On Wed, 28 Feb 2024 01:17:42 GMT
"Maurice Kinal -> Serak the Preparer"  wrote:

 MK>  :read !trans -b -no-ansi -t he "To pronounce it correctly, I would
 MK> have to pull out your tongue." | sed 's/^ *//'
 MK> .จจจ จจจจจ จจ จจจจจ จจจจจ ,จจจจ จจ จจ
 MK> จจจจ จจจ
 MK>
 MK> Note that the above is Hebrew ... or should be ... and I took the
 MK> liberty of stripping out the leading spaces since a 4k terminal results

The hebrew letters appear to be fine under my nntp read Sylpheed.

จ    +1

However, an nntp reader is outside the scope of what fidonet is.


 MK> As shown the Hebrew Subject line is exactly 46 utf-8 characters which
 MK> *should* fit into a FTN Subject which can be up to 71 characters.
 MK> However a byte count yeilds 82 bytes which the typical FTN abandonware
 MK> will consider too long and chop off the last (actually first since
 MK> Hebrew is read right to left) 12 bytes.  Note I say bytes and not
 MK> characters since DOS-think types cannot count straight, especially when
 MK> they treat stings as numbers.  Very sad.

Solution: FTS should limit utf-8 chars only to the message body only.


 MK> So much for utf-8 and FTS-5003.  As is that document is sadly lacking
 MK> wrt utf-8 charcters and probably should have it's own documentation in
 MK> order to escape crippleware developers.

I would tend to agree.

Anyhow..  it seems that utf-8 was "approved" without the evidence of ample usage of utf-8.


 MK> Also there is no such thing as LATIN-1.  It is the end result of some
 MK> Gatesian nightmare you can never wake up from.

No comment.

---