reconsider which code points can be in hashtags

[Johann150]

I have often seen hashtags being recognized in something that was definitely not intended to be a hashtag.

for example:

T=EÔ"º8øzÄ?k=ÿ#Ô¥ÄÀ¬"Üpz_µkýQ<)ÝÑIµë|`®ÿfeäÁ©¶Æ×çcDØ6=²Áå7À¾l|<à¾,3;«V(Cµ#ÒlP;Â0·¶R»ÛW篻ø6®9ÊëDa+¼ôà¬WG´w¾½Èírs¡Ò+p\z¿L9ÊGÞ7îR

Also apparently some spacing characters may be part of hashtags which is definitely incorrect. For example a nonbreaking space (U+00A0) is recognized as part of a hashtag. https://genau.qwertqwefsday.eu/notes/901diers1g

[marihachi]

例としてあげているのは作ったものですか？実際にどれくらい発生するのかが重要です。

[Johann150]

Both examples are not by me. The first one is probably less common case, it was taken from https://genau.qwertqwefsday.eu/notes/8zvhj6kdbj

[syuilo]

#の前に空白があるか、行の先頭に無い限りハッシュタグと見なさないようにしても良いかも

[syuilo]

ただ主に日本語などの分かち書きではない言語で不便になるケースもあるかも
以下のいずれもハッシュタグと認識されなくなる

• ふー、#foo
• ふー(#foo)
• 「#foo ふー」

[Johann150]

For another example I also often see people from other Fediverse software trying to separate a hashtag from the rest of a word if they only want a part of the word to be the hashtag, e.g. #hash|tag. See for example https://genau.qwertqwefsday.eu/notes/8zwzta88ki

[marihachi]

そもそもハッシュタグを誤認識するパターンが稀なので、重要度は高くなさそう。

#の前に空白があるか、行の先頭に無い限りハッシュタグと見なさないようにしても良いかも

この案で対応するとしても、デメリットが大きい。

Also apparently some spacing characters may be part of hashtags which is definitely incorrect. For example a nonbreaking space (U+00A0) is recognized as part of a hashtag. https://genau.qwertqwefsday.eu/notes/901diers1g

これについては修正したほうが良さそう。