Re: [PATCH] Improve error reporting when serializing non-Unicode strings to JSON

[Philipp Stephani <p.stephani2@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1282 bytes --]

Eli Zaretskii <eliz@gnu.org> schrieb am Sa., 23. Dez. 2017 um 16:34 Uhr:

> > From: Philipp Stephani <p.stephani2@gmail.com>
> > Date: Sat, 23 Dec 2017 15:19:17 +0000
> > Cc: phst@google.com, emacs-devel@gnu.org
> >
> >  So let's use check_utf_8, as its downsides don't sound serious to me,
> >
> > Well it needs to be rewritten significantly to take a char*, length
> argument instead of the coding_system
> > struct.
>
> But creating a coding_system struct requires a single call to
> setup_coding_system, which doesn't sound too bad.
>

Unfortunately it's not that easy. coding_system has many fields that
setup_coding_system doesn't initialize.
I've tried using check_utf_8, but it doesn't seem to work:

static void
json_check_utf8 (Lisp_Object string)
{
  eassert (!STRING_MULTIBYTE (string));
  struct coding_system coding;
  setup_coding_system (Qutf_8_unix, &coding);
  /* We initialize only the fields that check_utf_8 accesses.  */
  coding.src_pos = 0;
  coding.src_pos_byte = 0;
  coding.src_chars = SCHARS (string);
  coding.src_bytes = SBYTES (string);
  coding.src_object = string;
  coding.eol_seen = EOL_SEEN_NONE;
  CHECK_TYPE (check_utf_8 (&coding) >= 0, Qutf_8_string_p, string);
}

This apparently fails to detect the unibyte string "u\xFFv" as invalid
UTF-8.

[-- Attachment #2: Type: text/html, Size: 2045 bytes --]