Launchpad should warn when non-ASCII characters are introduced in a merge proposal

[Junien F]

Hello,

We had an incident yesterday caused copy/pasting a parameter from a man page.
The parameter was :
--delete‐delay
which is very different from :
--delete-delay

Can you spot the difference ?
In the first case, the character between "delete" and "delay" is "‐", aka U+2010, named "hyphen".
In the second case, it's the boring ASCII dash/minus sign/hyphen.

The introduced parameter did not get accepted by the program and it breaking a script.

One layer of protection against this could be implemented in Launchpad, with a warning when a merge proposal introduces a non-ASCII character in a code base.

What are your thoughts on this ?

Note that non-ASCII characters can also be used to carry attacks through various ways, e.g. https://www.blazeinfosec.com/post/homographs-attack/ (there are many other articles about this)

Thanks