With large objects, use of -fpic will fail to link on certain
architectures due to size limitations of the global offset table (GOT).
For example, the link stage for glfw-wayland.so fails on sparc64, as
below. Using -fPIC avoids these size limitations.
./glfw/osmesa_context.c:73:(.text+0xf2c): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `swapIntervalEGL':
./glfw/egl_context.c:238:(.text+0xf68): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `destroyContextOSMesa':
./glfw/osmesa_context.c:80:(.text+0x10a4): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `data_offer_action':
./glfw/wl_window.c:1692:(.text+0x15e4): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `data_offer_source_actions':
./glfw/wl_window.c:1683:(.text+0x1648): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `keyboardHandleRepeatInfo':
./glfw/wl_init.c:471:(.text+0x1704): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `checkScaleChange':
./glfw/wl_window.c:55:(.text+0x1768): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `primary_selection_copy_callback_done':
./glfw/wl_window.c:1833:(.text+0x18bc): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `clipboard_copy_callback_done':
./glfw/wl_window.c:1825:(.text+0x1910): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `primary_selection_source_canceled':
./glfw/wl_window.c:1582:(.text+0x1968): relocation truncated to fit: R_SPARC_GOT13 against symbol `_glfw' defined in .bss section in /tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o
/tmp/glfw-wayland.so.SYvdxr.ltrans0.ltrans.o: in function `data_source_canceled':
./glfw/wl_window.c:1576:(.text+0x19c0): additional relocation overflows omitted from the output
collect2: error: ld returned 1 exit status
XKB has various options for using keys to switch the keyboard layout.
When these are set, XKB will return keysyms which are unknown to GLFW,
so kitty will fallback to using a keymap without the options set, which
causes the keys to be interpreted as the original keysyms.
However, when these options are set, kitty shouldn't interpret the keys.
Therefore, check for some specific keysyms and return before translating
to GLFW keysyms.
There may be more keysyms which should be ignored, but as far as I can
see from https://gitlab.freedesktop.org/xkeyboard-config/xkeyboard-config/-/blob/xkeyboard-config-2.29/symbols/group
these are the ones which can be triggered by XKB options.
Fixes#2519
The display reader count only increments when wl_display_prepare_read
succeeds. Calling wl_display_cancel_read when wl_display_prepare_read
has not succeeded results in a negative reader count, which does not
have well-defined behavior.
Calling wl_display_cancel_read immediately ensures that other readers
waiting on us will have a chance to wake up and discover the error in a
timely manner.
`glfwGetCocoaKeyEquivalent()` in `glfw/cocoa_window.m` expects the returned characters to be of type `unichar`, which won't work for all unicode characters because it is defined as `unsigned short` according to https://developer.apple.com/documentation/foundation/unichar?language=objc, which is only guaranteed to be at least 16 bits in size. The code calling this function also expects the encoding to be UTF-16.
When I added the various keys in https://github.com/kovidgoyal/kitty/pull/1928, I missed these facts. This means, that `glfwGetCocoaKeyEquivalent()` will behave unexpectedly when called with any of the new-ish keys. Luckily this function is currently only used for determining the macOS shortcut for `new_os_window` but I plan on using it more in the future.
Some of the constants, e.g. `NSBackspaceCharacter` are UTF-16 constants, so we can't just use UTF-8 everywhere.
I fixed the problem by using either UTF-8 characters packed into a `uint32_t` or UTF-16 characters in a `unichar` and then converting them to a UTF-8 encoded char string.
`NSEventModifierFlagNumericPad` isn't guaranteed to fit in a `unichar`, which made this undefined behaviour. It also didn't work. I tried to make it work using `NSEventModifierFlagNumericPad` as a modifier instead, as can be seen in this commit, but couldn't get it to work either because the constants used are native key codes and not unicode characters. Therefore the numpad keys will be removed in the next commit.
Kovid Goyal
James McCoy
Trygve Aaberge
Luflosi
Kenny Levinsen