Forum info: https://forums.limesurvey.org/forum/design-issues/121205-validation-regex-including-unicode-characters

OK, so what if use decodeHtml or htmlspecialchars_decode for html decoding the "within" parameter in the LEMRegexMatch function.
https://github.com/LimeSurvey/LimeSurvey/blob/7092a6f600d2019811b0a722d09475101c37cd85/assets/packages/expressions/em_javascript.js#L847-L858

Comments?

@DenisChenu What do you think?

Are you sure it work ?

My opinion is :
We use preg_match with PHP : https://github.com/LimeSurvey/LimeSurvey/blob/7092a6f600d2019811b0a722d09475101c37cd85/application/helpers/expressions/em_core_helper.php#L3050-L3058
Why we use different option in JS

Why not https://locutus.io/php/pcre/preg_match/ directly …
And for unicode : https://stackoverflow.com/a/12897222/2239406

Maybe start by remove gimy modifier …

PS : still wait a stable 4.X to create some own function (for example : a real word counter …)

Decoding html before running regex. This (a decode string) is similar to what the PHP side regex function gets.

PR: https://github.com/LimeSurvey/LimeSurvey/pull/1499

Tested the issue before pulling the PR, Issue exist. Tested the issue after pulling the PR, below are my findings:
The Scenario fails in below cases:
/^[\x00E1A-Z0-9\s]+$/

/^[\x{00E1}A-Z0-9\s]+$/

/^[\u00E1A-Z0-9\s]+$/

Screenshot 1: Includes scenarios where there is no Unicode included and with "á" Unicode
Screenshot 2: Includes scenarios where it works with Unicode "á" but fails for above scenarios

Please refer the attachment for more details

Actually the codes mentioned above where wrong , I tried giving as below :
^[\u00E1A-Z0-9\s]+$

^[\xE1A-Z0-9\s]+$

Working as expected

Fix committed to 3.x-LTS branch: http://bugs.limesurvey.org/plugin.php?page=Source/view&id=30366

Fixed in Release 3.23.0+200813