On Sun, Jul 8, 2018 at 1:26 PM, <minf...@apache.org> wrote:
> Author: minfrin
> Date: Sun Jul 8 11:26:00 2018
> New Revision: 1835348
[]
> Added:
[]
> apr/apr/trunk/json/apr_json.c (with props)
[]
> +
> +static apr_status_t apr_json_decode_string(apr_json_scanner_t * self,
> apr_json_string_t * retval)
> +{
> + apr_status_t status = APR_SUCCESS;
> + apr_json_string_t string;
> + const char *p = self->p;
> + const char *e;
> + char *q;
> +
> + if (self->p >= self->e) {
> + status = APR_EOF;
> + goto out;
> + }
> +
> + self->p++; /* eat the leading '"' */
> +
> + /* advance past the \ " */
> + string.len = 0;
> + for (p = self->p, e = self->e; p < e;) {
> + if (*p == '"')
> + break;
> + else if (*p == '\\') {
> + p++;
> + if (p >= e) {
> + status = APR_EOF;
> + goto out;
> + }
> + if (*p == 'u') {
> + if (p + 4 >= e) {
> + status = APR_EOF;
> + goto out;
> + }
> + p += 5;
> + string.len += 4;/* an UTF-8 character spans at most 4 bytes
> */
> + break;
Why break here and ignore potential/subsequent '\u....' encodings in
the same json string?
> + }
> + else {
> + string.len++;
> + p++;
> + }
> + }
> + else {
> + string.len++;
> + p++;
> + }
> + }
> +
> + string.p = q = apr_pcalloc(self->pool, string.len + 1);
> + e = p;
(below this anything from 'e', thus after the first '\u....', seems to
be ignored)
Regards,
Yann.
