I'm writing bindings to libcurl for the R programming language.
Strings in R (const char*) can be stored in one of two ways: UTF-8 or
native encoding. On unix platforms these are typically identical, but
on e.g. windows (mingw-w64) the native encoding is ISO-8859-1
(latin1).

There are a few places where strings form R get copied directly into
libcurl fields, such as CURLOPT_URL, CURLFORM_FILE, and
CURLFORM_COPYCONTENTS. Does it matter which encoding I use for these
fields?

I personally rarely use non-ascii urls, but I imagine it is important
for many foreign sites. Does libcurl assume a particular encoding for
strings?
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2015-06-09