.
├── unit
│ └── …
└── compose
└── …
├── [rc] → configuration
├── cmd → command
├── [in] → start file
├── [ui-in] → start UI state in JSON format
├── [out] → end file
├── [ui-out] → output UI state in JSON format
├── [selections] → selection contents
├── [state] → selection states
└── [error] → error bypass
To test, just type run [test] in the test directory.
It will print each passing test. If a test fails, a unified context diff
is printed showing the test’s expected output and the actual output.
rc is optional
and should contain a sequence of commands,
e.g., set-option, define-command, declare-option.
rc is sourced and evaluated before the cmd key sequence is executed.
cmd is required
and should contain a key sequence that will edit the input buffer.
cmd is executed after the rc command sequence is sourced.
in is optional
and should contain the initial text loaded into the input buffer
for editing by the cmd key sequence.
ui-in is optional
and should contain the json-rpc commands sent by the UI at startup.
out is optional
and should contain the expected text generated by the cmd key sequence.
ui-out is optional
and should contain the expected UI JSON output.
selections is optional
and should contain the expected value of $kak_selections
(i.e., content of the selection separated by colons).
state is optional
and should contain the expected value of $kak_selections_desc
(i.e., range of the selections separated by colons).
If the actual output text, selections, state, or ui-out
does not match the expected content in the corresponding file,
the unit test will fail.
If there is no output, selections, state, or ui-out file,
then the unit test will always succeed.
If there is an error file, which could be empty,
then the unit test will always fail.