Add --show-password flag to user reset-password

[dsXLII]

First go at a PR to address #342 .

[danielbachhuber]

Great start!

[danielbachhuber]

@dsXLII Still planning to finish this up?

[dsXLII]

I'm still here, just distracted by the day job :/

Question about porcelain. For the new app password example, since you can only create one app password at a time the output is unambiguous. But here, since you can reset multiple users' passwords at the same time, it's less clear. (If I do a wp user reset-password 1 2 3 4 four passwords would be output, but it's not obvious that the passwords go in the same order as the user arguments.) What's the best way to handle that? Output them in that order, and document it? Don't allow --porcelain for multiple user arguments?

[danielbachhuber]

I'm still here, just distracted by the day job :/

No worries!

Question about porcelain.

Good question! You are correct: we don't yet have prior art within WP-CLI.

Let's follow git status and output each password on a new line:

$ git status --porcelain
?? test1
?? test2

[dsXLII]

TODO: Learn enough behat to have this check the output (it should be a password, currently "exactly 12 characters, no whitespace, etc")

[danielbachhuber]

The check doesn't need to be super accurate. In other cases, we simply confirm that STDOUT is not empty.

Behat probably doesn't support checking "exactly 12 characters, no whitespace, etc". In order to perform such a check, you'd need to write a standalone PHP file to execute the check, pass the password to the PHP file, and then have the PHP file exit with 0 or 1 depending on whether it passed or failed. But, doing so is probably overkill for this particular test.

[dsXLII]

Added a --porcelain flag. (If you do something impossible, like wp user reset-password 1 2 3 on a site where user 3 doesn't exist, it still outputs a warning, then the new passwords for users 1 and 2. I didn't compare that to the behavior for other places where there's a porcelain option.)

The test almost certainly doesn't work - I assume we should check the output to make sure it's a password (exactly 12 characters, all members of the charset used for passwords, etc), but maybe I'm overthinking it and really only need to make sure it doesn't contain "Warning:" ?

[dsmith4-godaddy]

I think this is ready for a more thorough review. (Resolved the phpcs issues, and a merge conflict from the "change passwords to be 24 characters" PR.)

[danielbachhuber]

Great work on this, @dsXLII ! Sorry I missed your comment earlier.

I made some minor formatting tweaks with 6bb7681