Null pointer dereference in X11 initialization error message

[ttencate]

Tested versions

master (9d1cbab)

System information

n/a

Issue description

godot/platform/linuxbsd/x11/display_server_x11.cpp

Lines 6032 to 6034 in 9d1cbab

	context_rd = nullptr;
	r_error = ERR_CANT_CREATE;
	ERR_FAIL_MSG(vformat("Could not initialize %s", context_rd->get_api_name()));

Steps to reproduce

Didn't actually try to trigger this, but my compiler warned about it.

Minimal reproduction project (MRP)

n/a

[timothyqiu]

FYI: It won't crash in practice because all get_api_name() simply return a string literal.

[akien-mga]

I noticed this too, as commented on the PR which introduced it:

@RandomShaper I'm running into this warning with GCC 12.3.0 on Linux when building with scons warnings=extra werror=yes tests=yes verbose=yes:

In file included from ./core/os/mutex.h:34, 
                 from ./core/os/thread_safe.h:34, 
                 from ./core/object/message_queue.h:35, 
                 from ./core/object/object.h:35, 
                 from ./core/variant/binder_common.h:35, 
                 from ./core/object/method_bind.h:34, 
                 from ./core/object/class_db.h:34, 
                 from ./core/object/ref_counted.h:34, 
                 from ./core/io/resource_uid.h:34, 
                 from ./core/io/resource.h:34, 
                 from ./core/input/input_event.h:35, 
                 from ./core/input/input.h:34, 
                 from platform/linuxbsd/joypad_linux.h:36, 
                 from platform/linuxbsd/x11/display_server_x11.h:36, 
                 from platform/linuxbsd/x11/display_server_x11.cpp:31: 
platform/linuxbsd/x11/display_server_x11.cpp: In constructor 'DisplayServerX11::DisplayServerX11(const String&, DisplayServer::WindowMode, DisplayServer::VSyncMode, uint32_t, const Vector2i*, const Vector2i&, in
t, Error&)': 
platform/linuxbsd/x11/display_server_x11.cpp:6034:45: error: 'this' pointer is null [-Werror=nonnull] 
 6034 |                         ERR_FAIL_MSG(vformat("Could not initialize %s", context_rd->get_api_name())); 
      |                                      ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 
./core/error/error_macros.h:595:95: note: in definition of macro 'ERR_FAIL_MSG' 
  595 |                 _err_print_error(FUNCTION_STR, __FILE__, __LINE__, "Method/function failed.", m_msg); \ 
      |                                                                                               ^~~~~

Weirdly enough, I don't reproduce it on my main laptop with the same distro and GCC version, with dev_build=yes dev_mode=yes scu_build=all.

Haven't looked into what the warning means yet.