Browser Support • Re: 34.2.0: Problem with rendering Arabic text on Linux

Best guess without doing any direct testing is this is a result of PR 2980 - the focus was on timestamps, but I also added it to CallICU() when wsCompat is true. I didn't expect it'd apply in regular text display.

Luckily there's a handful of other ICU whitespace functions we can test to be less replacive (u_isWhitespace, u_isJavaSpaceChar, u_isspace, and u_isblank) should this hypothesis prove true. If it's not this, I'm not sure what; since everything else should have only targeted date/time and the languages specifically mentioned in the change log.