In tests/config.rs there's something like 6 tests named with a variation of hello_world2. They're testing something useful, but it's not clear what from the way the code is currently structured. We should clean these up:
hello_world: produces a listing of what ssh -G output looks like to the terminal. So far I've used this mostly as a point of reference when other tests fail. Is there another way to give the output we expected to see context when one of the other tests fail?hello_world2: produces the default ssh config, but doesn't check it against anything yethello_world3, hello_world_4, hello_world5: all of these have some variation on the theme of "what will the ssh binary parse this weird config option to?" (for many of these, there's a matching behavior in ssh2-config, but we're not demonstrating that explicitly with any tests)
In addition, quoted_token_config and multi_line_quoted_token_fails both more or less fall into that last bucket, though they're a little clearer about what they're testing and actually asserting something meaningful.
It seems like there's room to pull out a test macro that's something like:
given this ssh config
when openssh parses it
and ssh2-config parses it
then they produce the same result
In
tests/config.rsthere's something like 6 tests named with a variation ofhello_world2. They're testing something useful, but it's not clear what from the way the code is currently structured. We should clean these up:hello_world: produces a listing of whatssh -Goutput looks like to the terminal. So far I've used this mostly as a point of reference when other tests fail. Is there another way to give the output we expected to see context when one of the other tests fail?hello_world2: produces the default ssh config, but doesn't check it against anything yethello_world3,hello_world_4,hello_world5: all of these have some variation on the theme of "what will thesshbinary parse this weird config option to?" (for many of these, there's a matching behavior inssh2-config, but we're not demonstrating that explicitly with any tests)In addition,
quoted_token_configandmulti_line_quoted_token_failsboth more or less fall into that last bucket, though they're a little clearer about what they're testing and actually asserting something meaningful.It seems like there's room to pull out a test macro that's something like: