Inspired by Georgios Petasis, I made some tests responding data other
than html.
I use a json return.
The are the commands with different param values and their results:
<?
headers type "application/json;charset=UTF-8"
fconfigure stdout {*}$param
# An Ä and a Bullet (=Äߦ)
puts "\"\u00c4\u2022\""
?>
set param | Sent data
"" | iso8859-1 + <lf>
-encoding utf-8 | utf-8 + <lf>
-encoding binary | "Ä""<lf><space> - very strange
-encoding identity | utf-8 + <lf><space>
-translation binary | "Ä""<lf><space> - very strange
So, utf-8 and identity gave good results.
The results with "binary" are quite unexpected. The bullet is translated
to '"'. The "Ä" is coded in iso8859-1.
The following script also gives a good utf-8 result:
<?
headers type "application/json;charset=UTF-8"
fconfigure stdout -translation binary
# An Ä and a Bullet (=Äߦ)
puts [encoding convertto utf-8 "\"\u00c4\u2022\""]
?>
I tested it with curl as client.
Here are my questions:
- what is the perfect stdout encoding ?
utf-8 or identity for utf-8 data and binary for binary data (for example
images).
- why is the rivet default not utf-8 as this is ok for xhtml which is
the default data type ?
Enjoy,
Harald
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
