fixed a typo in unit test, and made boolean options default to false ins... #19
Conversation
…instead of null. added bool unit test
|
couldn't find the dev branch you mention in the contribution note... hopefully that's fine. |
|
Looks like the remote branch has been deleted. That's fine. Thanks! Having said that, default support was actually just merged in a few days ago. Can you incorporate that into the pull request so we get a clean merge? |
|
Cool! Ok, will do. |
… and made the inverse of the default specified when the value is determined
|
Made the appropriate merge. I added logic to inverse the bool default value, and of course the default is false if not specified |
| @@ -342,7 +342,7 @@ public function parse() | |||
|
|
|||
| $option = $this->getOption($name); | |||
| if ($option->isBoolean()) { | |||
| $keyvals[$name] = true; | |||
| $keyvals[$name] = !$option->getDefault();// inverse of the default, as expected | |||
There was a problem hiding this comment.
Hmm. This makes total sense, but I can imagine this potentially causing headaches if the developer isn't fully aware of this. Feels a bit awkward. Let me think on this. Totally open to hearing an argument for this case.
|
Sure, np. The way I see it in order to support bool options that are interpreted as either true or false (giving more CLI flexibility), this is essentially changing what the default initial value is, since it's always going to be the inverse of the initial "default" value. So either we enforce this behavior using the default() API, or we'd have to modify the boolean() option to allow overriding the "default" value via some other means, i.e. ...
// some way of defining what the "default" value is...
$cmd->option('d')->boolean(false); // now our -d option is defined as false? if detected |
|
Also, how would the default() API function without this inverse logic boiled into it? // it'll always be true whether -d is passed or not?
$cmd->option('d')->boolean()->default(true); |
fixed a typo in unit test, and made boolean options default to false ins...
|
Thanks. That works for me. For some reason I was getting hung up on options that imply negation in their name (e.g. --no-pager), but this totally makes sense. Bumped to v0.2.5. |
...tead of null. added bool unit test