utf8-encode

== "czech chars: ìšèøžýáíé"} comment: {More info: http://czyborra.com/utf/ } ] shift: func [ "Takes a base-2 binary string and shifts bits" data [string! binary!] places [integer!] /left /right ][ data: enbase/base data 2 either right [ remove/part tail data negate places data: head insert/dup head data #"0" places ][ remove/part data places insert/dup tail data #"0" places ] return debase/base data 2 ] utf8-encode: func[ "Encodes the string data to UTF-8" str [any-string!] "string to encode" /local c ][ str: to-binary str forall str [ if #{79} < c: to-binary to-char first str [ remove str insert str join (#{c0} or shift/right c 6) (c and #{3F} or #{80}) str: next str ] ] to-string head str ]

because it can be replaced by multiplication/division with powers of 2? is rarely needed, and performance is not so critical. something like [ shift: [ 1 2 4 8 16 ..] my-number * shift/2 ] ?

Why convert to a binary string? Why not something like: shift: func [ "shifts bits in an integer, by default to the right" data [integer!] places [integer!] /right /left ] [ return data * ( 2 ** either left [ places ] [ 0 - places ] ]

[cut code] Very clever... but it's not utf-8. Don't you think you should translate ISO-2 characters into UTC-2 encoding? ethanak

== "chars: ìšèøžýáíé"} comment: {More info: http://czyborra.com/utf/ } ] utf8-encode: func[ "Encodes the string data to UTF-8 (from Latin-1)" str [any-string!] "string to encode" /local c ][ str: to binary! str forall str [ if 127 < c: first str [ change str to char! (c and 63 or 128) c: enbase/base to binary! to char! c 2 remove/part tail c -6 c: head insert/dup head c #"0" 6 str: insert str (#{c0} or debase/base c 2) ] ] to string! head str ] Now just to find how to encode Latin2 charset to have real czech chars:(

And this is mine (more fast and more criptic :-) I do not know utf8, i copied the logic of your code. This works also under actual View 1.2.1.3.1 I think parse is the best way to go. utf8-encode: func[ "Encodes the string data to UTF-8" str [any-string!] "string to encode" /local c h ][ ;if you remove 'copy you can change the original string parse/all copy str [ any [ h: skip ( if 127 < c: first h [ h: change h c / 64 or 192 insert h c and 63 or 128 ] ) :h skip ] ] head h ] --- Ciao Romano

BRR> [cut code] BRR> Very clever... but it's not utf-8. Don't you think you should BRR> translate ISO-2 characters into UTC-2 encoding? BRR> ethanak hmm... I was looking at it just a few minutes and there was same result as gives me utf8_encode() function in PHP, but you are right because the script I've sent does not solve my problems with our czech extended characters because it converts just from Latin-1 to UTF-8. The problem is that I don't know how to convert ISO-2 to UTC-2 (yet) but if I will have some time, I will try to find some documentation and...

RPT> And this is mine (more fast and more criptic :-) RPT> I do not know utf8, i copied the logic of your code. RPT> This works also under actual View 1.2.1.3.1 RPT> I think parse is the best way to go. RPT> utf8-encode: func[ RPT> "Encodes the string data to UTF-8" RPT> str [any-string!] "string to encode" RPT> /local c h RPT> ][ RPT> ;if you remove 'copy you can change the original string RPT> parse/all copy str [ RPT> any [ RPT> h: skip ( RPT> if 127 < c: first h [ RPT> h: change h c / 64 or 192 RPT> insert h c and 63 or 128 RPT> ] RPT> ) RPT> :h RPT> skip RPT> ] RPT> ] RPT> head h RPT> ] RPT> --- RPT> Ciao RPT> Romano Great.... I have to say... you are the winner:-) I did some tests and your code is the fastest. If you don't mind I will upload the script to the library.

No problem at all! --- Ciao Romano