 25.  In case it matters, I use XMing as my X
> server, but that is unchanged from my usage on the other platforms.
>
> Does this behavior ring a bell with anyone?  I must be missing something.
> Thanks.

There was a problem similar to this with Emacs 28, I think, maybe also with
29. I actually ran into it myself, and I remember finding the solution, but
I don't remember the details... If I'm not mistaken, it was something about
the toolkit being used in combination with the windowing system. Could be a
Wayland version of Emacs on X, or the other way around, I don't remember...

FWIW, I'm running Emacs 29 compiled with PGTK on Gnome+Wayland, and the
issue does not occur anymore.

There is some discussion at
https://emacs.stackexchange.com/questions/77667/emacs-starts-in-extremely-tiny-window
and https://debbugs.gnu.org/cgi/bugreport.cgi?bug=67654, including some
suggestions for work-arounds.

-- 
Joost Kremers
Life has its moments

Re: Odd problem with window geometry

[Al Eisner]

Thanks, these links look very useful, with something I can try out.
(The second link implies the problem can also occur with emacs-29,
but the "timer" workaround looks promising.)  I'll report back.

Thanks, Al

On Tue, 3 Dec 2024, Joost Kremers wrote:

> On Sun, Dec 01 2024, Al Eisner wrote:
>> I have had a consistent problem in starting up emacs on a Rocky 9 Linux
>> platform.  (I have no control over the machine, but it appears to be
>> using emacs-27.2.)  If I specify window size (with either -g or --geometry)
>> on the command line, I first see a window of appropriate size (which
>> changes when I change the values), but then as initialization completes
>> it collapses to a window with only 3 or 4 lines; there is also a small
>> change in width.  This happens whether or not I have a .emacs file.  (I
>> have no other initializtion file.)  I have longed used emacs on both
>> Redhat=6 Linix and Linux OS-7, with no such problem, but I think the
>> emacs version there was 25.  In case it matters, I use XMing as my X
>> server, but that is unchanged from my usage on the other platforms.
>>
>> Does this behavior ring a bell with anyone?  I must be missing something.
>> Thanks.
>
> There was a problem similar to this with Emacs 28, I think, maybe also with
> 29. I actually ran into it myself, and I remember finding the solution, but
> I don't remember the details... If I'm not mistaken, it was something about
> the toolkit being used in combination with the windowing system. Could be a
> Wayland version of Emacs on X, or the other way around, I don't remember...
>
> FWIW, I'm running Emacs 29 compiled with PGTK on Gnome+Wayland, and the
> issue does not occur anymore.
>
> There is some discussion at
> https://emacs.stackexchange.com/questions/77667/emacs-starts-in-extremely-tiny-window
> and https://debbugs.gnu.org/cgi/bugreport.cgi?bug=67654, including some
> suggestions for work-arounds.
>
> --
> Joost Kremers
> Life has its moments
>

-- 

Re: Odd problem with window geometry

[Al Eisner]

Well, I've tried the two time-delay patches in the second link below
without success.  I suspect I am missing something and they may not be
executing.  For example, I increased the sleep time on the suggestion in 
the last response in that thread, but the startup doesn't seem to be
sleeping....

Al

On Tue, 3 Dec 2024, Joost Kremers wrote:

> BEWARE: This email originated outside of our organization. DO NOT CLICK links or attachments unless you recognize the sender and know the content is safe.
>
>
> On Sun, Dec 01 2024, Al Eisner wrote:
>> I have had a consistent problem in starting up emacs on a Rocky 9 Linux
>> platform.  (I have no control over the machine, but it appears to be
>> using emacs-27.2.)  If I specify window size (with either -g or --geometry)
>> on the command line, I first see a window of appropriate size (which
>> changes when I change the values), but then as initialization completes
>> it collapses to a window with only 3 or 4 lines; there is also a small
>> change in width.  This happens whether or not I have a .emacs file.  (I
>> have no other initializtion file.)  I have longed used emacs on both
>> Redhat=6 Linix and Linux OS-7, with no such problem, but I think the
>> emacs version there was 25.  In case it matters, I use XMing as my X
>> server, but that is unchanged from my usage on the other platforms.
>>
>> Does this behavior ring a bell with anyone?  I must be missing something.
>> Thanks.
>
> There was a problem similar to this with Emacs 28, I think, maybe also with
> 29. I actually ran into it myself, and I remember finding the solution, but
> I don't remember the details... If I'm not mistaken, it was something about
> the toolkit being used in combination with the windowing system. Could be a
> Wayland version of Emacs on X, or the other way around, I don't remember...
>
> FWIW, I'm running Emacs 29 compiled with PGTK on Gnome+Wayland, and the
> issue does not occur anymore.
>
> There is some discussion at
> https://emacs.stackexchange.com/questions/77667/emacs-starts-in-extremely-tiny-window
> and https://debbugs.gnu.org/cgi/bugreport.cgi?bug=67654, including some
> suggestions for work-arounds.
>
> --
> Joost Kremers
> Life has its moments
>

-- 

Re: Odd problem with window geometry

[Joel Reicher]

Al Eisner <eisner@slac.stanford.edu> writes:

> I have had a consistent problem in starting up emacs on a Rocky 
> 9 Linux platform.  (I have no control over the machine, but it 
> appears to be using emacs-27.2.)  If I specify window size (with 
> either -g or --geometry) on the command line, I first see a 
> window of appropriate size (which changes when I change the 
> values), but then as initialization completes it collapses to a 
> window with only 3 or 4 lines; there is also a small change in 
> width.  This happens whether or not I have a .emacs file.  (I 
> have no other initializtion file.)  I have longed used emacs on 
> both Redhat=6 Linix and Linux OS-7, with no such problem, but I 
> think the emacs version there was 25.  In case it matters, I use 
> XMing as my X server, but that is unchanged from my usage on the 
> other platforms.
>
> Does this behavior ring a bell with anyone?  I must be missing 
> something.

This is just a stab in the dark, but I've noticed a weird geometry 
change when the toolbar is being removed. Maybe there's also one 
happening when it's being drawn? Perhaps experiment with having it 
enabled or disabled? And maybe same for the menu bar?

Regards,

        - Joel

Re: Odd problem with window geometry

[Al Eisner]

On Wed, 4 Dec 2024, Joel Reicher wrote:

(I initially wrote)

>> I have had a consistent problem in starting up emacs on a Rocky
>> 9 Linux platform.  (I have no control over the machine, but it
>> appears to be using emacs-27.2.)  If I specify window size (with
>> either -g or --geometry) on the command line, I first see a
>> window of appropriate size (which changes when I change the
>> values), but then as initialization completes it collapses to a
>> window with only 3 or 4 lines; there is also a small change in
>> width.  This happens whether or not I have a .emacs file.  (I
>> have no other initializtion file.)  I have longed used emacs on
>> both Redhat=6 Linix and Linux OS-7, with no such problem, but I
>> think the emacs version there was 25.  In case it matters, I use
>> XMing as my X server, but that is unchanged from my usage on the
>> other platforms.
>> 
>> Does this behavior ring a bell with anyone?  I must be missing
>> something.
>
> This is just a stab in the dark, but I've noticed a weird geometry
> change when the toolbar is being removed. Maybe there's also one
> happening when it's being drawn? Perhaps experiment with having it
> enabled or disabled? And maybe same for the menu bar?
>
> Regards,

I need the menu bar and prefer to keep the toolbar as well.

But I've adequately figured this out.  This page in the Emacs Lisp manual
proved very helpful: 
https://www.gnu.org/software/emacs/manual/html_node/emacs/Frame-Parameters.html

If instead of specifying geometry on the command line I define it by adding
to default-frame-alist in my .emacs file, I see three steps when I start
emacs:  first a plausible initial window, second it shrinks to the problem
window, and third it changes to my selected defaults (provided I have selected
a fixed-width font). The manual does not mention that second step, which 
still seems anomalous, but I don't think I need to worry about it.

The one remaining issue is that if I additionally specify geometry on the
command line (which the manual says affects the initial window, not the
default window), step 1 respects that as an initial window, step 2 is still
the proble, but then for some reason step 3 is ignored, so I'm left with 
the 3-or-4-line window.  It is possible that using a timer, as suggested
in a previous thread linked by Joost, will deal with this issue.  But 
I don't really need to include geometry on the command line.

Thanks for all the responses.