yajl_encode.c File Reference

#include <assert.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include "yajl_encode.h"

Include dependency graph for yajl_encode.c:

Go to the source code of this file.

Macros
#define	ADV_PTR   s++; if (!(len--)) return 0;

Functions
void	yajl_string_encode (const yajl_print_t print, void *ctx, const unsigned char *str, size_t len, int escape_solidus)
void	yajl_string_decode (yajl_buf buf, const unsigned char *str, size_t len)
int	yajl_string_validate_utf8 (const unsigned char *s, size_t len)

Macro Definition Documentation

#define ADV_PTR   s++; if (!(len--)) return 0;

Definition at line 180 of file yajl_encode.c.

Function Documentation

void yajl_string_decode	(	yajl_buf	buf,
		const unsigned char *	str,
		size_t	len
	)

Definition at line 116 of file yajl_encode.c.

{
    size_t beg = 0;
    size_t end = 0;

    while (end < len) {
        if (str[end] == '\\') {
            char utf8Buf[5];
            const char * unescaped = "?";
            yajl_buf_append(buf, str + beg, end - beg);
            switch (str[++end]) {
                case 'r': unescaped = "\r"; break;
                case 'n': unescaped = "\n"; break;
                case '\\': unescaped = "\\"; break;
                case '/': unescaped = "/"; break;
                case '"': unescaped = "\""; break;
                case 'f': unescaped = "\f"; break;
                case 'b': unescaped = "\b"; break;
                case 't': unescaped = "\t"; break;
                case 'u': {
                    unsigned int codepoint = 0;
                    hexToDigit(&codepoint, str + ++end);
                    end+=3;
                    /* check if this is a surrogate */
                    if ((codepoint & 0xFC00) == 0xD800) {
                        end++;
                        if (str[end] == '\\' && str[end + 1] == 'u') {
                            unsigned int surrogate = 0;
                            hexToDigit(&surrogate, str + end + 2);
                            codepoint =
                                (((codepoint & 0x3F) << 10) |
                                 ((((codepoint >> 6) & 0xF) + 1) << 16) |
                                 (surrogate & 0x3FF));
                            end += 5;
                        } else {
                            unescaped = "?";
                            break;
                        }
                    }

                    Utf32toUtf8(codepoint, utf8Buf);
                    unescaped = utf8Buf;

                    if (codepoint == 0) {
                        yajl_buf_append(buf, unescaped, 1);
                        beg = ++end;
                        continue;
                    }

                    break;
                }
                default:
                    assert("this should never happen" == NULL);
            }
            yajl_buf_append(buf, unescaped, (unsigned int)strlen(unescaped));
            beg = ++end;
        } else {
            end++;
        }
    }
    yajl_buf_append(buf, str + beg, end - beg);
}

void yajl_string_encode	(	const yajl_print_t	print,
		void *	ctx,
		const unsigned char *	str,
		size_t	len,
		int	escape_solidus
	)

Definition at line 32 of file yajl_encode.c.

{
    size_t beg = 0;
    size_t end = 0;
    char hexBuf[7];
    hexBuf[0] = '\\'; hexBuf[1] = 'u'; hexBuf[2] = '0'; hexBuf[3] = '0';
    hexBuf[6] = 0;

    while (end < len) {
        const char * escaped = NULL;
        switch (str[end]) {
            case '\r': escaped = "\\r"; break;
            case '\n': escaped = "\\n"; break;
            case '\\': escaped = "\\\\"; break;
            /* it is not required to escape a solidus in JSON:
             * read sec. 2.5: http://www.ietf.org/rfc/rfc4627.txt
             * specifically, this production from the grammar:
             *   unescaped = %x20-21 / %x23-5B / %x5D-10FFFF
             */
            case '/': if (escape_solidus) escaped = "\\/"; break;
            case '"': escaped = "\\\""; break;
            case '\f': escaped = "\\f"; break;
            case '\b': escaped = "\\b"; break;
            case '\t': escaped = "\\t"; break;
            default:
                if ((unsigned char) str[end] < 32) {
                    CharToHex(str[end], hexBuf + 4);
                    escaped = hexBuf;
                }
                break;
        }
        if (escaped != NULL) {
            print(ctx, (const char *) (str + beg), end - beg);
            print(ctx, escaped, (unsigned int)strlen(escaped));
            beg = ++end;
        } else {
            ++end;
        }
    }
    print(ctx, (const char *) (str + beg), end - beg);
}

int yajl_string_validate_utf8	(	const unsigned char *	s,
		size_t	len
	)

Definition at line 182 of file yajl_encode.c.

{
    if (!len) return 1;
    if (!s) return 0;

    while (len--) {
        /* single byte */
        if (*s <= 0x7f) {
            /* noop */
        }
        /* two byte */
        else if ((*s >> 5) == 0x6) {
            ADV_PTR;
            if (!((*s >> 6) == 0x2)) return 0;
        }
        /* three byte */
        else if ((*s >> 4) == 0x0e) {
            ADV_PTR;
            if (!((*s >> 6) == 0x2)) return 0;
            ADV_PTR;
            if (!((*s >> 6) == 0x2)) return 0;
        }
        /* four byte */
        else if ((*s >> 3) == 0x1e) {
            ADV_PTR;
            if (!((*s >> 6) == 0x2)) return 0;
            ADV_PTR;
            if (!((*s >> 6) == 0x2)) return 0;
            ADV_PTR;
            if (!((*s >> 6) == 0x2)) return 0;
        } else {
            return 0;
        }

        s++;
    }

    return 1;
}