[REBOL] Re: ISO-8859-1 question
From: kgd03011::nifty::ne::jp at: 24-Jan-2001 17:30
Hello MirclMax,
>A page I used to grab and parse through Rebol using a
>page: read to-url <urlhere>
>command .. seems to have switched from sending its content as ISO-8859-1
>(Latin1) and seems to be doing it as UTF-8 now.
>I really need the content now stored in "page" to be ISO-8859-1. So, can
>anyone tell me a way to force the "read" to pull it down as ISO-8859-1?
>Barring that, does anyone have any functions to convert "page" to it? (I
>would actually need another function to go in the reverse direction I think)
This is something I've wanted to have a whack at for quite a while, so
I threw something together. It seems to work OK, but I haven't tested it
on illegal or broken UTF-8 to see what it does in such cases. Anything that
can't be expressed as ISO-8859-1 is converted to "?", but you can easily
modify it to substitute some other string, or ignore such characters.
You can use it like this:
page: utf-iso read <some url>
Cheers,
Eric
utf-iso: func [
{convert a string from UTF-8 encoding to ISO-8859-1}
s [string!]
/local res ascii skipn skipped stretch one iso
] compose [
normal: (make bitset! [#"^(0)" - #"^(7F)"])
iso: (make bitset! [#"^(C2)" - #"^(C3)"])
skipn: (make bitset! [#"^(80)" - #"^(FF)"])
skipped: (make bitset! [#"^(80)" - #"^(BF)"])
res: copy ""
parse/all s [
any [
copy stretch some normal (append res stretch) |
copy one iso copy stretch skipped
(append res to-char
(first one) - #"^(C0)" * #"^(40)" +
((first stretch)- #"^(80)")) |
skipn any skipped (append res "?") |
some skipped (append res "!")
]
]
res
]