First, just want to thank you for such an amazing library!
Some of our users have mixed ascii and non-ascii characters in their names, e.g. Sᴀᴜʟ123, which makes the DEFAULT_CALLBACKS.matcher function's regex fail, as it will either match all ascii or all non-ascii:
I've overridden this callback for now with this regex: /(?:^|\s)@((?:([A-Za-zÀ-ÿ0-9_'.+-]|[^\x00-\xff]))*)$/, which combines the two character classes into a single capturing group.
Is there a specific reason that ascii and non-ascii groupings should be considered separate?
Again, thank you for all your work on this project :)
First, just want to thank you for such an amazing library!
Some of our users have mixed ascii and non-ascii characters in their names, e.g. Sᴀᴜʟ123, which makes the
DEFAULT_CALLBACKS.matcher
function's regex fail, as it will either match all ascii or all non-ascii:https://github.com/ichord/At.js/blob/1b7a52011ec2571f73385d0c0d81a61003142050/src/default.coffee#L56
I've overridden this callback for now with this regex:
/(?:^|\s)@((?:([A-Za-zÀ-ÿ0-9_'.+-]|[^\x00-\xff]))*)$/
, which combines the two character classes into a single capturing group.Is there a specific reason that ascii and non-ascii groupings should be considered separate?
Again, thank you for all your work on this project :)