From: Nathan Beyer <nbeyer@...>
Date: 2012-07-25T08:12:44+09:00
Subject: Re: what are the C extension analogs of String#force_encoding and String#encode?

On Thu, Jul 19, 2012 at 8:42 AM, Brian Candler <lists@ruby-forum.com> wrote:
> Nathan Beyer wrote in post #1069253:
>> I wasn't able to find anything in the README.EXT
>> (http://svn.ruby-lang.org/repos/ruby/tags/v1_9_3_194/README.EXT).
>
> Welcome to ruby 1.9.x, where everything to do with string encodings is
> completely undocumented.

I've noticed ... :)

>
>> The implementation of String#force_encoding does a bit more than that:
>>
>> static VALUE
>> rb_str_force_encoding(VALUE str, VALUE enc)
>> {
>>     str_modifiable(str);
>>     rb_enc_associate(str, rb_to_encoding(enc));
>>     ENC_CODERANGE_CLEAR(str);
>>     return str;
>> }
>>
>> This implementation adds an invocation of the ENC_CODERANGE_CLEAR
>> macro and I'm not clear on what this does or when/why it would be
>> needed.
>
> I believe this is essentially the same as rb_str_modify(). It clears the
> cache of properties like 'ascii_only?' and 'valid_encoding?', so that
> next time someone queries them it has to scan the whole string.
>
>> = String#encode
>> I wasn't able to find any documentation about String transcoding
>> within C extensions. Based on the implementation of String#encode,
>> there seems to be a function 'rb_str_encode' in 'ruby/encoding.h' that
>> might be appropriate, but there's no documentation for the method and
>> I couldn't completely reverse engineer how the 'ecflags' and 'ecopts'
>> arguments are used.
>
> Maybe easier just to rb_funcall it?

That's what I've done, but I wanted to check if there was something
more appropriate I should be doing while in the C code.