Repository navigation
ZWNJ is not supported in hashtags #11761
Description
Expected behaviour
I expect ZWNJ to be included in hashtag. I need whole of "نرمافزار" be a hashtag.
Actual behaviour
By now, Hashtag only includes non-ZWNJ characters. So in the given example above, only "نرم" gets converted to link.
Steps to reproduce the problem
You just add #نرمافزار and see how it only includes نرم as hashtag.
Specifications
It is a general bug in Mastodon
Activity
nightpool commentedon Sep 4, 2019
you're gonna have to fight this one out with @marrus-sh, i have no idea what the correct behavior should be
trwnh commentedon Sep 5, 2019
FWIW Twitter/Instagram/etc do the same, and the general behavior of users is to use underscores between words in joined languages.
ahangarha commentedon Sep 5, 2019
This is not correct. Both twitter and Instagram support ZWNJ in hashtag. It is needed in Persian. ِYou may check this link from twitter.
Gargron commentedon Sep 5, 2019
So is this just about including
\u200Cin the hashtag regex or are there more characters than that?ahangarha commentedon Sep 5, 2019
I believe it only only this is needed to be there.
Fix hashtags being split by ZWNJ character
Fix hashtags being split by ZWNJ character (#11821)
marrus-sh commentedon Sep 23, 2019
late reply but concur that both ZWJ and ZWNJ should be included in hashtags; it is WORD JOINER and ZWSP which should not
i know some people used ZWJ/ZWNJ as hashtag-breakers in some cases but i have always advocated use of WJ instead
marrus-sh commentedon Sep 23, 2019
(A more lengthy response:
U+2060 WORD JOINER (WJ) and U+200B ZERO WIDTH SPACE (ZWSP) [in addition to U+FEFF ZERO WIDTH NO-BREAK SPACE (BOM), but it should never be used for this purpose] are invisible format characters for controlling the breaking (as in, line‑breaking) of text. Although there might conceivably be use‑cases for controlling (preventing) line‑breaking in hashtags, this is generally a layout concern which should be controlled through higher‑order mechanisms (CSS) and not a linguistic issue (something inherent to the hashtagʼs meaning). As invisible layout characters, they are good candidates for things like "preventing a
#from opening a hashtag by placing WJ after it", which is primarily a layout concern.U+200C ZERO WIDTH NON-JOINER (ZWNJ) and U+200D ZERO WIDTH JOINER (ZWJ) (which should probably also be added to the regex if not present @Gargron) are invisible format characters for controlling the ligation (how ligatures form) of text. This does have linguistic meaning in many scripts, including Arabic (and Persian), Devanagari, and Emoji (👩💻). So it should be allowed in hashtags and does form a significant and meaningful aspect of their value.)