* [Listview] Initial implementation for left to right packing.
* [Listview] fix movement in pack-left-to-right.
* Add element-next/prev keybinding and remap tab.
* [Listview] Change option name to 'flow' and use rofi orientation type.
* [Listview] Make listview work with reverse property.
* Update test for 2 new keybindings.
fixes: #1058
* rofi moved to libera.chat
* unify some theme headers and whitespace
* consistently use the same hashbang for bash
* improve obsolete way of testing equality
see https://github.com/koalaman/shellcheck/wiki/SC2268 for explanation
* use consistent function definitions in bash scripts
* remove duplicated author and fix whitespace
* quote all variables in scripts
fixes many warnings identified through shellcheck
* fix whitespace in scripts
* fix whitespace in non-source files
* fix spelling in non-source files
* fix whitespace and indentation in source files
* fix spelling in source files
* Add test for hidden meta data in script mode
The purpose of this is to provide support for "hidden" fields on a
script item that work for search but don't get displayed. This is
mostly to provide something similar to the optional display (but still
matchable) fields in drun like "categories" or "keywords". This also
enables the choice to display unicode icons but still allow for
searching for the keywords without needing to print them.
* Ignore the output file from test runs
* Add support for the "meta" field on script entries
This fields provides a matchable but unprinted string for entries in a
script mode list match. This means you can use one thing but provide
multiple options that can match that item without polluting the list
view or make confusing output.
* Add new test to test suite file
* tests: initialize char* in mode test
Currently, test_mode_result relies on undefined behavior.
The test calls mode_result, which checks whether the pointer is NULL.
However, the pointer was never initialized, so it may or may not be
NULL, depending on the compiler.
This caused a test failure on ppc64 and Fedora 28, apparently because in
this setting, gcc sets uninitialized pointers to NULL.
By initializing the pointer to the empty string, the behavior is defined
and the test passes on all architectures.
* mode: fix input pointer check in mode_result
Do not check whether *input (i.e., the char* the input points to) is
NULL, as this is valid. Instead, check whether the input itself is NULL.
* tests: make char* input arg in test_mode_result modifiable
The function mode_result expects a modifiable char*, initialize the
argument properly so it can be modified.
