bug#44002: 27.1; Small window height with VcXsrv X server when built with gtk3 toolkit

[J. Scott Berg]

I don't believe there are any gtk settings that would apply; I can make this 
happen on a very minimal Arch install without any desktop environment, as well 
as more complete installs of Arch and a RHEL7 clone with KDE/GNOME desktops 
installed.

Below I've shown the a the adjust_fram_size breaks. Just before the call with 
new_width=1280, new_height=1, a full window with the correct size, a menu bar, 
and otherwise a white background is visible. After that call, the window 
retains the same size, but below the menu there is a single text line (white) 
buffer, the status line, and a single text line (white) minibuffer; the 
remainder of the window below that is black. I showed the backtrace for that 
call. After the subsequent call, the window size jumps to approximately 5 text 
lines, keeping the same buffer/status/minibuffer as before, and the last 
two-ish text lines are black. After the subsequent call, I have an 
ordinary-looking emacs window except for the fact that the main buffer is only 
3 lines long.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=0x555555c35e50, 
new_width=10, new_height=9, inhibit=5, pretend=false,
    parameter=0x3c60) at frame.c:597
597     {
(gdb) c
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=f@entry=0x555556034200, 
new_width=160, new_height=340,
    inhibit=inhibit@entry=5, pretend=pretend@entry=true, 
parameter=parameter@entry=0xee50) at frame.c:597
597     {
(gdb) c
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=f@entry=0x555556034200, 
new_width=new_width@entry=-1,
    new_height=new_height@entry=-1, inhibit=1, pretend=pretend@entry=false, 
parameter=parameter@entry=0xcde0) at frame.c:597
597     {
(gdb) c
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=0x555556034200, 
new_width=-1, new_height=-1, inhibit=3, pretend=false,
    parameter=0xbf40) at frame.c:597
597     {
(gdb) c
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=0x555556034200, 
new_width=-1, new_height=-1, inhibit=3, pretend=false,
    parameter=0xbee0) at frame.c:597
597     {
(gdb)
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=f@entry=0x555556034200, 
new_width=new_width@entry=-1,
    new_height=new_height@entry=-1, inhibit=inhibit@entry=2, 
pretend=pretend@entry=false, parameter=parameter@entry=0x96c0)
    at frame.c:597
597     {
(gdb)
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=f@entry=0x555556034200, 
new_width=1280, new_height=1224,
    inhibit=inhibit@entry=0, pretend=pretend@entry=true, 
parameter=parameter@entry=0xee80) at frame.c:597
597     {
(gdb)
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=f@entry=0x555556034200, 
new_width=new_width@entry=1280,
    new_height=new_height@entry=1224, inhibit=inhibit@entry=5, 
pretend=pretend@entry=false, parameter=parameter@entry=0xf270)
    at frame.c:597
597     {
(gdb)
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=0x555556034200, 
new_width=1280, new_height=1, inhibit=5, pretend=false,
    parameter=0x3c60) at frame.c:597
597     {
(gdb) where
#0  adjust_frame_size (f=0x555556034200, new_width=1280, new_height=1, 
inhibit=5, pretend=false, parameter=0x3c60) at frame.c:597
#1  0x00005555555a9fe3 in change_frame_size_1
    (pixelwise=<optimized out>, safe=false, delay=false, pretend=false, 
new_height=<optimized out>, new_width=<optimized out>, f=<optimized out>) at 
lisp.h:1033
#2  change_frame_size
    (pixelwise=<optimized out>, safe=false, delay=false, pretend=false, 
new_height=<optimized out>, new_width=<optimized out>, f=<optimized out>) at 
dispnew.c:5830
#3  do_pending_window_change (safe=safe@entry=false) at dispnew.c:5756
#4  0x00005555555e275f in message3_nolog (m=<optimized out>) at xdisp.c:11019
#5  0x00005555555e29e8 in message3 (m=m@entry=0x555555ff32a4) at xdisp.c:10950
#6  0x00005555556f3d90 in Fmessage (args=0x7fffffffccb0, nargs=<optimized out>) 
at editfns.c:2891
#7  Fmessage (nargs=<optimized out>, args=0x7fffffffccb0) at editfns.c:2859
#8  0x00005555556fcb6b in Ffuncall (nargs=3, args=args@entry=0x7fffffffcca8) at 
lisp.h:2110
#9  0x00005555557350c8 in exec_byte_code
    (bytestr=<optimized out>, vector=<optimized out>, maxdepth=<optimized out>, 
args_template=<optimized out>, nargs=<optimized out>, args=<optimized out>) at 
bytecode.c:633
#10 0x00005555556fcaa7 in Ffuncall (nargs=1, args=args@entry=0x7fffffffd020) at 
eval.c:2809
#11 0x00005555557350c8 in exec_byte_code
    (bytestr=<optimized out>, vector=<optimized out>, maxdepth=<optimized out>, 
args_template=<optimized out>, nargs=<optimized out>, args=<optimized out>) at 
bytecode.c:633
#12 0x00005555556fcaa7 in Ffuncall (nargs=2, args=args@entry=0x7fffffffd828) at 
eval.c:2809
#13 0x00005555557350c8 in exec_byte_code
    (bytestr=<optimized out>, vector=<optimized out>, maxdepth=<optimized out>, 
args_template=<optimized out>, nargs=<optimized out>, args=<optimized out>) at 
bytecode.c:633
#14 0x00005555556fcaa7 in Ffuncall (nargs=1, args=args@entry=0x7fffffffe1e0) at 
eval.c:2809
#15 0x00005555557350c8 in exec_byte_code
    (bytestr=<optimized out>, vector=<optimized out>, maxdepth=<optimized out>, 
args_template=<optimized out>, nargs=<optimized out>, args=<optimized out>) at 
bytecode.c:633
#16 0x00005555556ffd3d in apply_lambda (fun=0x7ffff285eb25, args=<optimized 
out>, count=4) at eval.c:2927
#17 0x00005555556fedd6 in eval_sub (form=<optimized out>) at eval.c:2349
#18 0x00005555557008d9 in Feval (form=0x7ffff29bb68b, lexical=<optimized out>) 
at eval.c:2103
#19 0x00005555556fbcd7 in internal_condition_case
    (bfun=bfun@entry=0x555555680ac0 <top_level_2>, 
handlers=handlers@entry=0x90, hfun=hfun@entry=0x555555686300 <cmd_error>)
    at eval.c:1356
#20 0x00005555556819a6 in top_level_1 (ignore=ignore@entry=0x0) at lisp.h:1033
#21 0x00005555556fbc31 in internal_catch (tag=tag@entry=0xd500, 
func=func@entry=0x555555681980 <top_level_1>, arg=arg@entry=0x0)
    at eval.c:1117
#22 0x0000555555680a48 in command_loop () at lisp.h:1033
#23 0x0000555555685f16 in recursive_edit_1 () at keyboard.c:714
#24 0x0000555555686242 in Frecursive_edit () at keyboard.c:786
#25 0x00005555555a00f7 in main (argc=2, argv=<optimized out>) at emacs.c:2062
(gdb) c
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=f@entry=0x555556034200, 
new_width=new_width@entry=-1,
    new_height=new_height@entry=-1, inhibit=inhibit@entry=2, 
pretend=pretend@entry=false, parameter=parameter@entry=0xd350)
    at frame.c:597
597     {
(gdb) c
Continuing.

Thread 1 "emacs" hit Breakpoint 1, adjust_frame_size (f=0x555556034200, 
new_width=1280, new_height=170, inhibit=5, pretend=false,
    parameter=0x3c60) at frame.c:597
597     {
(gdb) c
Continuing.

> -----Original Message-----
> From: Eli Zaretskii <eliz@gnu.org>
> Sent: Friday, October 16, 2020 3:32 AM
> To: J. Scott Berg <jsberg-bnl@outlook.com>
> Cc: 44002@debbugs.gnu.org
> Subject: Re: bug#44002: 27.1; Small window height with VcXsrv X server when
> built with gtk3 toolkit
> 
> > From: "J. Scott Berg" <jsberg-bnl@outlook.com>
> > Date: Thu, 15 Oct 2020 01:55:44 +0000
> >
> > Run "emacs -Q" on a remote machine, the local X server is VcXsrv on
> Windows
> > The window that opens will only have 5 lines.
> > This occurs when emacs is built to use the gtk3 toolkit
> > Everything is fine on the local display of the Linux machine
> > Everything is fine if emacs is built to use the gtk2 toolkit
> 
> Do you have any GTK settings that could affect this, like font
> selection or window dimensions?
> 
> If that doesn't give any clue, is it possible to you to run Emacs
> under GDB, put a breakpoint in adjust_frame_size, and tell when and by
> which code are these small dimensions specified during startup?
> 
> Thanks.