/work/workdir/UnpackedTarball/md4c/src/md4c-html.c

Source
/*
 * MD4C: Markdown parser for C
 * (http://github.com/mity/md4c)
 *
 * Copyright (c) 2016-2024 Martin Mitáš
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 */

#include <stdio.h>
#include <string.h>

#include "md4c-html.h"
#include "entity.h"


#if !defined(__STDC_VERSION__) || __STDC_VERSION__ < 199409L
    /* C89/90 or old compilers in general may not understand "inline". */
    #if defined __GNUC__
        #define inline __inline__
    #elif defined _MSC_VER
        #define inline __inline
    #else
        #define inline
    #endif
#endif

#ifdef _WIN32
    #define snprintf _snprintf
#endif



typedef struct MD_HTML_tag MD_HTML;
struct MD_HTML_tag {
    void (*process_output)(const MD_CHAR*, MD_SIZE, void*);
    void* userdata;
    unsigned flags;
    int image_nesting_level;
    char escape_map[256];
};

#define NEED_HTML_ESC_FLAG   0x1
#define NEED_URL_ESC_FLAG    0x2


/*****************************************
 ***  HTML rendering helper functions  ***
 *****************************************/

#define ISDIGIT(ch)     ('0' <= (ch) && (ch) <= '9')
#define ISLOWER(ch)     ('a' <= (ch) && (ch) <= 'z')
#define ISUPPER(ch)     ('A' <= (ch) && (ch) <= 'Z')
#define ISALNUM(ch)     (ISLOWER(ch) || ISUPPER(ch) || ISDIGIT(ch))


static inline void
render_verbatim(MD_HTML* r, const MD_CHAR* text, MD_SIZE size)
{
    r->process_output(text, size, r->userdata);
}

/* Keep this as a macro. Most compiler should then be smart enough to replace
 * the strlen() call with a compile-time constant if the string is a C literal. */
#define RENDER_VERBATIM(r, verbatim)                                    \
        render_verbatim((r), (verbatim), (MD_SIZE) (strlen(verbatim)))


static void
render_html_escaped(MD_HTML* r, const MD_CHAR* data, MD_SIZE size)
{
    MD_OFFSET beg = 0;
    MD_OFFSET off = 0;

    /* Some characters need to be escaped in normal HTML text. */
    #define NEED_HTML_ESC(ch)   (r->escape_map[(unsigned char)(ch)] & NEED_HTML_ESC_FLAG)

    while(1) {
        /* Optimization: Use some loop unrolling. */
        while(off + 3 < size  &&  !NEED_HTML_ESC(data[off+0])  &&  !NEED_HTML_ESC(data[off+1])
                              &&  !NEED_HTML_ESC(data[off+2])  &&  !NEED_HTML_ESC(data[off+3]))
            off += 4;
        while(off < size  &&  !NEED_HTML_ESC(data[off]))
            off++;

        if(off > beg)
            render_verbatim(r, data + beg, off - beg);

        if(off < size) {
            switch(data[off]) {
                case '&':   RENDER_VERBATIM(r, "&amp;"); break;
                case '<':   RENDER_VERBATIM(r, "&lt;"); break;
                case '>':   RENDER_VERBATIM(r, "&gt;"); break;
                case '"':   RENDER_VERBATIM(r, "&quot;"); break;
            }
            off++;
        } else {
            break;
        }
        beg = off;
    }
}

static void
render_url_escaped(MD_HTML* r, const MD_CHAR* data, MD_SIZE size)
{
    static const MD_CHAR hex_chars[] = "0123456789ABCDEF";
    MD_OFFSET beg = 0;
    MD_OFFSET off = 0;

    /* Some characters need to be escaped in URL attributes. */
    #define NEED_URL_ESC(ch)    (r->escape_map[(unsigned char)(ch)] & NEED_URL_ESC_FLAG)

    while(1) {
        while(off < size  &&  !NEED_URL_ESC(data[off]))
            off++;
        if(off > beg)
            render_verbatim(r, data + beg, off - beg);

        if(off < size) {
            char hex[3];

            switch(data[off]) {
                case '&':   RENDER_VERBATIM(r, "&amp;"); break;
                default:
                    hex[0] = '%';
                    hex[1] = hex_chars[((unsigned)data[off] >> 4) & 0xf];
                    hex[2] = hex_chars[((unsigned)data[off] >> 0) & 0xf];
                    render_verbatim(r, hex, 3);
                    break;
            }
            off++;
        } else {
            break;
        }

        beg = off;
    }
}

static unsigned
hex_val(char ch)
{
    if('0' <= ch && ch <= '9')
        return ch - '0';
    if('A' <= ch && ch <= 'Z')
        return ch - 'A' + 10;
    else
        return ch - 'a' + 10;
}

static void
render_utf8_codepoint(MD_HTML* r, unsigned codepoint,
                      void (*fn_append)(MD_HTML*, const MD_CHAR*, MD_SIZE))
{
    static const MD_CHAR utf8_replacement_char[] = { (char)0xef, (char)0xbf, (char)0xbd };

    unsigned char utf8[4];
    size_t n;

    if(codepoint <= 0x7f) {
        n = 1;
        utf8[0] = codepoint;
    } else if(codepoint <= 0x7ff) {
        n = 2;
        utf8[0] = 0xc0 | ((codepoint >>  6) & 0x1f);
        utf8[1] = 0x80 + ((codepoint >>  0) & 0x3f);
    } else if(codepoint <= 0xffff) {
        n = 3;
        utf8[0] = 0xe0 | ((codepoint >> 12) & 0xf);
        utf8[1] = 0x80 + ((codepoint >>  6) & 0x3f);
        utf8[2] = 0x80 + ((codepoint >>  0) & 0x3f);
    } else {
        n = 4;
        utf8[0] = 0xf0 | ((codepoint >> 18) & 0x7);
        utf8[1] = 0x80 + ((codepoint >> 12) & 0x3f);
        utf8[2] = 0x80 + ((codepoint >>  6) & 0x3f);
        utf8[3] = 0x80 + ((codepoint >>  0) & 0x3f);
    }

    if(0 < codepoint  &&  codepoint <= 0x10ffff)
        fn_append(r, (char*)utf8, (MD_SIZE)n);
    else
        fn_append(r, utf8_replacement_char, 3);
}

/* Translate entity to its UTF-8 equivalent, or output the verbatim one
 * if such entity is unknown (or if the translation is disabled). */
static void
render_entity(MD_HTML* r, const MD_CHAR* text, MD_SIZE size,
              void (*fn_append)(MD_HTML*, const MD_CHAR*, MD_SIZE))
{
    if(r->flags & MD_HTML_FLAG_VERBATIM_ENTITIES) {
        render_verbatim(r, text, size);
        return;
    }

    /* We assume UTF-8 output is what is desired. */
    if(size > 3 && text[1] == '#') {
        unsigned codepoint = 0;

        if(text[2] == 'x' || text[2] == 'X') {
            /* Hexadecimal entity (e.g. "&#x1234abcd;")). */
            MD_SIZE i;
            for(i = 3; i < size-1; i++)
                codepoint = 16 * codepoint + hex_val(text[i]);
        } else {
            /* Decimal entity (e.g. "&1234;") */
            MD_SIZE i;
            for(i = 2; i < size-1; i++)
                codepoint = 10 * codepoint + (text[i] - '0');
        }

        render_utf8_codepoint(r, codepoint, fn_append);
        return;
    } else {
        /* Named entity (e.g. "&nbsp;"). */
        const ENTITY* ent;

        ent = entity_lookup(text, size);
        if(ent != NULL) {
            render_utf8_codepoint(r, ent->codepoints[0], fn_append);
            if(ent->codepoints[1])
                render_utf8_codepoint(r, ent->codepoints[1], fn_append);
            return;
        }
    }

    fn_append(r, text, size);
}

static void
render_attribute(MD_HTML* r, const MD_ATTRIBUTE* attr,
                 void (*fn_append)(MD_HTML*, const MD_CHAR*, MD_SIZE))
{
    int i;

    for(i = 0; attr->substr_offsets[i] < attr->size; i++) {
        MD_TEXTTYPE type = attr->substr_types[i];
        MD_OFFSET off = attr->substr_offsets[i];
        MD_SIZE size = attr->substr_offsets[i+1] - off;
        const MD_CHAR* text = attr->text + off;

        switch(type) {
            case MD_TEXT_NULLCHAR:  render_utf8_codepoint(r, 0x0000, render_verbatim); break;
            case MD_TEXT_ENTITY:    render_entity(r, text, size, fn_append); break;
            default:                fn_append(r, text, size); break;
        }
    }
}


static void
render_open_ol_block(MD_HTML* r, const MD_BLOCK_OL_DETAIL* det)
{
    char buf[64];

    if(det->start == 1) {
        RENDER_VERBATIM(r, "<ol>\n");
        return;
    }

    snprintf(buf, sizeof(buf), "<ol start=\"%u\">\n", det->start);
    RENDER_VERBATIM(r, buf);
}

static void
render_open_li_block(MD_HTML* r, const MD_BLOCK_LI_DETAIL* det)
{
    if(det->is_task) {
        RENDER_VERBATIM(r, "<li class=\"task-list-item\">"
                          "<input type=\"checkbox\" class=\"task-list-item-checkbox\" disabled");
        if(r->flags & MD_HTML_FLAG_XHTML) RENDER_VERBATIM(r, "=\"true\"");
        if(det->task_mark == 'x' || det->task_mark == 'X') {
            RENDER_VERBATIM(r, " checked");
            if(r->flags & MD_HTML_FLAG_XHTML) RENDER_VERBATIM(r, "=\"true\"");
        }
        RENDER_VERBATIM(r, (r->flags & MD_HTML_FLAG_XHTML) ? " />" : ">");
    } else {
        RENDER_VERBATIM(r, "<li>");
    }
}

static void
render_open_code_block(MD_HTML* r, const MD_BLOCK_CODE_DETAIL* det)
{
    RENDER_VERBATIM(r, "<pre><code");

    /* If known, output the HTML 5 attribute class="language-LANGNAME". */
    if(det->lang.text != NULL) {
        RENDER_VERBATIM(r, " class=\"");
        if(det->lang.size < 9  ||  strncmp(det->lang.text, "language-", 9) != 0) {
            RENDER_VERBATIM(r, "language-");
        }
        render_attribute(r, &det->lang, render_html_escaped);
        RENDER_VERBATIM(r, "\"");
    }

    RENDER_VERBATIM(r, ">");
}

static void
render_open_td_block(MD_HTML* r, const MD_CHAR* cell_type, const MD_BLOCK_TD_DETAIL* det)
{
    RENDER_VERBATIM(r, "<");
    RENDER_VERBATIM(r, cell_type);

    switch(det->align) {
        case MD_ALIGN_LEFT:     RENDER_VERBATIM(r, " align=\"left\">"); break;
        case MD_ALIGN_CENTER:   RENDER_VERBATIM(r, " align=\"center\">"); break;
        case MD_ALIGN_RIGHT:    RENDER_VERBATIM(r, " align=\"right\">"); break;
        default:                RENDER_VERBATIM(r, ">"); break;
    }
}

static void
render_open_a_span(MD_HTML* r, const MD_SPAN_A_DETAIL* det)
{
    RENDER_VERBATIM(r, "<a href=\"");
    render_attribute(r, &det->href, render_url_escaped);

    if(det->title.text != NULL) {
        RENDER_VERBATIM(r, "\" title=\"");
        render_attribute(r, &det->title, render_html_escaped);
    }

    RENDER_VERBATIM(r, "\">");
}

static void
render_open_img_span(MD_HTML* r, const MD_SPAN_IMG_DETAIL* det)
{
    RENDER_VERBATIM(r, "<img src=\"");
    render_attribute(r, &det->src, render_url_escaped);

    RENDER_VERBATIM(r, "\" alt=\"");
}

static void
render_close_img_span(MD_HTML* r, const MD_SPAN_IMG_DETAIL* det)
{
    if(det->title.text != NULL) {
        RENDER_VERBATIM(r, "\" title=\"");
        render_attribute(r, &det->title, render_html_escaped);
    }

    RENDER_VERBATIM(r, (r->flags & MD_HTML_FLAG_XHTML) ? "\" />" : "\">");
}

static void
render_open_wikilink_span(MD_HTML* r, const MD_SPAN_WIKILINK_DETAIL* det)
{
    RENDER_VERBATIM(r, "<x-wikilink data-target=\"");
    render_attribute(r, &det->target, render_html_escaped);

    RENDER_VERBATIM(r, "\">");
}


/**************************************
 ***  HTML renderer implementation  ***
 **************************************/

static int
enter_block_callback(MD_BLOCKTYPE type, void* detail, void* userdata)
{
    static const MD_CHAR* head[6] = { "<h1>", "<h2>", "<h3>", "<h4>", "<h5>", "<h6>" };
    MD_HTML* r = (MD_HTML*) userdata;

    switch(type) {
        case MD_BLOCK_DOC:      /* noop */ break;
        case MD_BLOCK_QUOTE:    RENDER_VERBATIM(r, "<blockquote>\n"); break;
        case MD_BLOCK_UL:       RENDER_VERBATIM(r, "<ul>\n"); break;
        case MD_BLOCK_OL:       render_open_ol_block(r, (const MD_BLOCK_OL_DETAIL*)detail); break;
        case MD_BLOCK_LI:       render_open_li_block(r, (const MD_BLOCK_LI_DETAIL*)detail); break;
        case MD_BLOCK_HR:       RENDER_VERBATIM(r, (r->flags & MD_HTML_FLAG_XHTML) ? "<hr />\n" : "<hr>\n"); break;
        case MD_BLOCK_H:        RENDER_VERBATIM(r, head[((MD_BLOCK_H_DETAIL*)detail)->level - 1]); break;
        case MD_BLOCK_CODE:     render_open_code_block(r, (const MD_BLOCK_CODE_DETAIL*) detail); break;
        case MD_BLOCK_HTML:     /* noop */ break;
        case MD_BLOCK_P:        RENDER_VERBATIM(r, "<p>"); break;
        case MD_BLOCK_TABLE:    RENDER_VERBATIM(r, "<table>\n"); break;
        case MD_BLOCK_THEAD:    RENDER_VERBATIM(r, "<thead>\n"); break;
        case MD_BLOCK_TBODY:    RENDER_VERBATIM(r, "<tbody>\n"); break;
        case MD_BLOCK_TR:       RENDER_VERBATIM(r, "<tr>\n"); break;
        case MD_BLOCK_TH:       render_open_td_block(r, "th", (MD_BLOCK_TD_DETAIL*)detail); break;
        case MD_BLOCK_TD:       render_open_td_block(r, "td", (MD_BLOCK_TD_DETAIL*)detail); break;
    }

    return 0;
}

static int
leave_block_callback(MD_BLOCKTYPE type, void* detail, void* userdata)
{
    static const MD_CHAR* head[6] = { "</h1>\n", "</h2>\n", "</h3>\n", "</h4>\n", "</h5>\n", "</h6>\n" };
    MD_HTML* r = (MD_HTML*) userdata;

    switch(type) {
        case MD_BLOCK_DOC:      /*noop*/ break;
        case MD_BLOCK_QUOTE:    RENDER_VERBATIM(r, "</blockquote>\n"); break;
        case MD_BLOCK_UL:       RENDER_VERBATIM(r, "</ul>\n"); break;
        case MD_BLOCK_OL:       RENDER_VERBATIM(r, "</ol>\n"); break;
        case MD_BLOCK_LI:       RENDER_VERBATIM(r, "</li>\n"); break;
        case MD_BLOCK_HR:       /*noop*/ break;
        case MD_BLOCK_H:        RENDER_VERBATIM(r, head[((MD_BLOCK_H_DETAIL*)detail)->level - 1]); break;
        case MD_BLOCK_CODE:     RENDER_VERBATIM(r, "</code></pre>\n"); break;
        case MD_BLOCK_HTML:     /* noop */ break;
        case MD_BLOCK_P:        RENDER_VERBATIM(r, "</p>\n"); break;
        case MD_BLOCK_TABLE:    RENDER_VERBATIM(r, "</table>\n"); break;
        case MD_BLOCK_THEAD:    RENDER_VERBATIM(r, "</thead>\n"); break;
        case MD_BLOCK_TBODY:    RENDER_VERBATIM(r, "</tbody>\n"); break;
        case MD_BLOCK_TR:       RENDER_VERBATIM(r, "</tr>\n"); break;
        case MD_BLOCK_TH:       RENDER_VERBATIM(r, "</th>\n"); break;
        case MD_BLOCK_TD:       RENDER_VERBATIM(r, "</td>\n"); break;
    }

    return 0;
}

static int
enter_span_callback(MD_SPANTYPE type, void* detail, void* userdata)
{
    MD_HTML* r = (MD_HTML*) userdata;
    int inside_img = (r->image_nesting_level > 0);

    /* We are inside a Markdown image label. Markdown allows to use any emphasis
     * and other rich contents in that context similarly as in any link label.
     *
     * However, unlike in the case of links (where that contents becomescontents
     * of the <a>...</a> tag), in the case of images the contents is supposed to
     * fall into the attribute alt: <img alt="...">.
     *
     * In that context we naturally cannot output nested HTML tags. So lets
     * suppress them and only output the plain text (i.e. what falls into text()
     * callback).
     *
     * CommonMark specification declares this a recommended practice for HTML
     * output.
     */
    if(type == MD_SPAN_IMG)
        r->image_nesting_level++;
    if(inside_img)
        return 0;

    switch(type) {
        case MD_SPAN_EM:                RENDER_VERBATIM(r, "<em>"); break;
        case MD_SPAN_STRONG:            RENDER_VERBATIM(r, "<strong>"); break;
        case MD_SPAN_U:                 RENDER_VERBATIM(r, "<u>"); break;
        case MD_SPAN_A:                 render_open_a_span(r, (MD_SPAN_A_DETAIL*) detail); break;
        case MD_SPAN_IMG:               render_open_img_span(r, (MD_SPAN_IMG_DETAIL*) detail); break;
        case MD_SPAN_CODE:              RENDER_VERBATIM(r, "<code>"); break;
        case MD_SPAN_DEL:               RENDER_VERBATIM(r, "<del>"); break;
        case MD_SPAN_LATEXMATH:         RENDER_VERBATIM(r, "<x-equation>"); break;
        case MD_SPAN_LATEXMATH_DISPLAY: RENDER_VERBATIM(r, "<x-equation type=\"display\">"); break;
        case MD_SPAN_WIKILINK:          render_open_wikilink_span(r, (MD_SPAN_WIKILINK_DETAIL*) detail); break;
    }

    return 0;
}

static int
leave_span_callback(MD_SPANTYPE type, void* detail, void* userdata)
{
    MD_HTML* r = (MD_HTML*) userdata;

    if(type == MD_SPAN_IMG)
        r->image_nesting_level--;
    if(r->image_nesting_level > 0)
        return 0;

    switch(type) {
        case MD_SPAN_EM:                RENDER_VERBATIM(r, "</em>"); break;
        case MD_SPAN_STRONG:            RENDER_VERBATIM(r, "</strong>"); break;
        case MD_SPAN_U:                 RENDER_VERBATIM(r, "</u>"); break;
        case MD_SPAN_A:                 RENDER_VERBATIM(r, "</a>"); break;
        case MD_SPAN_IMG:               render_close_img_span(r, (MD_SPAN_IMG_DETAIL*) detail); break;
        case MD_SPAN_CODE:              RENDER_VERBATIM(r, "</code>"); break;
        case MD_SPAN_DEL:               RENDER_VERBATIM(r, "</del>"); break;
        case MD_SPAN_LATEXMATH:         /*fall through*/
        case MD_SPAN_LATEXMATH_DISPLAY: RENDER_VERBATIM(r, "</x-equation>"); break;
        case MD_SPAN_WIKILINK:          RENDER_VERBATIM(r, "</x-wikilink>"); break;
    }

    return 0;
}

static int
text_callback(MD_TEXTTYPE type, const MD_CHAR* text, MD_SIZE size, void* userdata)
{
    MD_HTML* r = (MD_HTML*) userdata;

    switch(type) {
        case MD_TEXT_NULLCHAR:  render_utf8_codepoint(r, 0x0000, render_verbatim); break;
        case MD_TEXT_BR:        RENDER_VERBATIM(r, (r->image_nesting_level == 0
                                        ? ((r->flags & MD_HTML_FLAG_XHTML) ? "<br />\n" : "<br>\n")
                                        : " "));
                                break;
        case MD_TEXT_SOFTBR:    RENDER_VERBATIM(r, (r->image_nesting_level == 0 ? "\n" : " ")); break;
        case MD_TEXT_HTML:      render_verbatim(r, text, size); break;
        case MD_TEXT_ENTITY:    render_entity(r, text, size, render_html_escaped); break;
        default:                render_html_escaped(r, text, size); break;
    }

    return 0;
}

static void
debug_log_callback(const char* msg, void* userdata)
{
    MD_HTML* r = (MD_HTML*) userdata;
    if(r->flags & MD_HTML_FLAG_DEBUG)
        fprintf(stderr, "MD4C: %s\n", msg);
}

int
md_html(const MD_CHAR* input, MD_SIZE input_size,
        void (*process_output)(const MD_CHAR*, MD_SIZE, void*),
        void* userdata, unsigned parser_flags, unsigned renderer_flags)
{
    MD_HTML render = { process_output, userdata, renderer_flags, 0, { 0 } };
    int i;

    MD_PARSER parser = {
        0,
        parser_flags,
        enter_block_callback,
        leave_block_callback,
        enter_span_callback,
        leave_span_callback,
        text_callback,
        debug_log_callback,
        NULL
    };

    /* Build map of characters which need escaping. */
    for(i = 0; i < 256; i++) {
        unsigned char ch = (unsigned char) i;

        if(strchr("\"&<>", ch) != NULL)
            render.escape_map[i] |= NEED_HTML_ESC_FLAG;

        if(!ISALNUM(ch)  &&  strchr("~-_.+!*(),%#@?=;:/,+$", ch) == NULL)
            render.escape_map[i] |= NEED_URL_ESC_FLAG;
    }

    /* Consider skipping UTF-8 byte order mark (BOM). */
    if(renderer_flags & MD_HTML_FLAG_SKIP_UTF8_BOM  &&  sizeof(MD_CHAR) == 1) {
        static const MD_CHAR bom[3] = { (char)0xef, (char)0xbb, (char)0xbf };
        if(input_size >= sizeof(bom)  &&  memcmp(input, bom, sizeof(bom)) == 0) {
            input += sizeof(bom);
            input_size -= sizeof(bom);
        }
    }

    return md_parse(input, input_size, &parser, (void*) &render);
}


Coverage Report

Created: 2026-05-16 09:25

Line	Count	Source
1		/*
2		* MD4C: Markdown parser for C
3		* (http://github.com/mity/md4c)
4		*
5		* Copyright (c) 2016-2024 Martin Mitáš
6		*
7		* Permission is hereby granted, free of charge, to any person obtaining a
8		* copy of this software and associated documentation files (the "Software"),
9		* to deal in the Software without restriction, including without limitation
10		* the rights to use, copy, modify, merge, publish, distribute, sublicense,
11		* and/or sell copies of the Software, and to permit persons to whom the
12		* Software is furnished to do so, subject to the following conditions:
13		*
14		* The above copyright notice and this permission notice shall be included in
15		* all copies or substantial portions of the Software.
16		*
17		* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
18		* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
19		* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
20		* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
21		* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
22		* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
23		* IN THE SOFTWARE.
24		*/
25
26		#include <stdio.h>
27		#include <string.h>
28
29		#include "md4c-html.h"
30		#include "entity.h"
31
32
33		#if !defined(__STDC_VERSION__) \|\| __STDC_VERSION__ < 199409L
34		/* C89/90 or old compilers in general may not understand "inline". */
35		#if defined __GNUC__
36		#define inline __inline__
37		#elif defined _MSC_VER
38		#define inline __inline
39		#else
40		#define inline
41		#endif
42		#endif
43
44		#ifdef _WIN32
45		#define snprintf _snprintf
46		#endif
47
48
49
50		typedef struct MD_HTML_tag MD_HTML;
51		struct MD_HTML_tag {
52		void (process_output)(const MD_CHAR, MD_SIZE, void*);
53		void* userdata;
54		unsigned flags;
55		int image_nesting_level;
56		char escape_map[256];
57		};
58
59	0	#define NEED_HTML_ESC_FLAG 0x1
60	0	#define NEED_URL_ESC_FLAG 0x2
61
62
63		/*****************************************
64		* HTML rendering helper functions *
65		*****************************************/
66
67	0	#define ISDIGIT(ch) ('0' <= (ch) && (ch) <= '9')
68	0	#define ISLOWER(ch) ('a' <= (ch) && (ch) <= 'z')
69	0	#define ISUPPER(ch) ('A' <= (ch) && (ch) <= 'Z')
70	0	#define ISALNUM(ch) (ISLOWER(ch) \|\| ISUPPER(ch) \|\| ISDIGIT(ch))
71
72
73		static inline void
74		render_verbatim(MD_HTML* r, const MD_CHAR* text, MD_SIZE size)
75	0	{
76	0	r->process_output(text, size, r->userdata);
77	0	}
78
79		/* Keep this as a macro. Most compiler should then be smart enough to replace
80		* the strlen() call with a compile-time constant if the string is a C literal. */
81		#define RENDER_VERBATIM(r, verbatim) \
82	0	render_verbatim((r), (verbatim), (MD_SIZE) (strlen(verbatim)))
83
84
85		static void
86		render_html_escaped(MD_HTML* r, const MD_CHAR* data, MD_SIZE size)
87	0	{
88	0	MD_OFFSET beg = 0;
89	0	MD_OFFSET off = 0;
90
91		/* Some characters need to be escaped in normal HTML text. */
92	0	#define NEED_HTML_ESC(ch) (r->escape_map[(unsigned char)(ch)] & NEED_HTML_ESC_FLAG)
93
94	0	while(1) {
95		/* Optimization: Use some loop unrolling. */
96	0	while(off + 3 < size && !NEED_HTML_ESC(data[off+0]) && !NEED_HTML_ESC(data[off+1])
97	0	&& !NEED_HTML_ESC(data[off+2]) && !NEED_HTML_ESC(data[off+3]))
98	0	off += 4;
99	0	while(off < size && !NEED_HTML_ESC(data[off]))
100	0	off++;
101
102	0	if(off > beg)
103	0	render_verbatim(r, data + beg, off - beg);
104
105	0	if(off < size) {
106	0	switch(data[off]) {
107	0	case '&': RENDER_VERBATIM(r, "&"); break;
108	0	case '<': RENDER_VERBATIM(r, "<"); break;
109	0	case '>': RENDER_VERBATIM(r, ">"); break;
110	0	case '"': RENDER_VERBATIM(r, """); break;
111	0	}
112	0	off++;
113	0	} else {
114	0	break;
115	0	}
116	0	beg = off;
117	0	}
118	0	}
119
120		static void
121		render_url_escaped(MD_HTML* r, const MD_CHAR* data, MD_SIZE size)
122	0	{
123	0	static const MD_CHAR hex_chars[] = "0123456789ABCDEF";
124	0	MD_OFFSET beg = 0;
125	0	MD_OFFSET off = 0;
126
127		/* Some characters need to be escaped in URL attributes. */
128	0	#define NEED_URL_ESC(ch) (r->escape_map[(unsigned char)(ch)] & NEED_URL_ESC_FLAG)
129
130	0	while(1) {
131	0	while(off < size && !NEED_URL_ESC(data[off]))
132	0	off++;
133	0	if(off > beg)
134	0	render_verbatim(r, data + beg, off - beg);
135
136	0	if(off < size) {
137	0	char hex[3];
138
139	0	switch(data[off]) {
140	0	case '&': RENDER_VERBATIM(r, "&"); break;
141	0	default:
142	0	hex[0] = '%';
143	0	hex[1] = hex_chars[((unsigned)data[off] >> 4) & 0xf];
144	0	hex[2] = hex_chars[((unsigned)data[off] >> 0) & 0xf];
145	0	render_verbatim(r, hex, 3);
146	0	break;
147	0	}
148	0	off++;
149	0	} else {
150	0	break;
151	0	}
152
153	0	beg = off;
154	0	}
155	0	}
156
157		static unsigned
158		hex_val(char ch)
159	0	{
160	0	if('0' <= ch && ch <= '9')
161	0	return ch - '0';
162	0	if('A' <= ch && ch <= 'Z')
163	0	return ch - 'A' + 10;
164	0	else
165	0	return ch - 'a' + 10;
166	0	}
167
168		static void
169		render_utf8_codepoint(MD_HTML* r, unsigned codepoint,
170		void (fn_append)(MD_HTML, const MD_CHAR*, MD_SIZE))
171	0	{
172	0	static const MD_CHAR utf8_replacement_char[] = { (char)0xef, (char)0xbf, (char)0xbd };
173
174	0	unsigned char utf8[4];
175	0	size_t n;
176
177	0	if(codepoint <= 0x7f) {
178	0	n = 1;
179	0	utf8[0] = codepoint;
180	0	} else if(codepoint <= 0x7ff) {
181	0	n = 2;
182	0	utf8[0] = 0xc0 \| ((codepoint >> 6) & 0x1f);
183	0	utf8[1] = 0x80 + ((codepoint >> 0) & 0x3f);
184	0	} else if(codepoint <= 0xffff) {
185	0	n = 3;
186	0	utf8[0] = 0xe0 \| ((codepoint >> 12) & 0xf);
187	0	utf8[1] = 0x80 + ((codepoint >> 6) & 0x3f);
188	0	utf8[2] = 0x80 + ((codepoint >> 0) & 0x3f);
189	0	} else {
190	0	n = 4;
191	0	utf8[0] = 0xf0 \| ((codepoint >> 18) & 0x7);
192	0	utf8[1] = 0x80 + ((codepoint >> 12) & 0x3f);
193	0	utf8[2] = 0x80 + ((codepoint >> 6) & 0x3f);
194	0	utf8[3] = 0x80 + ((codepoint >> 0) & 0x3f);
195	0	}
196
197	0	if(0 < codepoint && codepoint <= 0x10ffff)
198	0	fn_append(r, (char*)utf8, (MD_SIZE)n);
199	0	else
200	0	fn_append(r, utf8_replacement_char, 3);
201	0	}
202
203		/* Translate entity to its UTF-8 equivalent, or output the verbatim one
204		* if such entity is unknown (or if the translation is disabled). */
205		static void
206		render_entity(MD_HTML* r, const MD_CHAR* text, MD_SIZE size,
207		void (fn_append)(MD_HTML, const MD_CHAR*, MD_SIZE))
208	0	{
209	0	if(r->flags & MD_HTML_FLAG_VERBATIM_ENTITIES) {
210	0	render_verbatim(r, text, size);
211	0	return;
212	0	}
213
214		/* We assume UTF-8 output is what is desired. */
215	0	if(size > 3 && text[1] == '#') {
216	0	unsigned codepoint = 0;
217
218	0	if(text[2] == 'x' \|\| text[2] == 'X') {
219		/* Hexadecimal entity (e.g. "&#x1234abcd;")). */
220	0	MD_SIZE i;
221	0	for(i = 3; i < size-1; i++)
222	0	codepoint = 16 * codepoint + hex_val(text[i]);
223	0	} else {
224		/* Decimal entity (e.g. "&1234;") */
225	0	MD_SIZE i;
226	0	for(i = 2; i < size-1; i++)
227	0	codepoint = 10 * codepoint + (text[i] - '0');
228	0	}
229
230	0	render_utf8_codepoint(r, codepoint, fn_append);
231	0	return;
232	0	} else {
233		/* Named entity (e.g. " "). */
234	0	const ENTITY* ent;
235
236	0	ent = entity_lookup(text, size);
237	0	if(ent != NULL) {
238	0	render_utf8_codepoint(r, ent->codepoints[0], fn_append);
239	0	if(ent->codepoints[1])
240	0	render_utf8_codepoint(r, ent->codepoints[1], fn_append);
241	0	return;
242	0	}
243	0	}
244
245	0	fn_append(r, text, size);
246	0	}
247
248		static void
249		render_attribute(MD_HTML* r, const MD_ATTRIBUTE* attr,
250		void (fn_append)(MD_HTML, const MD_CHAR*, MD_SIZE))
251	0	{
252	0	int i;
253
254	0	for(i = 0; attr->substr_offsets[i] < attr->size; i++) {
255	0	MD_TEXTTYPE type = attr->substr_types[i];
256	0	MD_OFFSET off = attr->substr_offsets[i];
257	0	MD_SIZE size = attr->substr_offsets[i+1] - off;
258	0	const MD_CHAR* text = attr->text + off;
259
260	0	switch(type) {
261	0	case MD_TEXT_NULLCHAR: render_utf8_codepoint(r, 0x0000, render_verbatim); break;
262	0	case MD_TEXT_ENTITY: render_entity(r, text, size, fn_append); break;
263	0	default: fn_append(r, text, size); break;
264	0	}
265	0	}
266	0	}
267
268
269		static void
270		render_open_ol_block(MD_HTML* r, const MD_BLOCK_OL_DETAIL* det)
271	0	{
272	0	char buf[64];
273
274	0	if(det->start == 1) {
275	0	RENDER_VERBATIM(r, "<ol>\n");
276	0	return;
277	0	}
278
279	0	snprintf(buf, sizeof(buf), "<ol start=\"%u\">\n", det->start);
280	0	RENDER_VERBATIM(r, buf);
281	0	}
282
283		static void
284		render_open_li_block(MD_HTML* r, const MD_BLOCK_LI_DETAIL* det)
285	0	{
286	0	if(det->is_task) {
287	0	RENDER_VERBATIM(r, "<li class=\"task-list-item\">"
288	0	"<input type=\"checkbox\" class=\"task-list-item-checkbox\" disabled");
289	0	if(r->flags & MD_HTML_FLAG_XHTML) RENDER_VERBATIM(r, "=\"true\"");
290	0	if(det->task_mark == 'x' \|\| det->task_mark == 'X') {
291	0	RENDER_VERBATIM(r, " checked");
292	0	if(r->flags & MD_HTML_FLAG_XHTML) RENDER_VERBATIM(r, "=\"true\"");
293	0	}
294	0	RENDER_VERBATIM(r, (r->flags & MD_HTML_FLAG_XHTML) ? " />" : ">");
295	0	} else {
296	0	RENDER_VERBATIM(r, "<li>");
297	0	}
298	0	}
299
300		static void
301		render_open_code_block(MD_HTML* r, const MD_BLOCK_CODE_DETAIL* det)
302	0	{
303	0	RENDER_VERBATIM(r, "<pre><code");
304
305		/* If known, output the HTML 5 attribute class="language-LANGNAME". */
306	0	if(det->lang.text != NULL) {
307	0	RENDER_VERBATIM(r, " class=\"");
308	0	if(det->lang.size < 9 \|\| strncmp(det->lang.text, "language-", 9) != 0) {
309	0	RENDER_VERBATIM(r, "language-");
310	0	}
311	0	render_attribute(r, &det->lang, render_html_escaped);
312	0	RENDER_VERBATIM(r, "\"");
313	0	}
314
315	0	RENDER_VERBATIM(r, ">");
316	0	}
317
318		static void
319		render_open_td_block(MD_HTML* r, const MD_CHAR* cell_type, const MD_BLOCK_TD_DETAIL* det)
320	0	{
321	0	RENDER_VERBATIM(r, "<");
322	0	RENDER_VERBATIM(r, cell_type);
323
324	0	switch(det->align) {
325	0	case MD_ALIGN_LEFT: RENDER_VERBATIM(r, " align=\"left\">"); break;
326	0	case MD_ALIGN_CENTER: RENDER_VERBATIM(r, " align=\"center\">"); break;
327	0	case MD_ALIGN_RIGHT: RENDER_VERBATIM(r, " align=\"right\">"); break;
328	0	default: RENDER_VERBATIM(r, ">"); break;
329	0	}
330	0	}
331
332		static void
333		render_open_a_span(MD_HTML* r, const MD_SPAN_A_DETAIL* det)
334	0	{
335	0	RENDER_VERBATIM(r, "<a href=\"");
336	0	render_attribute(r, &det->href, render_url_escaped);
337
338	0	if(det->title.text != NULL) {
339	0	RENDER_VERBATIM(r, "\" title=\"");
340	0	render_attribute(r, &det->title, render_html_escaped);
341	0	}
342
343	0	RENDER_VERBATIM(r, "\">");
344	0	}
345
346		static void
347		render_open_img_span(MD_HTML* r, const MD_SPAN_IMG_DETAIL* det)
348	0	{
349	0	RENDER_VERBATIM(r, "<img src=\"");
350	0	render_attribute(r, &det->src, render_url_escaped);
351
352	0	RENDER_VERBATIM(r, "\" alt=\"");
353	0	}
354
355		static void
356		render_close_img_span(MD_HTML* r, const MD_SPAN_IMG_DETAIL* det)
357	0	{
358	0	if(det->title.text != NULL) {
359	0	RENDER_VERBATIM(r, "\" title=\"");
360	0	render_attribute(r, &det->title, render_html_escaped);
361	0	}
362
363	0	RENDER_VERBATIM(r, (r->flags & MD_HTML_FLAG_XHTML) ? "\" />" : "\">");
364	0	}
365
366		static void
367		render_open_wikilink_span(MD_HTML* r, const MD_SPAN_WIKILINK_DETAIL* det)
368	0	{
369	0	RENDER_VERBATIM(r, "<x-wikilink data-target=\"");
370	0	render_attribute(r, &det->target, render_html_escaped);
371
372	0	RENDER_VERBATIM(r, "\">");
373	0	}
374
375
376		/**************************************
377		* HTML renderer implementation *
378		**************************************/
379
380		static int
381		enter_block_callback(MD_BLOCKTYPE type, void* detail, void* userdata)
382	0	{
383	0	static const MD_CHAR* head[6] = { "<h1>", "<h2>", "<h3>", "<h4>", "<h5>", "<h6>" };
384	0	MD_HTML* r = (MD_HTML*) userdata;
385
386	0	switch(type) {
387	0	case MD_BLOCK_DOC: /* noop */ break;
388	0	case MD_BLOCK_QUOTE: RENDER_VERBATIM(r, "<blockquote>\n"); break;
389	0	case MD_BLOCK_UL: RENDER_VERBATIM(r, "<ul>\n"); break;
390	0	case MD_BLOCK_OL: render_open_ol_block(r, (const MD_BLOCK_OL_DETAIL*)detail); break;
391	0	case MD_BLOCK_LI: render_open_li_block(r, (const MD_BLOCK_LI_DETAIL*)detail); break;
392	0	case MD_BLOCK_HR: RENDER_VERBATIM(r, (r->flags & MD_HTML_FLAG_XHTML) ? "<hr />\n" : "<hr>\n"); break;
393	0	case MD_BLOCK_H: RENDER_VERBATIM(r, head[((MD_BLOCK_H_DETAIL*)detail)->level - 1]); break;
394	0	case MD_BLOCK_CODE: render_open_code_block(r, (const MD_BLOCK_CODE_DETAIL*) detail); break;
395	0	case MD_BLOCK_HTML: /* noop */ break;
396	0	case MD_BLOCK_P: RENDER_VERBATIM(r, "<p>"); break;
397	0	case MD_BLOCK_TABLE: RENDER_VERBATIM(r, "<table>\n"); break;
398	0	case MD_BLOCK_THEAD: RENDER_VERBATIM(r, "<thead>\n"); break;
399	0	case MD_BLOCK_TBODY: RENDER_VERBATIM(r, "<tbody>\n"); break;
400	0	case MD_BLOCK_TR: RENDER_VERBATIM(r, "<tr>\n"); break;
401	0	case MD_BLOCK_TH: render_open_td_block(r, "th", (MD_BLOCK_TD_DETAIL*)detail); break;
402	0	case MD_BLOCK_TD: render_open_td_block(r, "td", (MD_BLOCK_TD_DETAIL*)detail); break;
403	0	}
404
405	0	return 0;
406	0	}
407
408		static int
409		leave_block_callback(MD_BLOCKTYPE type, void* detail, void* userdata)
410	0	{
411	0	static const MD_CHAR* head[6] = { "</h1>\n", "</h2>\n", "</h3>\n", "</h4>\n", "</h5>\n", "</h6>\n" };
412	0	MD_HTML* r = (MD_HTML*) userdata;
413
414	0	switch(type) {
415	0	case MD_BLOCK_DOC: /noop/ break;
416	0	case MD_BLOCK_QUOTE: RENDER_VERBATIM(r, "</blockquote>\n"); break;
417	0	case MD_BLOCK_UL: RENDER_VERBATIM(r, "</ul>\n"); break;
418	0	case MD_BLOCK_OL: RENDER_VERBATIM(r, "</ol>\n"); break;
419	0	case MD_BLOCK_LI: RENDER_VERBATIM(r, "</li>\n"); break;
420	0	case MD_BLOCK_HR: /noop/ break;
421	0	case MD_BLOCK_H: RENDER_VERBATIM(r, head[((MD_BLOCK_H_DETAIL*)detail)->level - 1]); break;
422	0	case MD_BLOCK_CODE: RENDER_VERBATIM(r, "</code></pre>\n"); break;
423	0	case MD_BLOCK_HTML: /* noop */ break;
424	0	case MD_BLOCK_P: RENDER_VERBATIM(r, "</p>\n"); break;
425	0	case MD_BLOCK_TABLE: RENDER_VERBATIM(r, "</table>\n"); break;
426	0	case MD_BLOCK_THEAD: RENDER_VERBATIM(r, "</thead>\n"); break;
427	0	case MD_BLOCK_TBODY: RENDER_VERBATIM(r, "</tbody>\n"); break;
428	0	case MD_BLOCK_TR: RENDER_VERBATIM(r, "</tr>\n"); break;
429	0	case MD_BLOCK_TH: RENDER_VERBATIM(r, "</th>\n"); break;
430	0	case MD_BLOCK_TD: RENDER_VERBATIM(r, "</td>\n"); break;
431	0	}
432
433	0	return 0;
434	0	}
435
436		static int
437		enter_span_callback(MD_SPANTYPE type, void* detail, void* userdata)
438	0	{
439	0	MD_HTML* r = (MD_HTML*) userdata;
440	0	int inside_img = (r->image_nesting_level > 0);
441
442		/* We are inside a Markdown image label. Markdown allows to use any emphasis
443		* and other rich contents in that context similarly as in any link label.
444		*
445		* However, unlike in the case of links (where that contents becomescontents
446		* of the <a>...</a> tag), in the case of images the contents is supposed to
447		* fall into the attribute alt: <img alt="...">.
448		*
449		* In that context we naturally cannot output nested HTML tags. So lets
450		* suppress them and only output the plain text (i.e. what falls into text()
451		* callback).
452		*
453		* CommonMark specification declares this a recommended practice for HTML
454		* output.
455		*/
456	0	if(type == MD_SPAN_IMG)
457	0	r->image_nesting_level++;
458	0	if(inside_img)
459	0	return 0;
460
461	0	switch(type) {
462	0	case MD_SPAN_EM: RENDER_VERBATIM(r, "<em>"); break;
463	0	case MD_SPAN_STRONG: RENDER_VERBATIM(r, "<strong>"); break;
464	0	case MD_SPAN_U: RENDER_VERBATIM(r, "<u>"); break;
465	0	case MD_SPAN_A: render_open_a_span(r, (MD_SPAN_A_DETAIL*) detail); break;
466	0	case MD_SPAN_IMG: render_open_img_span(r, (MD_SPAN_IMG_DETAIL*) detail); break;
467	0	case MD_SPAN_CODE: RENDER_VERBATIM(r, "<code>"); break;
468	0	case MD_SPAN_DEL: RENDER_VERBATIM(r, "<del>"); break;
469	0	case MD_SPAN_LATEXMATH: RENDER_VERBATIM(r, "<x-equation>"); break;
470	0	case MD_SPAN_LATEXMATH_DISPLAY: RENDER_VERBATIM(r, "<x-equation type=\"display\">"); break;
471	0	case MD_SPAN_WIKILINK: render_open_wikilink_span(r, (MD_SPAN_WIKILINK_DETAIL*) detail); break;
472	0	}
473
474	0	return 0;
475	0	}
476
477		static int
478		leave_span_callback(MD_SPANTYPE type, void* detail, void* userdata)
479	0	{
480	0	MD_HTML* r = (MD_HTML*) userdata;
481
482	0	if(type == MD_SPAN_IMG)
483	0	r->image_nesting_level--;
484	0	if(r->image_nesting_level > 0)
485	0	return 0;
486
487	0	switch(type) {
488	0	case MD_SPAN_EM: RENDER_VERBATIM(r, "</em>"); break;
489	0	case MD_SPAN_STRONG: RENDER_VERBATIM(r, "</strong>"); break;
490	0	case MD_SPAN_U: RENDER_VERBATIM(r, "</u>"); break;
491	0	case MD_SPAN_A: RENDER_VERBATIM(r, "</a>"); break;
492	0	case MD_SPAN_IMG: render_close_img_span(r, (MD_SPAN_IMG_DETAIL*) detail); break;
493	0	case MD_SPAN_CODE: RENDER_VERBATIM(r, "</code>"); break;
494	0	case MD_SPAN_DEL: RENDER_VERBATIM(r, "</del>"); break;
495	0	case MD_SPAN_LATEXMATH: /fall through/
496	0	case MD_SPAN_LATEXMATH_DISPLAY: RENDER_VERBATIM(r, "</x-equation>"); break;
497	0	case MD_SPAN_WIKILINK: RENDER_VERBATIM(r, "</x-wikilink>"); break;
498	0	}
499
500	0	return 0;
501	0	}
502
503		static int
504		text_callback(MD_TEXTTYPE type, const MD_CHAR* text, MD_SIZE size, void* userdata)
505	0	{
506	0	MD_HTML* r = (MD_HTML*) userdata;
507
508	0	switch(type) {
509	0	case MD_TEXT_NULLCHAR: render_utf8_codepoint(r, 0x0000, render_verbatim); break;
510	0	case MD_TEXT_BR: RENDER_VERBATIM(r, (r->image_nesting_level == 0
511	0	? ((r->flags & MD_HTML_FLAG_XHTML) ? "<br />\n" : "<br>\n")
512	0	: " "));
513	0	break;
514	0	case MD_TEXT_SOFTBR: RENDER_VERBATIM(r, (r->image_nesting_level == 0 ? "\n" : " ")); break;
515	0	case MD_TEXT_HTML: render_verbatim(r, text, size); break;
516	0	case MD_TEXT_ENTITY: render_entity(r, text, size, render_html_escaped); break;
517	0	default: render_html_escaped(r, text, size); break;
518	0	}
519
520	0	return 0;
521	0	}
522
523		static void
524		debug_log_callback(const char* msg, void* userdata)
525	0	{
526	0	MD_HTML* r = (MD_HTML*) userdata;
527	0	if(r->flags & MD_HTML_FLAG_DEBUG)
528	0	fprintf(stderr, "MD4C: %s\n", msg);
529	0	}
530
531		int
532		md_html(const MD_CHAR* input, MD_SIZE input_size,
533		void (process_output)(const MD_CHAR, MD_SIZE, void*),
534		void* userdata, unsigned parser_flags, unsigned renderer_flags)
535	0	{
536	0	MD_HTML render = { process_output, userdata, renderer_flags, 0, { 0 } };
537	0	int i;
538
539	0	MD_PARSER parser = {
540	0	0,
541	0	parser_flags,
542	0	enter_block_callback,
543	0	leave_block_callback,
544	0	enter_span_callback,
545	0	leave_span_callback,
546	0	text_callback,
547	0	debug_log_callback,
548	0	NULL
549	0	};
550
551		/* Build map of characters which need escaping. */
552	0	for(i = 0; i < 256; i++) {
553	0	unsigned char ch = (unsigned char) i;
554
555	0	if(strchr("\"&<>", ch) != NULL)
556	0	render.escape_map[i] \|= NEED_HTML_ESC_FLAG;
557
558	0	if(!ISALNUM(ch) && strchr("~-_.+!*(),%#@?=;:/,+$", ch) == NULL)
559	0	render.escape_map[i] \|= NEED_URL_ESC_FLAG;
560	0	}
561
562		/* Consider skipping UTF-8 byte order mark (BOM). */
563	0	if(renderer_flags & MD_HTML_FLAG_SKIP_UTF8_BOM && sizeof(MD_CHAR) == 1) {
564	0	static const MD_CHAR bom[3] = { (char)0xef, (char)0xbb, (char)0xbf };
565	0	if(input_size >= sizeof(bom) && memcmp(input, bom, sizeof(bom)) == 0) {
566	0	input += sizeof(bom);
567	0	input_size -= sizeof(bom);
568	0	}
569	0	}
570
571	0	return md_parse(input, input_size, &parser, (void*) &render);
572	0	}
573