At patchlevel 9.1.1725, fatal error for Wayland in Normal build only

56 views
Skip to first unread message

Tony Mechelynck

unread,
Sep 1, 2025, 5:43:55 PM (5 days ago) Sep 1
to Christian Brabandt, vim_dev
In file included from clipboard.c:36:
wayland.h:17:10: fatal error: wayland-client.h: No such file or directory
17 | #include <wayland-client.h>
| ^~~~~~~~~~~~~~~~~~
compilation terminated.


This is a Normal build with Motif GUI. I suspect a possible missing
include directory in the compile parameters for the Normal build. This
compile failure means the build stops then and there without trying to
compile additional modules or /a fortiori/ to link.

Huge (implicitly +wayland +wayland_clipboard +wayland_focus_steal) and
Tiny (implicitly -wayland -wayland_clipboard -wayland_focus_steal)
build OK.

The Huge builds (with +++wayland) have (among others)
-I/usr/include/wayland in the compile arguments and -lwayland-client
in the link arguments.

I'm going (for the time being) to uncomment --without-wayland in the
"Normal" configure arguments then "build reconfig".

Best regards,
Tony.

John

unread,
Sep 1, 2025, 8:34:42 PM (4 days ago) Sep 1
to vim...@googlegroups.com
I have this build failure on archlinux at the same patch level (non-GUI
huge build):
<snip>
gcc -Iproto -DHAVE_CONFIG_H -I/usr/include/gtk-2.0
-I/usr/lib/x86_64-linux-gnu/gtk-2.0/include -I/usr/include/pango-1.0
-I/usr/include/glib-2.0 -I/usr/lib/x86_64-linux-gnu/glib-2.0/include
-I/usr/include/harfbuzz -I/usr/include/freetype2 -I/usr/include/libpng16
-I/usr/include/libmount -I/usr/include/blkid -I/usr/include/fribidi
-I/usr/include/uuid -I/usr/include/cairo -I/usr/include/pixman-1
-I/usr/include/gdk-pixbuf-2.0 -I/usr/include/x86_64-linux-gnu
-I/usr/include/atk-1.0 -pthread -c -I. -Iproto -DHAVE_CONFIG_H     -O
-Wall -Wextra -Wshadow -Wstrict-prototypes -Wmissing-prototypes
-Wno-deprecated-declarations -Wno-error=missing-field-initializers
-Wno-error=maybe-uninitialized -Wunused-but-set-variable  -g -O0
-DABORT_ON_INTERNAL_ERROR -DEXITFREE -fsanitize-recover=all
-fsanitize=address -fsanitize=undefined -fno-omit-frame-pointer
 -D_REENTRANT -o objects/clipboard.o clipboard.c
In file included from clipboard.c:36:
wayland.h:20:11: fatal error:
auto/wayland/wlr-data-control-unstable-v1.h: No such file or directory
   20 | # include "auto/wayland/wlr-data-control-unstable-v1.h"
      |           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[2]: *** [Makefile:3237: objects/clipboard.o] Error 1
make[2]: Leaving directory '/home/john/vim_dev/linux/vim-9.1.1725/src'
make[1]: *** [Makefile:2037: reconfig] Error 2
make[1]: Leaving directory '/home/john/vim_dev/linux/vim-9.1.1725/src'
make: *** [Makefile:29: first] Error 2
</snip>

Cheers
John

Christian Brabandt

unread,
Sep 2, 2025, 3:10:22 AM (4 days ago) Sep 2
to vim...@googlegroups.com
It seems to be the same issue as this one:
https://github.com/vim/vim/issues/18187

Try to go into src/auto/wayland and run make manually there.

I will revert that patch probably.

Thanks,
Christian
--
Demographic polls show that you have lost credibility across the board.
Especially with those 14 year-old Valley girls.

Tony Mechelynck

unread,
Sep 2, 2025, 4:51:11 AM (4 days ago) Sep 2
to vim...@googlegroups.com
This may solve John's problem, it doesn't solve mine ("No such file or
directory" error for #include <wayland-client.h> in Normal build
only). I notice that the compile arguments for the Normal build (with
Motif GUI) include no other -I than -I. -Iproto while those for the
Huge build (with GTK3 GUI) include quite a lot of them; one of them is
-I/usr/include/wayland — and indeed, wayland-client.h exists in
/usr/include/wayland.

Best regards,
Tony.

Tony Mechelynck

unread,
Sep 2, 2025, 11:55:05 PM (3 days ago) Sep 2
to Christian Brabandt, vim...@googlegroups.com
At patchlevel 9.1.1730 (and, I suppose, because of patch 1726 having
undone patch 1725) my Normal build compiles again with no error
without the need to configure it as --without-wayland. The
+wayland_focus_steal compile-time setting apparently doesn't exist
anymore.

By looking at the log from the Normal build's "make reconfig" (with
Motif GUI) I notice that it has now spontaneously run "make" in the
auto/wayland subdirectory, and that an argument
-I/usr/include/wayland, not mentioned in the "Compilation" line at the
bottom of the Normal build's ":version" output, is present only in the
compile lines for wayland.c and for the four .c files from the
auto/wayland/ subdirectory. Tricky, but AFAICT it works.

In the log from the Huge build (with GTK3), I see that there that same
-I argument is present in the compile lines for all modules among all
other -I arguments (and is mentioned there by ":version"), and in
addition, for these five modules only, it is present a second time not
mentioned by ":version" between -D_FORTIFY_SOURCE=1 (the last
compilation argument mentioned by "version") and the -o something.o
something.c arguments specific to each module.

Best regards,
Tony.

Christian Brabandt

unread,
Sep 3, 2025, 2:15:30 AM (3 days ago) Sep 3
to Tony Mechelynck, vim...@googlegroups.com

On Mi, 03 Sep 2025, Tony Mechelynck wrote:

> At patchlevel 9.1.1730 (and, I suppose, because of patch 1726 having
> undone patch 1725) my Normal build compiles again with no error
> without the need to configure it as --without-wayland. The
> +wayland_focus_steal compile-time setting apparently doesn't exist
> anymore.
>
> By looking at the log from the Normal build's "make reconfig" (with
> Motif GUI) I notice that it has now spontaneously run "make" in the
> auto/wayland subdirectory, and that an argument
> -I/usr/include/wayland, not mentioned in the "Compilation" line at the
> bottom of the Normal build's ":version" output, is present only in the
> compile lines for wayland.c and for the four .c files from the
> auto/wayland/ subdirectory. Tricky, but AFAICT it works.

Yes, that's what it was before the patch.

> In the log from the Huge build (with GTK3), I see that there that same
> -I argument is present in the compile lines for all modules among all
> other -I arguments (and is mentioned there by ":version"), and in
> addition, for these five modules only, it is present a second time not
> mentioned by ":version" between -D_FORTIFY_SOURCE=1 (the last
> compilation argument mentioned by "version") and the -o something.o
> something.c arguments specific to each module.

Do you think this is a problem?


Thanks,
Christian
--
It is impossible to make anything foolproof because fools are so ingenious.

Tony Mechelynck

unread,
Sep 3, 2025, 11:33:49 AM (3 days ago) Sep 3
to Tony Mechelynck, vim...@googlegroups.com
No I don't. "Belt and suspenders" policy, or as a French sailors'
proverb, "Trop fort n'a jamais manqué".
>
>
> Thanks,
> Christian
> --
> It is impossible to make anything foolproof because fools are so ingenious.

:-D :-D :-D

Best regards,
Tony.
Reply all
Reply to author
Forward
0 new messages