From b6d455a02bd338e9dc0faa09d4d8177ecd8d569a Mon Sep 17 00:00:00 2001 From: Petr Mrázek Date: Sun, 10 Apr 2016 15:53:05 +0200 Subject: NOISSUE reorganize and document libraries --- libraries/hoedown/src/autolink.c | 281 +++ libraries/hoedown/src/buffer.c | 308 ++++ libraries/hoedown/src/document.c | 2958 ++++++++++++++++++++++++++++++ libraries/hoedown/src/escape.c | 188 ++ libraries/hoedown/src/html.c | 754 ++++++++ libraries/hoedown/src/html_blocks.c | 240 +++ libraries/hoedown/src/html_smartypants.c | 435 +++++ libraries/hoedown/src/stack.c | 79 + libraries/hoedown/src/version.c | 9 + 9 files changed, 5252 insertions(+) create mode 100644 libraries/hoedown/src/autolink.c create mode 100644 libraries/hoedown/src/buffer.c create mode 100644 libraries/hoedown/src/document.c create mode 100644 libraries/hoedown/src/escape.c create mode 100644 libraries/hoedown/src/html.c create mode 100644 libraries/hoedown/src/html_blocks.c create mode 100644 libraries/hoedown/src/html_smartypants.c create mode 100644 libraries/hoedown/src/stack.c create mode 100644 libraries/hoedown/src/version.c (limited to 'libraries/hoedown/src') diff --git a/libraries/hoedown/src/autolink.c b/libraries/hoedown/src/autolink.c new file mode 100644 index 00000000..9bc7fad5 --- /dev/null +++ b/libraries/hoedown/src/autolink.c @@ -0,0 +1,281 @@ +#include "hoedown/autolink.h" + +#include +#include +#include +#include + +#ifndef _MSC_VER +#include +#else +#define strncasecmp _strnicmp +#endif + +int +hoedown_autolink_is_safe(const uint8_t *data, size_t size) +{ + static const size_t valid_uris_count = 6; + static const char *valid_uris[] = { + "http://", "https://", "/", "#", "ftp://", "mailto:" + }; + static const size_t valid_uris_size[] = { 7, 8, 1, 1, 6, 7 }; + size_t i; + + for (i = 0; i < valid_uris_count; ++i) { + size_t len = valid_uris_size[i]; + + if (size > len && + strncasecmp((char *)data, valid_uris[i], len) == 0 && + isalnum(data[len])) + return 1; + } + + return 0; +} + +static size_t +autolink_delim(uint8_t *data, size_t link_end, size_t max_rewind, size_t size) +{ + uint8_t cclose, copen = 0; + size_t i; + + for (i = 0; i < link_end; ++i) + if (data[i] == '<') { + link_end = i; + break; + } + + while (link_end > 0) { + if (strchr("?!.,:", data[link_end - 1]) != NULL) + link_end--; + + else if (data[link_end - 1] == ';') { + size_t new_end = link_end - 2; + + while (new_end > 0 && isalpha(data[new_end])) + new_end--; + + if (new_end < link_end - 2 && data[new_end] == '&') + link_end = new_end; + else + link_end--; + } + else break; + } + + if (link_end == 0) + return 0; + + cclose = data[link_end - 1]; + + switch (cclose) { + case '"': copen = '"'; break; + case '\'': copen = '\''; break; + case ')': copen = '('; break; + case ']': copen = '['; break; + case '}': copen = '{'; break; + } + + if (copen != 0) { + size_t closing = 0; + size_t opening = 0; + size_t i = 0; + + /* Try to close the final punctuation sign in this same line; + * if we managed to close it outside of the URL, that means that it's + * not part of the URL. If it closes inside the URL, that means it + * is part of the URL. + * + * Examples: + * + * foo http://www.pokemon.com/Pikachu_(Electric) bar + * => http://www.pokemon.com/Pikachu_(Electric) + * + * foo (http://www.pokemon.com/Pikachu_(Electric)) bar + * => http://www.pokemon.com/Pikachu_(Electric) + * + * foo http://www.pokemon.com/Pikachu_(Electric)) bar + * => http://www.pokemon.com/Pikachu_(Electric)) + * + * (foo http://www.pokemon.com/Pikachu_(Electric)) bar + * => foo http://www.pokemon.com/Pikachu_(Electric) + */ + + while (i < link_end) { + if (data[i] == copen) + opening++; + else if (data[i] == cclose) + closing++; + + i++; + } + + if (closing != opening) + link_end--; + } + + return link_end; +} + +static size_t +check_domain(uint8_t *data, size_t size, int allow_short) +{ + size_t i, np = 0; + + if (!isalnum(data[0])) + return 0; + + for (i = 1; i < size - 1; ++i) { + if (strchr(".:", data[i]) != NULL) np++; + else if (!isalnum(data[i]) && data[i] != '-') break; + } + + if (allow_short) { + /* We don't need a valid domain in the strict sense (with + * least one dot; so just make sure it's composed of valid + * domain characters and return the length of the the valid + * sequence. */ + return i; + } else { + /* a valid domain needs to have at least a dot. + * that's as far as we get */ + return np ? i : 0; + } +} + +size_t +hoedown_autolink__www( + size_t *rewind_p, + hoedown_buffer *link, + uint8_t *data, + size_t max_rewind, + size_t size, + unsigned int flags) +{ + size_t link_end; + + if (max_rewind > 0 && !ispunct(data[-1]) && !isspace(data[-1])) + return 0; + + if (size < 4 || memcmp(data, "www.", strlen("www.")) != 0) + return 0; + + link_end = check_domain(data, size, 0); + + if (link_end == 0) + return 0; + + while (link_end < size && !isspace(data[link_end])) + link_end++; + + link_end = autolink_delim(data, link_end, max_rewind, size); + + if (link_end == 0) + return 0; + + hoedown_buffer_put(link, data, link_end); + *rewind_p = 0; + + return (int)link_end; +} + +size_t +hoedown_autolink__email( + size_t *rewind_p, + hoedown_buffer *link, + uint8_t *data, + size_t max_rewind, + size_t size, + unsigned int flags) +{ + size_t link_end, rewind; + int nb = 0, np = 0; + + for (rewind = 0; rewind < max_rewind; ++rewind) { + uint8_t c = data[-1 - rewind]; + + if (isalnum(c)) + continue; + + if (strchr(".+-_", c) != NULL) + continue; + + break; + } + + if (rewind == 0) + return 0; + + for (link_end = 0; link_end < size; ++link_end) { + uint8_t c = data[link_end]; + + if (isalnum(c)) + continue; + + if (c == '@') + nb++; + else if (c == '.' && link_end < size - 1) + np++; + else if (c != '-' && c != '_') + break; + } + + if (link_end < 2 || nb != 1 || np == 0 || + !isalpha(data[link_end - 1])) + return 0; + + link_end = autolink_delim(data, link_end, max_rewind, size); + + if (link_end == 0) + return 0; + + hoedown_buffer_put(link, data - rewind, link_end + rewind); + *rewind_p = rewind; + + return link_end; +} + +size_t +hoedown_autolink__url( + size_t *rewind_p, + hoedown_buffer *link, + uint8_t *data, + size_t max_rewind, + size_t size, + unsigned int flags) +{ + size_t link_end, rewind = 0, domain_len; + + if (size < 4 || data[1] != '/' || data[2] != '/') + return 0; + + while (rewind < max_rewind && isalpha(data[-1 - rewind])) + rewind++; + + if (!hoedown_autolink_is_safe(data - rewind, size + rewind)) + return 0; + + link_end = strlen("://"); + + domain_len = check_domain( + data + link_end, + size - link_end, + flags & HOEDOWN_AUTOLINK_SHORT_DOMAINS); + + if (domain_len == 0) + return 0; + + link_end += domain_len; + while (link_end < size && !isspace(data[link_end])) + link_end++; + + link_end = autolink_delim(data, link_end, max_rewind, size); + + if (link_end == 0) + return 0; + + hoedown_buffer_put(link, data - rewind, link_end + rewind); + *rewind_p = rewind; + + return link_end; +} diff --git a/libraries/hoedown/src/buffer.c b/libraries/hoedown/src/buffer.c new file mode 100644 index 00000000..1c7ba55a --- /dev/null +++ b/libraries/hoedown/src/buffer.c @@ -0,0 +1,308 @@ +#include "hoedown/buffer.h" + +#include +#include +#include +#include + +void * +hoedown_malloc(size_t size) +{ + void *ret = malloc(size); + + if (!ret) { + fprintf(stderr, "Allocation failed.\n"); + abort(); + } + + return ret; +} + +void * +hoedown_calloc(size_t nmemb, size_t size) +{ + void *ret = calloc(nmemb, size); + + if (!ret) { + fprintf(stderr, "Allocation failed.\n"); + abort(); + } + + return ret; +} + +void * +hoedown_realloc(void *ptr, size_t size) +{ + void *ret = realloc(ptr, size); + + if (!ret) { + fprintf(stderr, "Allocation failed.\n"); + abort(); + } + + return ret; +} + +void +hoedown_buffer_init( + hoedown_buffer *buf, + size_t unit, + hoedown_realloc_callback data_realloc, + hoedown_free_callback data_free, + hoedown_free_callback buffer_free) +{ + assert(buf); + + buf->data = NULL; + buf->size = buf->asize = 0; + buf->unit = unit; + buf->data_realloc = data_realloc; + buf->data_free = data_free; + buf->buffer_free = buffer_free; +} + +void +hoedown_buffer_uninit(hoedown_buffer *buf) +{ + assert(buf && buf->unit); + buf->data_free(buf->data); +} + +hoedown_buffer * +hoedown_buffer_new(size_t unit) +{ + hoedown_buffer *ret = hoedown_malloc(sizeof (hoedown_buffer)); + hoedown_buffer_init(ret, unit, hoedown_realloc, free, free); + return ret; +} + +void +hoedown_buffer_free(hoedown_buffer *buf) +{ + if (!buf) return; + assert(buf && buf->unit); + + buf->data_free(buf->data); + + if (buf->buffer_free) + buf->buffer_free(buf); +} + +void +hoedown_buffer_reset(hoedown_buffer *buf) +{ + assert(buf && buf->unit); + + buf->data_free(buf->data); + buf->data = NULL; + buf->size = buf->asize = 0; +} + +void +hoedown_buffer_grow(hoedown_buffer *buf, size_t neosz) +{ + size_t neoasz; + assert(buf && buf->unit); + + if (buf->asize >= neosz) + return; + + neoasz = buf->asize + buf->unit; + while (neoasz < neosz) + neoasz += buf->unit; + + buf->data = (uint8_t *) buf->data_realloc(buf->data, neoasz); + buf->asize = neoasz; +} + +void +hoedown_buffer_put(hoedown_buffer *buf, const uint8_t *data, size_t size) +{ + assert(buf && buf->unit); + + if (buf->size + size > buf->asize) + hoedown_buffer_grow(buf, buf->size + size); + + memcpy(buf->data + buf->size, data, size); + buf->size += size; +} + +void +hoedown_buffer_puts(hoedown_buffer *buf, const char *str) +{ + hoedown_buffer_put(buf, (const uint8_t *)str, strlen(str)); +} + +void +hoedown_buffer_putc(hoedown_buffer *buf, uint8_t c) +{ + assert(buf && buf->unit); + + if (buf->size >= buf->asize) + hoedown_buffer_grow(buf, buf->size + 1); + + buf->data[buf->size] = c; + buf->size += 1; +} + +int +hoedown_buffer_putf(hoedown_buffer *buf, FILE *file) +{ + assert(buf && buf->unit); + + while (!(feof(file) || ferror(file))) { + hoedown_buffer_grow(buf, buf->size + buf->unit); + buf->size += fread(buf->data + buf->size, 1, buf->unit, file); + } + + return ferror(file); +} + +void +hoedown_buffer_set(hoedown_buffer *buf, const uint8_t *data, size_t size) +{ + assert(buf && buf->unit); + + if (size > buf->asize) + hoedown_buffer_grow(buf, size); + + memcpy(buf->data, data, size); + buf->size = size; +} + +void +hoedown_buffer_sets(hoedown_buffer *buf, const char *str) +{ + hoedown_buffer_set(buf, (const uint8_t *)str, strlen(str)); +} + +int +hoedown_buffer_eq(const hoedown_buffer *buf, const uint8_t *data, size_t size) +{ + if (buf->size != size) return 0; + return memcmp(buf->data, data, size) == 0; +} + +int +hoedown_buffer_eqs(const hoedown_buffer *buf, const char *str) +{ + return hoedown_buffer_eq(buf, (const uint8_t *)str, strlen(str)); +} + +int +hoedown_buffer_prefix(const hoedown_buffer *buf, const char *prefix) +{ + size_t i; + + for (i = 0; i < buf->size; ++i) { + if (prefix[i] == 0) + return 0; + + if (buf->data[i] != prefix[i]) + return buf->data[i] - prefix[i]; + } + + return 0; +} + +void +hoedown_buffer_slurp(hoedown_buffer *buf, size_t size) +{ + assert(buf && buf->unit); + + if (size >= buf->size) { + buf->size = 0; + return; + } + + buf->size -= size; + memmove(buf->data, buf->data + size, buf->size); +} + +const char * +hoedown_buffer_cstr(hoedown_buffer *buf) +{ + assert(buf && buf->unit); + + if (buf->size < buf->asize && buf->data[buf->size] == 0) + return (char *)buf->data; + + hoedown_buffer_grow(buf, buf->size + 1); + buf->data[buf->size] = 0; + + return (char *)buf->data; +} + +void +hoedown_buffer_printf(hoedown_buffer *buf, const char *fmt, ...) +{ + va_list ap; + int n; + + assert(buf && buf->unit); + + if (buf->size >= buf->asize) + hoedown_buffer_grow(buf, buf->size + 1); + + va_start(ap, fmt); + n = vsnprintf((char *)buf->data + buf->size, buf->asize - buf->size, fmt, ap); + va_end(ap); + + if (n < 0) { +#ifndef _MSC_VER + return; +#else + va_start(ap, fmt); + n = _vscprintf(fmt, ap); + va_end(ap); +#endif + } + + if ((size_t)n >= buf->asize - buf->size) { + hoedown_buffer_grow(buf, buf->size + n + 1); + + va_start(ap, fmt); + n = vsnprintf((char *)buf->data + buf->size, buf->asize - buf->size, fmt, ap); + va_end(ap); + } + + if (n < 0) + return; + + buf->size += n; +} + +void hoedown_buffer_put_utf8(hoedown_buffer *buf, unsigned int c) { + unsigned char unichar[4]; + + assert(buf && buf->unit); + + if (c < 0x80) { + hoedown_buffer_putc(buf, c); + } + else if (c < 0x800) { + unichar[0] = 192 + (c / 64); + unichar[1] = 128 + (c % 64); + hoedown_buffer_put(buf, unichar, 2); + } + else if (c - 0xd800u < 0x800) { + HOEDOWN_BUFPUTSL(buf, "\xef\xbf\xbd"); + } + else if (c < 0x10000) { + unichar[0] = 224 + (c / 4096); + unichar[1] = 128 + (c / 64) % 64; + unichar[2] = 128 + (c % 64); + hoedown_buffer_put(buf, unichar, 3); + } + else if (c < 0x110000) { + unichar[0] = 240 + (c / 262144); + unichar[1] = 128 + (c / 4096) % 64; + unichar[2] = 128 + (c / 64) % 64; + unichar[3] = 128 + (c % 64); + hoedown_buffer_put(buf, unichar, 4); + } + else { + HOEDOWN_BUFPUTSL(buf, "\xef\xbf\xbd"); + } +} diff --git a/libraries/hoedown/src/document.c b/libraries/hoedown/src/document.c new file mode 100644 index 00000000..8ba82e47 --- /dev/null +++ b/libraries/hoedown/src/document.c @@ -0,0 +1,2958 @@ +#include "hoedown/document.h" + +#include +#include +#include +#include + +#include "hoedown/stack.h" + +#ifndef _MSC_VER +#include +#else +#define strncasecmp _strnicmp +#endif + +#define REF_TABLE_SIZE 8 + +#define BUFFER_BLOCK 0 +#define BUFFER_SPAN 1 + +#define HOEDOWN_LI_END 8 /* internal list flag */ + +const char *hoedown_find_block_tag(const char *str, unsigned int len); + +/*************** + * LOCAL TYPES * + ***************/ + +/* link_ref: reference to a link */ +struct link_ref { + unsigned int id; + + hoedown_buffer *link; + hoedown_buffer *title; + + struct link_ref *next; +}; + +/* footnote_ref: reference to a footnote */ +struct footnote_ref { + unsigned int id; + + int is_used; + unsigned int num; + + hoedown_buffer *contents; +}; + +/* footnote_item: an item in a footnote_list */ +struct footnote_item { + struct footnote_ref *ref; + struct footnote_item *next; +}; + +/* footnote_list: linked list of footnote_item */ +struct footnote_list { + unsigned int count; + struct footnote_item *head; + struct footnote_item *tail; +}; + +/* char_trigger: function pointer to render active chars */ +/* returns the number of chars taken care of */ +/* data is the pointer of the beginning of the span */ +/* offset is the number of valid chars before data */ +typedef size_t +(*char_trigger)(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); + +static size_t char_emphasis(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_quote(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_linebreak(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_codespan(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_escape(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_entity(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_langle_tag(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_autolink_url(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_autolink_email(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_autolink_www(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_link(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_superscript(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); +static size_t char_math(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size); + +enum markdown_char_t { + MD_CHAR_NONE = 0, + MD_CHAR_EMPHASIS, + MD_CHAR_CODESPAN, + MD_CHAR_LINEBREAK, + MD_CHAR_LINK, + MD_CHAR_LANGLE, + MD_CHAR_ESCAPE, + MD_CHAR_ENTITY, + MD_CHAR_AUTOLINK_URL, + MD_CHAR_AUTOLINK_EMAIL, + MD_CHAR_AUTOLINK_WWW, + MD_CHAR_SUPERSCRIPT, + MD_CHAR_QUOTE, + MD_CHAR_MATH +}; + +static char_trigger markdown_char_ptrs[] = { + NULL, + &char_emphasis, + &char_codespan, + &char_linebreak, + &char_link, + &char_langle_tag, + &char_escape, + &char_entity, + &char_autolink_url, + &char_autolink_email, + &char_autolink_www, + &char_superscript, + &char_quote, + &char_math +}; + +struct hoedown_document { + hoedown_renderer md; + hoedown_renderer_data data; + + struct link_ref *refs[REF_TABLE_SIZE]; + struct footnote_list footnotes_found; + struct footnote_list footnotes_used; + uint8_t active_char[256]; + hoedown_stack work_bufs[2]; + hoedown_extensions ext_flags; + size_t max_nesting; + int in_link_body; +}; + +/*************************** + * HELPER FUNCTIONS * + ***************************/ + +static hoedown_buffer * +newbuf(hoedown_document *doc, int type) +{ + static const size_t buf_size[2] = {256, 64}; + hoedown_buffer *work = NULL; + hoedown_stack *pool = &doc->work_bufs[type]; + + if (pool->size < pool->asize && + pool->item[pool->size] != NULL) { + work = pool->item[pool->size++]; + work->size = 0; + } else { + work = hoedown_buffer_new(buf_size[type]); + hoedown_stack_push(pool, work); + } + + return work; +} + +static void +popbuf(hoedown_document *doc, int type) +{ + doc->work_bufs[type].size--; +} + +static void +unscape_text(hoedown_buffer *ob, hoedown_buffer *src) +{ + size_t i = 0, org; + while (i < src->size) { + org = i; + while (i < src->size && src->data[i] != '\\') + i++; + + if (i > org) + hoedown_buffer_put(ob, src->data + org, i - org); + + if (i + 1 >= src->size) + break; + + hoedown_buffer_putc(ob, src->data[i + 1]); + i += 2; + } +} + +static unsigned int +hash_link_ref(const uint8_t *link_ref, size_t length) +{ + size_t i; + unsigned int hash = 0; + + for (i = 0; i < length; ++i) + hash = tolower(link_ref[i]) + (hash << 6) + (hash << 16) - hash; + + return hash; +} + +static struct link_ref * +add_link_ref( + struct link_ref **references, + const uint8_t *name, size_t name_size) +{ + struct link_ref *ref = hoedown_calloc(1, sizeof(struct link_ref)); + + ref->id = hash_link_ref(name, name_size); + ref->next = references[ref->id % REF_TABLE_SIZE]; + + references[ref->id % REF_TABLE_SIZE] = ref; + return ref; +} + +static struct link_ref * +find_link_ref(struct link_ref **references, uint8_t *name, size_t length) +{ + unsigned int hash = hash_link_ref(name, length); + struct link_ref *ref = NULL; + + ref = references[hash % REF_TABLE_SIZE]; + + while (ref != NULL) { + if (ref->id == hash) + return ref; + + ref = ref->next; + } + + return NULL; +} + +static void +free_link_refs(struct link_ref **references) +{ + size_t i; + + for (i = 0; i < REF_TABLE_SIZE; ++i) { + struct link_ref *r = references[i]; + struct link_ref *next; + + while (r) { + next = r->next; + hoedown_buffer_free(r->link); + hoedown_buffer_free(r->title); + free(r); + r = next; + } + } +} + +static struct footnote_ref * +create_footnote_ref(struct footnote_list *list, const uint8_t *name, size_t name_size) +{ + struct footnote_ref *ref = hoedown_calloc(1, sizeof(struct footnote_ref)); + + ref->id = hash_link_ref(name, name_size); + + return ref; +} + +static int +add_footnote_ref(struct footnote_list *list, struct footnote_ref *ref) +{ + struct footnote_item *item = hoedown_calloc(1, sizeof(struct footnote_item)); + if (!item) + return 0; + item->ref = ref; + + if (list->head == NULL) { + list->head = list->tail = item; + } else { + list->tail->next = item; + list->tail = item; + } + list->count++; + + return 1; +} + +static struct footnote_ref * +find_footnote_ref(struct footnote_list *list, uint8_t *name, size_t length) +{ + unsigned int hash = hash_link_ref(name, length); + struct footnote_item *item = NULL; + + item = list->head; + + while (item != NULL) { + if (item->ref->id == hash) + return item->ref; + item = item->next; + } + + return NULL; +} + +static void +free_footnote_ref(struct footnote_ref *ref) +{ + hoedown_buffer_free(ref->contents); + free(ref); +} + +static void +free_footnote_list(struct footnote_list *list, int free_refs) +{ + struct footnote_item *item = list->head; + struct footnote_item *next; + + while (item) { + next = item->next; + if (free_refs) + free_footnote_ref(item->ref); + free(item); + item = next; + } +} + + +/* + * Check whether a char is a Markdown spacing char. + + * Right now we only consider spaces the actual + * space and a newline: tabs and carriage returns + * are filtered out during the preprocessing phase. + * + * If we wanted to actually be UTF-8 compliant, we + * should instead extract an Unicode codepoint from + * this character and check for space properties. + */ +static int +_isspace(int c) +{ + return c == ' ' || c == '\n'; +} + +/* is_empty_all: verify that all the data is spacing */ +static int +is_empty_all(const uint8_t *data, size_t size) +{ + size_t i = 0; + while (i < size && _isspace(data[i])) i++; + return i == size; +} + +/* + * Replace all spacing characters in data with spaces. As a special + * case, this collapses a newline with the previous space, if possible. + */ +static void +replace_spacing(hoedown_buffer *ob, const uint8_t *data, size_t size) +{ + size_t i = 0, mark; + hoedown_buffer_grow(ob, size); + while (1) { + mark = i; + while (i < size && data[i] != '\n') i++; + hoedown_buffer_put(ob, data + mark, i - mark); + + if (i >= size) break; + + if (!(i > 0 && data[i-1] == ' ')) + hoedown_buffer_putc(ob, ' '); + i++; + } +} + +/**************************** + * INLINE PARSING FUNCTIONS * + ****************************/ + +/* is_mail_autolink • looks for the address part of a mail autolink and '>' */ +/* this is less strict than the original markdown e-mail address matching */ +static size_t +is_mail_autolink(uint8_t *data, size_t size) +{ + size_t i = 0, nb = 0; + + /* address is assumed to be: [-@._a-zA-Z0-9]+ with exactly one '@' */ + for (i = 0; i < size; ++i) { + if (isalnum(data[i])) + continue; + + switch (data[i]) { + case '@': + nb++; + + case '-': + case '.': + case '_': + break; + + case '>': + return (nb == 1) ? i + 1 : 0; + + default: + return 0; + } + } + + return 0; +} + +/* tag_length • returns the length of the given tag, or 0 is it's not valid */ +static size_t +tag_length(uint8_t *data, size_t size, hoedown_autolink_type *autolink) +{ + size_t i, j; + + /* a valid tag can't be shorter than 3 chars */ + if (size < 3) return 0; + + /* begins with a '<' optionally followed by '/', followed by letter or number */ + if (data[0] != '<') return 0; + i = (data[1] == '/') ? 2 : 1; + + if (!isalnum(data[i])) + return 0; + + /* scheme test */ + *autolink = HOEDOWN_AUTOLINK_NONE; + + /* try to find the beginning of an URI */ + while (i < size && (isalnum(data[i]) || data[i] == '.' || data[i] == '+' || data[i] == '-')) + i++; + + if (i > 1 && data[i] == '@') { + if ((j = is_mail_autolink(data + i, size - i)) != 0) { + *autolink = HOEDOWN_AUTOLINK_EMAIL; + return i + j; + } + } + + if (i > 2 && data[i] == ':') { + *autolink = HOEDOWN_AUTOLINK_NORMAL; + i++; + } + + /* completing autolink test: no spacing or ' or " */ + if (i >= size) + *autolink = HOEDOWN_AUTOLINK_NONE; + + else if (*autolink) { + j = i; + + while (i < size) { + if (data[i] == '\\') i += 2; + else if (data[i] == '>' || data[i] == '\'' || + data[i] == '"' || data[i] == ' ' || data[i] == '\n') + break; + else i++; + } + + if (i >= size) return 0; + if (i > j && data[i] == '>') return i + 1; + /* one of the forbidden chars has been found */ + *autolink = HOEDOWN_AUTOLINK_NONE; + } + + /* looking for something looking like a tag end */ + while (i < size && data[i] != '>') i++; + if (i >= size) return 0; + return i + 1; +} + +/* parse_inline • parses inline markdown elements */ +static void +parse_inline(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size) +{ + size_t i = 0, end = 0, consumed = 0; + hoedown_buffer work = { 0, 0, 0, 0, NULL, NULL, NULL }; + uint8_t *active_char = doc->active_char; + + if (doc->work_bufs[BUFFER_SPAN].size + + doc->work_bufs[BUFFER_BLOCK].size > doc->max_nesting) + return; + + while (i < size) { + /* copying inactive chars into the output */ + while (end < size && active_char[data[end]] == 0) + end++; + + if (doc->md.normal_text) { + work.data = data + i; + work.size = end - i; + doc->md.normal_text(ob, &work, &doc->data); + } + else + hoedown_buffer_put(ob, data + i, end - i); + + if (end >= size) break; + i = end; + + end = markdown_char_ptrs[ (int)active_char[data[end]] ](ob, doc, data + i, i - consumed, size - i); + if (!end) /* no action from the callback */ + end = i + 1; + else { + i += end; + end = i; + consumed = i; + } + } +} + +/* is_escaped • returns whether special char at data[loc] is escaped by '\\' */ +static int +is_escaped(uint8_t *data, size_t loc) +{ + size_t i = loc; + while (i >= 1 && data[i - 1] == '\\') + i--; + + /* odd numbers of backslashes escapes data[loc] */ + return (loc - i) % 2; +} + +/* find_emph_char • looks for the next emph uint8_t, skipping other constructs */ +static size_t +find_emph_char(uint8_t *data, size_t size, uint8_t c) +{ + size_t i = 0; + + while (i < size) { + while (i < size && data[i] != c && data[i] != '[' && data[i] != '`') + i++; + + if (i == size) + return 0; + + /* not counting escaped chars */ + if (is_escaped(data, i)) { + i++; continue; + } + + if (data[i] == c) + return i; + + /* skipping a codespan */ + if (data[i] == '`') { + size_t span_nb = 0, bt; + size_t tmp_i = 0; + + /* counting the number of opening backticks */ + while (i < size && data[i] == '`') { + i++; span_nb++; + } + + if (i >= size) return 0; + + /* finding the matching closing sequence */ + bt = 0; + while (i < size && bt < span_nb) { + if (!tmp_i && data[i] == c) tmp_i = i; + if (data[i] == '`') bt++; + else bt = 0; + i++; + } + + /* not a well-formed codespan; use found matching emph char */ + if (i >= size) return tmp_i; + } + /* skipping a link */ + else if (data[i] == '[') { + size_t tmp_i = 0; + uint8_t cc; + + i++; + while (i < size && data[i] != ']') { + if (!tmp_i && data[i] == c) tmp_i = i; + i++; + } + + i++; + while (i < size && _isspace(data[i])) + i++; + + if (i >= size) + return tmp_i; + + switch (data[i]) { + case '[': + cc = ']'; break; + + case '(': + cc = ')'; break; + + default: + if (tmp_i) + return tmp_i; + else + continue; + } + + i++; + while (i < size && data[i] != cc) { + if (!tmp_i && data[i] == c) tmp_i = i; + i++; + } + + if (i >= size) + return tmp_i; + + i++; + } + } + + return 0; +} + +/* parse_emph1 • parsing single emphase */ +/* closed by a symbol not preceded by spacing and not followed by symbol */ +static size_t +parse_emph1(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size, uint8_t c) +{ + size_t i = 0, len; + hoedown_buffer *work = 0; + int r; + + /* skipping one symbol if coming from emph3 */ + if (size > 1 && data[0] == c && data[1] == c) i = 1; + + while (i < size) { + len = find_emph_char(data + i, size - i, c); + if (!len) return 0; + i += len; + if (i >= size) return 0; + + if (data[i] == c && !_isspace(data[i - 1])) { + + if (doc->ext_flags & HOEDOWN_EXT_NO_INTRA_EMPHASIS) { + if (i + 1 < size && isalnum(data[i + 1])) + continue; + } + + work = newbuf(doc, BUFFER_SPAN); + parse_inline(work, doc, data, i); + + if (doc->ext_flags & HOEDOWN_EXT_UNDERLINE && c == '_') + r = doc->md.underline(ob, work, &doc->data); + else + r = doc->md.emphasis(ob, work, &doc->data); + + popbuf(doc, BUFFER_SPAN); + return r ? i + 1 : 0; + } + } + + return 0; +} + +/* parse_emph2 • parsing single emphase */ +static size_t +parse_emph2(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size, uint8_t c) +{ + size_t i = 0, len; + hoedown_buffer *work = 0; + int r; + + while (i < size) { + len = find_emph_char(data + i, size - i, c); + if (!len) return 0; + i += len; + + if (i + 1 < size && data[i] == c && data[i + 1] == c && i && !_isspace(data[i - 1])) { + work = newbuf(doc, BUFFER_SPAN); + parse_inline(work, doc, data, i); + + if (c == '~') + r = doc->md.strikethrough(ob, work, &doc->data); + else if (c == '=') + r = doc->md.highlight(ob, work, &doc->data); + else + r = doc->md.double_emphasis(ob, work, &doc->data); + + popbuf(doc, BUFFER_SPAN); + return r ? i + 2 : 0; + } + i++; + } + return 0; +} + +/* parse_emph3 • parsing single emphase */ +/* finds the first closing tag, and delegates to the other emph */ +static size_t +parse_emph3(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size, uint8_t c) +{ + size_t i = 0, len; + int r; + + while (i < size) { + len = find_emph_char(data + i, size - i, c); + if (!len) return 0; + i += len; + + /* skip spacing preceded symbols */ + if (data[i] != c || _isspace(data[i - 1])) + continue; + + if (i + 2 < size && data[i + 1] == c && data[i + 2] == c && doc->md.triple_emphasis) { + /* triple symbol found */ + hoedown_buffer *work = newbuf(doc, BUFFER_SPAN); + + parse_inline(work, doc, data, i); + r = doc->md.triple_emphasis(ob, work, &doc->data); + popbuf(doc, BUFFER_SPAN); + return r ? i + 3 : 0; + + } else if (i + 1 < size && data[i + 1] == c) { + /* double symbol found, handing over to emph1 */ + len = parse_emph1(ob, doc, data - 2, size + 2, c); + if (!len) return 0; + else return len - 2; + + } else { + /* single symbol found, handing over to emph2 */ + len = parse_emph2(ob, doc, data - 1, size + 1, c); + if (!len) return 0; + else return len - 1; + } + } + return 0; +} + +/* parse_math • parses a math span until the given ending delimiter */ +static size_t +parse_math(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size, const char *end, size_t delimsz, int displaymode) +{ + hoedown_buffer text = { NULL, 0, 0, 0, NULL, NULL, NULL }; + size_t i = delimsz; + + if (!doc->md.math) + return 0; + + /* find ending delimiter */ + while (1) { + while (i < size && data[i] != (uint8_t)end[0]) + i++; + + if (i >= size) + return 0; + + if (!is_escaped(data, i) && !(i + delimsz > size) + && memcmp(data + i, end, delimsz) == 0) + break; + + i++; + } + + /* prepare buffers */ + text.data = data + delimsz; + text.size = i - delimsz; + + /* if this is a $$ and MATH_EXPLICIT is not active, + * guess whether displaymode should be enabled from the context */ + i += delimsz; + if (delimsz == 2 && !(doc->ext_flags & HOEDOWN_EXT_MATH_EXPLICIT)) + displaymode = is_empty_all(data - offset, offset) && is_empty_all(data + i, size - i); + + /* call callback */ + if (doc->md.math(ob, &text, displaymode, &doc->data)) + return i; + + return 0; +} + +/* char_emphasis • single and double emphasis parsing */ +static size_t +char_emphasis(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + uint8_t c = data[0]; + size_t ret; + + if (doc->ext_flags & HOEDOWN_EXT_NO_INTRA_EMPHASIS) { + if (offset > 0 && !_isspace(data[-1]) && data[-1] != '>' && data[-1] != '(') + return 0; + } + + if (size > 2 && data[1] != c) { + /* spacing cannot follow an opening emphasis; + * strikethrough and highlight only takes two characters '~~' */ + if (c == '~' || c == '=' || _isspace(data[1]) || (ret = parse_emph1(ob, doc, data + 1, size - 1, c)) == 0) + return 0; + + return ret + 1; + } + + if (size > 3 && data[1] == c && data[2] != c) { + if (_isspace(data[2]) || (ret = parse_emph2(ob, doc, data + 2, size - 2, c)) == 0) + return 0; + + return ret + 2; + } + + if (size > 4 && data[1] == c && data[2] == c && data[3] != c) { + if (c == '~' || c == '=' || _isspace(data[3]) || (ret = parse_emph3(ob, doc, data + 3, size - 3, c)) == 0) + return 0; + + return ret + 3; + } + + return 0; +} + + +/* char_linebreak • '\n' preceded by two spaces (assuming linebreak != 0) */ +static size_t +char_linebreak(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + if (offset < 2 || data[-1] != ' ' || data[-2] != ' ') + return 0; + + /* removing the last space from ob and rendering */ + while (ob->size && ob->data[ob->size - 1] == ' ') + ob->size--; + + return doc->md.linebreak(ob, &doc->data) ? 1 : 0; +} + + +/* char_codespan • '`' parsing a code span (assuming codespan != 0) */ +static size_t +char_codespan(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + hoedown_buffer work = { NULL, 0, 0, 0, NULL, NULL, NULL }; + size_t end, nb = 0, i, f_begin, f_end; + + /* counting the number of backticks in the delimiter */ + while (nb < size && data[nb] == '`') + nb++; + + /* finding the next delimiter */ + i = 0; + for (end = nb; end < size && i < nb; end++) { + if (data[end] == '`') i++; + else i = 0; + } + + if (i < nb && end >= size) + return 0; /* no matching delimiter */ + + /* trimming outside spaces */ + f_begin = nb; + while (f_begin < end && data[f_begin] == ' ') + f_begin++; + + f_end = end - nb; + while (f_end > nb && data[f_end-1] == ' ') + f_end--; + + /* real code span */ + if (f_begin < f_end) { + work.data = data + f_begin; + work.size = f_end - f_begin; + + if (!doc->md.codespan(ob, &work, &doc->data)) + end = 0; + } else { + if (!doc->md.codespan(ob, 0, &doc->data)) + end = 0; + } + + return end; +} + +/* char_quote • '"' parsing a quote */ +static size_t +char_quote(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + size_t end, nq = 0, i, f_begin, f_end; + + /* counting the number of quotes in the delimiter */ + while (nq < size && data[nq] == '"') + nq++; + + /* finding the next delimiter */ + end = nq; + while (1) { + i = end; + end += find_emph_char(data + end, size - end, '"'); + if (end == i) return 0; /* no matching delimiter */ + i = end; + while (end < size && data[end] == '"' && end - i < nq) end++; + if (end - i >= nq) break; + } + + /* trimming outside spaces */ + f_begin = nq; + while (f_begin < end && data[f_begin] == ' ') + f_begin++; + + f_end = end - nq; + while (f_end > nq && data[f_end-1] == ' ') + f_end--; + + /* real quote */ + if (f_begin < f_end) { + hoedown_buffer *work = newbuf(doc, BUFFER_SPAN); + parse_inline(work, doc, data + f_begin, f_end - f_begin); + + if (!doc->md.quote(ob, work, &doc->data)) + end = 0; + popbuf(doc, BUFFER_SPAN); + } else { + if (!doc->md.quote(ob, 0, &doc->data)) + end = 0; + } + + return end; +} + + +/* char_escape • '\\' backslash escape */ +static size_t +char_escape(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + static const char *escape_chars = "\\`*_{}[]()#+-.!:|&<>^~=\"$"; + hoedown_buffer work = { 0, 0, 0, 0, NULL, NULL, NULL }; + size_t w; + + if (size > 1) { + if (data[1] == '\\' && (doc->ext_flags & HOEDOWN_EXT_MATH) && + size > 2 && (data[2] == '(' || data[2] == '[')) { + const char *end = (data[2] == '[') ? "\\\\]" : "\\\\)"; + w = parse_math(ob, doc, data, offset, size, end, 3, data[2] == '['); + if (w) return w; + } + + if (strchr(escape_chars, data[1]) == NULL) + return 0; + + if (doc->md.normal_text) { + work.data = data + 1; + work.size = 1; + doc->md.normal_text(ob, &work, &doc->data); + } + else hoedown_buffer_putc(ob, data[1]); + } else if (size == 1) { + hoedown_buffer_putc(ob, data[0]); + } + + return 2; +} + +/* char_entity • '&' escaped when it doesn't belong to an entity */ +/* valid entities are assumed to be anything matching &#?[A-Za-z0-9]+; */ +static size_t +char_entity(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + size_t end = 1; + hoedown_buffer work = { 0, 0, 0, 0, NULL, NULL, NULL }; + + if (end < size && data[end] == '#') + end++; + + while (end < size && isalnum(data[end])) + end++; + + if (end < size && data[end] == ';') + end++; /* real entity */ + else + return 0; /* lone '&' */ + + if (doc->md.entity) { + work.data = data; + work.size = end; + doc->md.entity(ob, &work, &doc->data); + } + else hoedown_buffer_put(ob, data, end); + + return end; +} + +/* char_langle_tag • '<' when tags or autolinks are allowed */ +static size_t +char_langle_tag(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + hoedown_buffer work = { NULL, 0, 0, 0, NULL, NULL, NULL }; + hoedown_autolink_type altype = HOEDOWN_AUTOLINK_NONE; + size_t end = tag_length(data, size, &altype); + int ret = 0; + + work.data = data; + work.size = end; + + if (end > 2) { + if (doc->md.autolink && altype != HOEDOWN_AUTOLINK_NONE) { + hoedown_buffer *u_link = newbuf(doc, BUFFER_SPAN); + work.data = data + 1; + work.size = end - 2; + unscape_text(u_link, &work); + ret = doc->md.autolink(ob, u_link, altype, &doc->data); + popbuf(doc, BUFFER_SPAN); + } + else if (doc->md.raw_html) + ret = doc->md.raw_html(ob, &work, &doc->data); + } + + if (!ret) return 0; + else return end; +} + +static size_t +char_autolink_www(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + hoedown_buffer *link, *link_url, *link_text; + size_t link_len, rewind; + + if (!doc->md.link || doc->in_link_body) + return 0; + + link = newbuf(doc, BUFFER_SPAN); + + if ((link_len = hoedown_autolink__www(&rewind, link, data, offset, size, HOEDOWN_AUTOLINK_SHORT_DOMAINS)) > 0) { + link_url = newbuf(doc, BUFFER_SPAN); + HOEDOWN_BUFPUTSL(link_url, "http://"); + hoedown_buffer_put(link_url, link->data, link->size); + + ob->size -= rewind; + if (doc->md.normal_text) { + link_text = newbuf(doc, BUFFER_SPAN); + doc->md.normal_text(link_text, link, &doc->data); + doc->md.link(ob, link_text, link_url, NULL, &doc->data); + popbuf(doc, BUFFER_SPAN); + } else { + doc->md.link(ob, link, link_url, NULL, &doc->data); + } + popbuf(doc, BUFFER_SPAN); + } + + popbuf(doc, BUFFER_SPAN); + return link_len; +} + +static size_t +char_autolink_email(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + hoedown_buffer *link; + size_t link_len, rewind; + + if (!doc->md.autolink || doc->in_link_body) + return 0; + + link = newbuf(doc, BUFFER_SPAN); + + if ((link_len = hoedown_autolink__email(&rewind, link, data, offset, size, 0)) > 0) { + ob->size -= rewind; + doc->md.autolink(ob, link, HOEDOWN_AUTOLINK_EMAIL, &doc->data); + } + + popbuf(doc, BUFFER_SPAN); + return link_len; +} + +static size_t +char_autolink_url(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + hoedown_buffer *link; + size_t link_len, rewind; + + if (!doc->md.autolink || doc->in_link_body) + return 0; + + link = newbuf(doc, BUFFER_SPAN); + + if ((link_len = hoedown_autolink__url(&rewind, link, data, offset, size, 0)) > 0) { + ob->size -= rewind; + doc->md.autolink(ob, link, HOEDOWN_AUTOLINK_NORMAL, &doc->data); + } + + popbuf(doc, BUFFER_SPAN); + return link_len; +} + +/* char_link • '[': parsing a link, a footnote or an image */ +static size_t +char_link(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + int is_img = (offset && data[-1] == '!' && !is_escaped(data - offset, offset - 1)); + int is_footnote = (doc->ext_flags & HOEDOWN_EXT_FOOTNOTES && data[1] == '^'); + size_t i = 1, txt_e, link_b = 0, link_e = 0, title_b = 0, title_e = 0; + hoedown_buffer *content = NULL; + hoedown_buffer *link = NULL; + hoedown_buffer *title = NULL; + hoedown_buffer *u_link = NULL; + size_t org_work_size = doc->work_bufs[BUFFER_SPAN].size; + int ret = 0, in_title = 0, qtype = 0; + + /* checking whether the correct renderer exists */ + if ((is_footnote && !doc->md.footnote_ref) || (is_img && !doc->md.image) + || (!is_img && !is_footnote && !doc->md.link)) + goto cleanup; + + /* looking for the matching closing bracket */ + i += find_emph_char(data + i, size - i, ']'); + txt_e = i; + + if (i < size && data[i] == ']') i++; + else goto cleanup; + + /* footnote link */ + if (is_footnote) { + hoedown_buffer id = { NULL, 0, 0, 0, NULL, NULL, NULL }; + struct footnote_ref *fr; + + if (txt_e < 3) + goto cleanup; + + id.data = data + 2; + id.size = txt_e - 2; + + fr = find_footnote_ref(&doc->footnotes_found, id.data, id.size); + + /* mark footnote used */ + if (fr && !fr->is_used) { + if(!add_footnote_ref(&doc->footnotes_used, fr)) + goto cleanup; + fr->is_used = 1; + fr->num = doc->footnotes_used.count; + + /* render */ + if (doc->md.footnote_ref) + ret = doc->md.footnote_ref(ob, fr->num, &doc->data); + } + + goto cleanup; + } + + /* skip any amount of spacing */ + /* (this is much more laxist than original markdown syntax) */ + while (i < size && _isspace(data[i])) + i++; + + /* inline style link */ + if (i < size && data[i] == '(') { + size_t nb_p; + + /* skipping initial spacing */ + i++; + + while (i < size && _isspace(data[i])) + i++; + + link_b = i; + + /* looking for link end: ' " ) */ + /* Count the number of open parenthesis */ + nb_p = 0; + + while (i < size) { + if (data[i] == '\\') i += 2; + else if (data[i] == '(' && i != 0) { + nb_p++; i++; + } + else if (data[i] == ')') { + if (nb_p == 0) break; + else nb_p--; i++; + } else if (i >= 1 && _isspace(data[i-1]) && (data[i] == '\'' || data[i] == '"')) break; + else i++; + } + + if (i >= size) goto cleanup; + link_e = i; + + /* looking for title end if present */ + if (data[i] == '\'' || data[i] == '"') { + qtype = data[i]; + in_title = 1; + i++; + title_b = i; + + while (i < size) { + if (data[i] == '\\') i += 2; + else if (data[i] == qtype) {in_title = 0; i++;} + else if ((data[i] == ')') && !in_title) break; + else i++; + } + + if (i >= size) goto cleanup; + + /* skipping spacing after title */ + title_e = i - 1; + while (title_e > title_b && _isspace(data[title_e])) + title_e--; + + /* checking for closing quote presence */ + if (data[title_e] != '\'' && data[title_e] != '"') { + title_b = title_e = 0; + link_e = i; + } + } + + /* remove spacing at the end of the link */ + while (link_e > link_b && _isspace(data[link_e - 1])) + link_e--; + + /* remove optional angle brackets around the link */ + if (data[link_b] == '<') link_b++; + if (data[link_e - 1] == '>') link_e--; + + /* building escaped link and title */ + if (link_e > link_b) { + link = newbuf(doc, BUFFER_SPAN); + hoedown_buffer_put(link, data + link_b, link_e - link_b); + } + + if (title_e > title_b) { + title = newbuf(doc, BUFFER_SPAN); + hoedown_buffer_put(title, data + title_b, title_e - title_b); + } + + i++; + } + + /* reference style link */ + else if (i < size && data[i] == '[') { + hoedown_buffer *id = newbuf(doc, BUFFER_SPAN); + struct link_ref *lr; + + /* looking for the id */ + i++; + link_b = i; + while (i < size && data[i] != ']') i++; + if (i >= size) goto cleanup; + link_e = i; + + /* finding the link_ref */ + if (link_b == link_e) + replace_spacing(id, data + 1, txt_e - 1); + else + hoedown_buffer_put(id, data + link_b, link_e - link_b); + + lr = find_link_ref(doc->refs, id->data, id->size); + if (!lr) + goto cleanup; + + /* keeping link and title from link_ref */ + link = lr->link; + title = lr->title; + i++; + } + + /* shortcut reference style link */ + else { + hoedown_buffer *id = newbuf(doc, BUFFER_SPAN); + struct link_ref *lr; + + /* crafting the id */ + replace_spacing(id, data + 1, txt_e - 1); + + /* finding the link_ref */ + lr = find_link_ref(doc->refs, id->data, id->size); + if (!lr) + goto cleanup; + + /* keeping link and title from link_ref */ + link = lr->link; + title = lr->title; + + /* rewinding the spacing */ + i = txt_e + 1; + } + + /* building content: img alt is kept, only link content is parsed */ + if (txt_e > 1) { + content = newbuf(doc, BUFFER_SPAN); + if (is_img) { + hoedown_buffer_put(content, data + 1, txt_e - 1); + } else { + /* disable autolinking when parsing inline the + * content of a link */ + doc->in_link_body = 1; + parse_inline(content, doc, data + 1, txt_e - 1); + doc->in_link_body = 0; + } + } + + if (link) { + u_link = newbuf(doc, BUFFER_SPAN); + unscape_text(u_link, link); + } + + /* calling the relevant rendering function */ + if (is_img) { + if (ob->size && ob->data[ob->size - 1] == '!') + ob->size -= 1; + + ret = doc->md.image(ob, u_link, title, content, &doc->data); + } else { + ret = doc->md.link(ob, content, u_link, title, &doc->data); + } + + /* cleanup */ +cleanup: + doc->work_bufs[BUFFER_SPAN].size = (int)org_work_size; + return ret ? i : 0; +} + +static size_t +char_superscript(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + size_t sup_start, sup_len; + hoedown_buffer *sup; + + if (!doc->md.superscript) + return 0; + + if (size < 2) + return 0; + + if (data[1] == '(') { + sup_start = 2; + sup_len = find_emph_char(data + 2, size - 2, ')') + 2; + + if (sup_len == size) + return 0; + } else { + sup_start = sup_len = 1; + + while (sup_len < size && !_isspace(data[sup_len])) + sup_len++; + } + + if (sup_len - sup_start == 0) + return (sup_start == 2) ? 3 : 0; + + sup = newbuf(doc, BUFFER_SPAN); + parse_inline(sup, doc, data + sup_start, sup_len - sup_start); + doc->md.superscript(ob, sup, &doc->data); + popbuf(doc, BUFFER_SPAN); + + return (sup_start == 2) ? sup_len + 1 : sup_len; +} + +static size_t +char_math(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t offset, size_t size) +{ + /* double dollar */ + if (size > 1 && data[1] == '$') + return parse_math(ob, doc, data, offset, size, "$$", 2, 1); + + /* single dollar allowed only with MATH_EXPLICIT flag */ + if (doc->ext_flags & HOEDOWN_EXT_MATH_EXPLICIT) + return parse_math(ob, doc, data, offset, size, "$", 1, 0); + + return 0; +} + +/********************************* + * BLOCK-LEVEL PARSING FUNCTIONS * + *********************************/ + +/* is_empty • returns the line length when it is empty, 0 otherwise */ +static size_t +is_empty(const uint8_t *data, size_t size) +{ + size_t i; + + for (i = 0; i < size && data[i] != '\n'; i++) + if (data[i] != ' ') + return 0; + + return i + 1; +} + +/* is_hrule • returns whether a line is a horizontal rule */ +static int +is_hrule(uint8_t *data, size_t size) +{ + size_t i = 0, n = 0; + uint8_t c; + + /* skipping initial spaces */ + if (size < 3) return 0; + if (data[0] == ' ') { i++; + if (data[1] == ' ') { i++; + if (data[2] == ' ') { i++; } } } + + /* looking at the hrule uint8_t */ + if (i + 2 >= size + || (data[i] != '*' && data[i] != '-' && data[i] != '_')) + return 0; + c = data[i]; + + /* the whole line must be the char or space */ + while (i < size && data[i] != '\n') { + if (data[i] == c) n++; + else if (data[i] != ' ') + return 0; + + i++; + } + + return n >= 3; +} + +/* check if a line is a code fence; return the + * end of the code fence. if passed, width of + * the fence rule and character will be returned */ +static size_t +is_codefence(uint8_t *data, size_t size, size_t *width, uint8_t *chr) +{ + size_t i = 0, n = 1; + uint8_t c; + + /* skipping initial spaces */ + if (size < 3) + return 0; + + if (data[0] == ' ') { i++; + if (data[1] == ' ') { i++; + if (data[2] == ' ') { i++; } } } + + /* looking at the hrule uint8_t */ + c = data[i]; + if (i + 2 >= size || !(c=='~' || c=='`')) + return 0; + + /* the fence must be that same character */ + while (++i < size && data[i] == c) + ++n; + + if (n < 3) + return 0; + + if (width) *width = n; + if (chr) *chr = c; + return i; +} + +/* expects single line, checks if it's a codefence and extracts language */ +static size_t +parse_codefence(uint8_t *data, size_t size, hoedown_buffer *lang, size_t *width, uint8_t *chr) +{ + size_t i, w, lang_start; + + i = w = is_codefence(data, size, width, chr); + if (i == 0) + return 0; + + while (i < size && _isspace(data[i])) + i++; + + lang_start = i; + + while (i < size && !_isspace(data[i])) + i++; + + lang->data = data + lang_start; + lang->size = i - lang_start; + + /* Avoid parsing a codespan as a fence */ + i = lang_start + 2; + while (i < size && !(data[i] == *chr && data[i-1] == *chr && data[i-2] == *chr)) i++; + if (i < size) return 0; + + return w; +} + +/* is_atxheader • returns whether the line is a hash-prefixed header */ +static int +is_atxheader(hoedown_document *doc, uint8_t *data, size_t size) +{ + if (data[0] != '#') + return 0; + + if (doc->ext_flags & HOEDOWN_EXT_SPACE_HEADERS) { + size_t level = 0; + + while (level < size && level < 6 && data[level] == '#') + level++; + + if (level < size && data[level] != ' ') + return 0; + } + + return 1; +} + +/* is_headerline • returns whether the line is a setext-style hdr underline */ +static int +is_headerline(uint8_t *data, size_t size) +{ + size_t i = 0; + + /* test of level 1 header */ + if (data[i] == '=') { + for (i = 1; i < size && data[i] == '='; i++); + while (i < size && data[i] == ' ') i++; + return (i >= size || data[i] == '\n') ? 1 : 0; } + + /* test of level 2 header */ + if (data[i] == '-') { + for (i = 1; i < size && data[i] == '-'; i++); + while (i < size && data[i] == ' ') i++; + return (i >= size || data[i] == '\n') ? 2 : 0; } + + return 0; +} + +static int +is_next_headerline(uint8_t *data, size_t size) +{ + size_t i = 0; + + while (i < size && data[i] != '\n') + i++; + + if (++i >= size) + return 0; + + return is_headerline(data + i, size - i); +} + +/* prefix_quote • returns blockquote prefix length */ +static size_t +prefix_quote(uint8_t *data, size_t size) +{ + size_t i = 0; + if (i < size && data[i] == ' ') i++; + if (i < size && data[i] == ' ') i++; + if (i < size && data[i] == ' ') i++; + + if (i < size && data[i] == '>') { + if (i + 1 < size && data[i + 1] == ' ') + return i + 2; + + return i + 1; + } + + return 0; +} + +/* prefix_code • returns prefix length for block code*/ +static size_t +prefix_code(uint8_t *data, size_t size) +{ + if (size > 3 && data[0] == ' ' && data[1] == ' ' + && data[2] == ' ' && data[3] == ' ') return 4; + + return 0; +} + +/* prefix_oli • returns ordered list item prefix */ +static size_t +prefix_oli(uint8_t *data, size_t size) +{ + size_t i = 0; + + if (i < size && data[i] == ' ') i++; + if (i < size && data[i] == ' ') i++; + if (i < size && data[i] == ' ') i++; + + if (i >= size || data[i] < '0' || data[i] > '9') + return 0; + + while (i < size && data[i] >= '0' && data[i] <= '9') + i++; + + if (i + 1 >= size || data[i] != '.' || data[i + 1] != ' ') + return 0; + + if (is_next_headerline(data + i, size - i)) + return 0; + + return i + 2; +} + +/* prefix_uli • returns ordered list item prefix */ +static size_t +prefix_uli(uint8_t *data, size_t size) +{ + size_t i = 0; + + if (i < size && data[i] == ' ') i++; + if (i < size && data[i] == ' ') i++; + if (i < size && data[i] == ' ') i++; + + if (i + 1 >= size || + (data[i] != '*' && data[i] != '+' && data[i] != '-') || + data[i + 1] != ' ') + return 0; + + if (is_next_headerline(data + i, size - i)) + return 0; + + return i + 2; +} + + +/* parse_block • parsing of one block, returning next uint8_t to parse */ +static void parse_block(hoedown_buffer *ob, hoedown_document *doc, + uint8_t *data, size_t size); + + +/* parse_blockquote • handles parsing of a blockquote fragment */ +static size_t +parse_blockquote(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size) +{ + size_t beg, end = 0, pre, work_size = 0; + uint8_t *work_data = 0; + hoedown_buffer *out = 0; + + out = newbuf(doc, BUFFER_BLOCK); + beg = 0; + while (beg < size) { + for (end = beg + 1; end < size && data[end - 1] != '\n'; end++); + + pre = prefix_quote(data + beg, end - beg); + + if (pre) + beg += pre; /* skipping prefix */ + + /* empty line followed by non-quote line */ + else if (is_empty(data + beg, end - beg) && + (end >= size || (prefix_quote(data + end, size - end) == 0 && + !is_empty(data + end, size - end)))) + break; + + if (beg < end) { /* copy into the in-place working buffer */ + /* hoedown_buffer_put(work, data + beg, end - beg); */ + if (!work_data) + work_data = data + beg; + else if (data + beg != work_data + work_size) + memmove(work_data + work_size, data + beg, end - beg); + work_size += end - beg; + } + beg = end; + } + + parse_block(out, doc, work_data, work_size); + if (doc->md.blockquote) + doc->md.blockquote(ob, out, &doc->data); + popbuf(doc, BUFFER_BLOCK); + return end; +} + +static size_t +parse_htmlblock(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size, int do_render); + +/* parse_blockquote • handles parsing of a regular paragraph */ +static size_t +parse_paragraph(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size) +{ + hoedown_buffer work = { NULL, 0, 0, 0, NULL, NULL, NULL }; + size_t i = 0, end = 0; + int level = 0; + + work.data = data; + + while (i < size) { + for (end = i + 1; end < size && data[end - 1] != '\n'; end++) /* empty */; + + if (is_empty(data + i, size - i)) + break; + + if ((level = is_headerline(data + i, size - i)) != 0) + break; + + if (is_atxheader(doc, data + i, size - i) || + is_hrule(data + i, size - i) || + prefix_quote(data + i, size - i)) { + end = i; + break; + } + + i = end; + } + + work.size = i; + while (work.size && data[work.size - 1] == '\n') + work.size--; + + if (!level) { + hoedown_buffer *tmp = newbuf(doc, BUFFER_BLOCK); + parse_inline(tmp, doc, work.data, work.size); + if (doc->md.paragraph) + doc->md.paragraph(ob, tmp, &doc->data); + popbuf(doc, BUFFER_BLOCK); + } else { + hoedown_buffer *header_work; + + if (work.size) { + size_t beg; + i = work.size; + work.size -= 1; + + while (work.size && data[work.size] != '\n') + work.size -= 1; + + beg = work.size + 1; + while (work.size && data[work.size - 1] == '\n') + work.size -= 1; + + if (work.size > 0) { + hoedown_buffer *tmp = newbuf(doc, BUFFER_BLOCK); + parse_inline(tmp, doc, work.data, work.size); + + if (doc->md.paragraph) + doc->md.paragraph(ob, tmp, &doc->data); + + popbuf(doc, BUFFER_BLOCK); + work.data += beg; + work.size = i - beg; + } + else work.size = i; + } + + header_work = newbuf(doc, BUFFER_SPAN); + parse_inline(header_work, doc, work.data, work.size); + + if (doc->md.header) + doc->md.header(ob, header_work, (int)level, &doc->data); + + popbuf(doc, BUFFER_SPAN); + } + + return end; +} + +/* parse_fencedcode • handles parsing of a block-level code fragment */ +static size_t +parse_fencedcode(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size) +{ + hoedown_buffer text = { 0, 0, 0, 0, NULL, NULL, NULL }; + hoedown_buffer lang = { 0, 0, 0, 0, NULL, NULL, NULL }; + size_t i = 0, text_start, line_start; + size_t w, w2; + size_t width, width2; + uint8_t chr, chr2; + + /* parse codefence line */ + while (i < size && data[i] != '\n') + i++; + + w = parse_codefence(data, i, &lang, &width, &chr); + if (!w) + return 0; + + /* search for end */ + i++; + text_start = i; + while ((line_start = i) < size) { + while (i < size && data[i] != '\n') + i++; + + w2 = is_codefence(data + line_start, i - line_start, &width2, &chr2); + if (w == w2 && width == width2 && chr == chr2 && + is_empty(data + (line_start+w), i - (line_start+w))) + break; + + i++; + } + + text.data = data + text_start; + text.size = line_start - text_start; + + if (doc->md.blockcode) + doc->md.blockcode(ob, text.size ? &text : NULL, lang.size ? &lang : NULL, &doc->data); + + return i; +} + +static size_t +parse_blockcode(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size) +{ + size_t beg, end, pre; + hoedown_buffer *work = 0; + + work = newbuf(doc, BUFFER_BLOCK); + + beg = 0; + while (beg < size) { + for (end = beg + 1; end < size && data[end - 1] != '\n'; end++) {}; + pre = prefix_code(data + beg, end - beg); + + if (pre) + beg += pre; /* skipping prefix */ + else if (!is_empty(data + beg, end - beg)) + /* non-empty non-prefixed line breaks the pre */ + break; + + if (beg < end) { + /* verbatim copy to the working buffer, + escaping entities */ + if (is_empty(data + beg, end - beg)) + hoedown_buffer_putc(work, '\n'); + else hoedown_buffer_put(work, data + beg, end - beg); + } + beg = end; + } + + while (work->size && work->data[work->size - 1] == '\n') + work->size -= 1; + + hoedown_buffer_putc(work, '\n'); + + if (doc->md.blockcode) + doc->md.blockcode(ob, work, NULL, &doc->data); + + popbuf(doc, BUFFER_BLOCK); + return beg; +} + +/* parse_listitem • parsing of a single list item */ +/* assuming initial prefix is already removed */ +static size_t +parse_listitem(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size, hoedown_list_flags *flags) +{ + hoedown_buffer *work = 0, *inter = 0; + size_t beg = 0, end, pre, sublist = 0, orgpre = 0, i; + int in_empty = 0, has_inside_empty = 0, in_fence = 0; + + /* keeping track of the first indentation prefix */ + while (orgpre < 3 && orgpre < size && data[orgpre] == ' ') + orgpre++; + + beg = prefix_uli(data, size); + if (!beg) + beg = prefix_oli(data, size); + + if (!beg) + return 0; + + /* skipping to the beginning of the following line */ + end = beg; + while (end < size && data[end - 1] != '\n') + end++; + + /* getting working buffers */ + work = newbuf(doc, BUFFER_SPAN); + inter = newbuf(doc, BUFFER_SPAN); + + /* putting the first line into the working buffer */ + hoedown_buffer_put(work, data + beg, end - beg); + beg = end; + + /* process the following lines */ + while (beg < size) { + size_t has_next_uli = 0, has_next_oli = 0; + + end++; + + while (end < size && data[end - 1] != '\n') + end++; + + /* process an empty line */ + if (is_empty(data + beg, end - beg)) { + in_empty = 1; + beg = end; + continue; + } + + /* calculating the indentation */ + i = 0; + while (i < 4 && beg + i < end && data[beg + i] == ' ') + i++; + + pre = i; + + if (doc->ext_flags & HOEDOWN_EXT_FENCED_CODE) { + if (is_codefence(data + beg + i, end - beg - i, NULL, NULL)) + in_fence = !in_fence; + } + + /* Only check for new list items if we are **not** inside + * a fenced code block */ + if (!in_fence) { + has_next_uli = prefix_uli(data + beg + i, end - beg - i); + has_next_oli = prefix_oli(data + beg + i, end - beg - i); + } + + /* checking for a new item */ + if ((has_next_uli && !is_hrule(data + beg + i, end - beg - i)) || has_next_oli) { + if (in_empty) + has_inside_empty = 1; + + /* the following item must have the same (or less) indentation */ + if (pre <= orgpre) { + /* if the following item has different list type, we end this list */ + if (in_empty && ( + ((*flags & HOEDOWN_LIST_ORDERED) && has_next_uli) || + (!(*flags & HOEDOWN_LIST_ORDERED) && has_next_oli))) + *flags |= HOEDOWN_LI_END; + + break; + } + + if (!sublist) + sublist = work->size; + } + /* joining only indented stuff after empty lines; + * note that now we only require 1 space of indentation + * to continue a list */ + else if (in_empty && pre == 0) { + *flags |= HOEDOWN_LI_END; + break; + } + + if (in_empty) { + hoedown_buffer_putc(work, '\n'); + has_inside_empty = 1; + in_empty = 0; + } + + /* adding the line without prefix into the working buffer */ + hoedown_buffer_put(work, data + beg + i, end - beg - i); + beg = end; + } + + /* render of li contents */ + if (has_inside_empty) + *flags |= HOEDOWN_LI_BLOCK; + + if (*flags & HOEDOWN_LI_BLOCK) { + /* intermediate render of block li */ + if (sublist && sublist < work->size) { + parse_block(inter, doc, work->data, sublist); + parse_block(inter, doc, work->data + sublist, work->size - sublist); + } + else + parse_block(inter, doc, work->data, work->size); + } else { + /* intermediate render of inline li */ + if (sublist && sublist < work->size) { + parse_inline(inter, doc, work->data, sublist); + parse_block(inter, doc, work->data + sublist, work->size - sublist); + } + else + parse_inline(inter, doc, work->data, work->size); + } + + /* render of li itself */ + if (doc->md.listitem) + doc->md.listitem(ob, inter, *flags, &doc->data); + + popbuf(doc, BUFFER_SPAN); + popbuf(doc, BUFFER_SPAN); + return beg; +} + + +/* parse_list • parsing ordered or unordered list block */ +static size_t +parse_list(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size, hoedown_list_flags flags) +{ + hoedown_buffer *work = 0; + size_t i = 0, j; + + work = newbuf(doc, BUFFER_BLOCK); + + while (i < size) { + j = parse_listitem(work, doc, data + i, size - i, &flags); + i += j; + + if (!j || (flags & HOEDOWN_LI_END)) + break; + } + + if (doc->md.list) + doc->md.list(ob, work, flags, &doc->data); + popbuf(doc, BUFFER_BLOCK); + return i; +} + +/* parse_atxheader • parsing of atx-style headers */ +static size_t +parse_atxheader(hoedown_buffer *ob, hoedown_document *doc, uint8_t *data, size_t size) +{ + size_t level = 0; + size_t i, end, skip; + + while (level < size && level < 6 && data[level] == '#') + level++; + + for (i = level; i < size && data[i] == ' '; i++); + + for (end = i; end < size && data[end] != '\n'; end++); + skip = end; + + while (end && data[end - 1] == '#') + end--; + + while (end && data[end - 1] == ' ') + end--; + + if (end > i) { + hoedown_buffer *work = newbuf(doc, BUFFER_SPAN); + + parse_inline(work, doc, data + i, end - i); + + if (doc->md.header) + doc->md.header(ob, work, (int)level, &doc->data); + + popbuf(doc, BUFFER_SPAN); + } + + return skip; +} + +/* parse_footnote_def • parse a single footnote definition */ +static void +parse_footnote_def(hoedown_buffer *ob, hoedown_document *doc, unsigned int num, uint8_t *data, size_t size) +{ + hoedown_buffer *work = 0; + work = newbuf(doc, BUFFER_SPAN); + + parse_block(work, doc, data, size); + + if (doc->md.footnote_def) + doc->md.footnote_def(ob, work, num, &doc->data); + popbuf(doc, BUFFER_SPAN); +} + +/* parse_footnote_list • render the contents of the footnotes */ +static void +parse_footnote_list(hoedown_buffer *ob, hoedown_document *doc, struct footnote_list *footnotes) +{ + hoedown_buffer *work = 0; + struct footnote_item *item; + struct footnote_ref *ref; + + if (footnotes->count == 0) + return; + + work = newbuf(doc, BUFFER_BLOCK); + + item = footnotes->head; + while (item) { + ref = item->ref; + parse_footnote_def(work, doc, ref->num, ref->contents->data, ref->contents->size); + item = item->next; + } + + if (doc->md.footnotes) + doc->md.footnotes(ob, work, &doc->data); + popbuf(doc, BUFFER_BLOCK); +} + +/* htmlblock_is_end • check for end of HTML block : ( *)\n */ +/* returns tag length on match, 0 otherwise */ +/* assumes data starts with "<" */ +static size_t +htmlblock_is_end( + const char *tag, + size_t tag_len, + hoedown_document *doc, + uint8_t *data, + size_t size) +{ + size_t i = tag_len + 3, w; + + /* try to match the end tag */ + /* note: we're not considering tags like "" which are still valid */ + if (i > size || + data[1] != '/' || + strncasecmp((char *)data + 2, tag, tag_len) != 0 || + data[tag_len + 2] != '>') + return 0; + + /* rest of the line must be empty */ + if ((w = is_empty(data + i, size - i)) == 0 && i < size) + return 0; + + return i + w; +} + +/* htmlblock_find_end • try to find HTML block ending tag */ +/* returns the length on match, 0 otherwise */ +static size_t +htmlblock_find_end( + const char *tag, + size_t tag_len, + hoedown_document *doc, + uint8_t *data, + size_t size) +{ + size_t i = 0, w; + + while (1) { + while (i < size && data[i] != '<') i++; + if (i >= size) return 0; + + w = htmlblock_is_end(tag, tag_len, doc, data + i, size - i); + if (w) return i + w; + i++; + } +} + +/* htmlblock_find_end_strict • try to find end of HTML block in strict mode */ +/* (it must be an unindented line, and have a blank line afterwads) */ +/* returns the length on match, 0 otherwise */ +static size_t +htmlblock_fin