bug#24239: 25.1; Potential regression: Multibyte text in HTTP request

[Dmitry Gutov]

On 08/16/2016 09:23 AM, Damien Cassou wrote:

> - what will happen to my multibyte characters (such as ⇒) that I'm using
>   for months and that worked well with Emacs 24?

Unibyte != ASCII. Use encode-coding-string.

> - where is the behavior change discussed in the Emacs' changelog?

We normally don't mention bug fixes in there. Binding url-request-data 
to a string with multibyte characters never actually worked properly (it 
resulted in a wrong Content-Length value). Some servers may have ignored 
that, but some may have even corrupted data silently.

> - where is the unibyte conversion requirement discussed in the url
>   library documentation?

Patch welcome?