commonmarker 0.17.13 → 0.23.4

data/ext/commonmarker/blocks.c

@@ -36,6 +36,10 @@ static bool S_last_line_blank(const cmark_node *node) {
   return (node->flags & CMARK_NODE__LAST_LINE_BLANK) != 0;
 }
 
+static bool S_last_line_checked(const cmark_node *node) {
+  return (node->flags & CMARK_NODE__LAST_LINE_CHECKED) != 0;
+}
+
 static CMARK_INLINE cmark_node_type S_type(const cmark_node *node) {
   return (cmark_node_type)node->type;
 }
@@ -47,6 +51,10 @@ static void S_set_last_line_blank(cmark_node *node, bool is_blank) {
     node->flags &= ~CMARK_NODE__LAST_LINE_BLANK;
 }
 
+static void S_set_last_line_checked(cmark_node *node) {
+  node->flags |= CMARK_NODE__LAST_LINE_CHECKED;
+}
+
 static CMARK_INLINE bool S_is_line_end_char(char c) {
   return (c == '\n' || c == '\r');
 }
@@ -121,8 +129,6 @@ static void cmark_parser_reset(cmark_parser *parser) {
   parser->root = document;
   parser->current = document;
 
-  parser->last_buffer_ended_with_cr = false;
-
   parser->syntax_extensions = saved_exts;
   parser->inline_syntax_extensions = saved_inline_exts;
   parser->options = saved_options;
@@ -234,19 +240,35 @@ static void remove_trailing_blank_lines(cmark_strbuf *ln) {
 
 // Check to see if a node ends with a blank line, descending
 // if needed into lists and sublists.
-static bool ends_with_blank_line(cmark_node *node) {
-  cmark_node *cur = node;
-  while (cur != NULL) {
-    if (S_last_line_blank(cur)) {
-      return true;
-    }
-    if (S_type(cur) == CMARK_NODE_LIST || S_type(cur) == CMARK_NODE_ITEM) {
-      cur = cur->last_child;
-    } else {
-      cur = NULL;
-    }
+static bool S_ends_with_blank_line(cmark_node *node) {
+  if (S_last_line_checked(node)) {
+    return(S_last_line_blank(node));
+  } else if ((S_type(node) == CMARK_NODE_LIST ||
+              S_type(node) == CMARK_NODE_ITEM) && node->last_child) {
+    S_set_last_line_checked(node);
+    return(S_ends_with_blank_line(node->last_child));
+  } else {
+    S_set_last_line_checked(node);
+    return (S_last_line_blank(node));
   }
-  return false;
+}
+
+// returns true if content remains after link defs are resolved.
+static bool resolve_reference_link_definitions(
+		cmark_parser *parser,
+                cmark_node *b) {
+  bufsize_t pos;
+  cmark_strbuf *node_content = &b->content;
+  cmark_chunk chunk = {node_content->ptr, node_content->size, 0};
+  while (chunk.len && chunk.data[0] == '[' &&
+         (pos = cmark_parse_reference_inline(parser->mem, &chunk,
+					     parser->refmap))) {
+
+    chunk.data += pos;
+    chunk.len -= pos;
+  }
+  cmark_strbuf_drop(node_content, (node_content->size - chunk.len));
+  return !is_blank(&b->content, 0);
 }
 
 static cmark_node *finalize(cmark_parser *parser, cmark_node *b) {
@@ -254,6 +276,7 @@ static cmark_node *finalize(cmark_parser *parser, cmark_node *b) {
   cmark_node *item;
   cmark_node *subitem;
   cmark_node *parent;
+  bool has_content;
 
   parent = b->parent;
   assert(b->flags &
@@ -283,15 +306,8 @@ static cmark_node *finalize(cmark_parser *parser, cmark_node *b) {
   switch (S_type(b)) {
   case CMARK_NODE_PARAGRAPH:
   {
-    cmark_chunk chunk = {node_content->ptr, node_content->size, 0};
-    while (chunk.len && chunk.data[0] == '[' &&
-           (pos = cmark_parse_reference_inline(parser->mem, &chunk, parser->refmap))) {
-
-      chunk.data += pos;
-      chunk.len -= pos;
-    }
-    cmark_strbuf_drop(node_content, (node_content->size - chunk.len));
-    if (is_blank(node_content, 0)) {
+    has_content = resolve_reference_link_definitions(parser, b);
+    if (!has_content) {
       // remove blank node (former reference def)
       cmark_node_free(b);
     }
@@ -343,7 +359,8 @@ static cmark_node *finalize(cmark_parser *parser, cmark_node *b) {
       // spaces between them:
       subitem = item->first_child;
       while (subitem) {
-        if (ends_with_blank_line(subitem) && (item->next || subitem->next)) {
+        if ((item->next || subitem->next) &&
+            S_ends_with_blank_line(subitem)) {
           b->as.list.tight = false;
           break;
         }
@@ -451,7 +468,6 @@ static void process_footnotes(cmark_parser *parser) {
   while ((ev_type = cmark_iter_next(iter)) != CMARK_EVENT_DONE) {
     cur = cmark_iter_get_node(iter);
     if (ev_type == CMARK_EVENT_EXIT && cur->type == CMARK_NODE_FOOTNOTE_DEFINITION) {
-      cmark_node_unlink(cur);
       cmark_footnote_create(map, cur);
     }
   }
@@ -468,6 +484,15 @@ static void process_footnotes(cmark_parser *parser) {
         if (!footnote->ix)
           footnote->ix = ++ix;
 
+        // store a reference to this footnote reference's footnote definition
+        // this is used by renderers when generating label ids
+        cur->parent_footnote_def = footnote->node;
+
+        // keep track of a) count of how many times this footnote def has been
+        // referenced, and b) which reference index this footnote ref is at.
+        // this is used by renderers when generating links and backreferences.
+        cur->footnote.ref_ix = ++footnote->node->footnote.def_count;
+
         char n[32];
         snprintf(n, sizeof(n), "%d", footnote->ix);
         cmark_chunk_free(parser->mem, &cur->as.literal);
@@ -498,13 +523,16 @@ static void process_footnotes(cmark_parser *parser) {
     qsort(map->sorted, map->size, sizeof(cmark_map_entry *), sort_footnote_by_ix);
     for (unsigned int i = 0; i < map->size; ++i) {
       cmark_footnote *footnote = (cmark_footnote *)map->sorted[i];
-      if (!footnote->ix)
+      if (!footnote->ix) {
+        cmark_node_unlink(footnote->node);
         continue;
+      }
       cmark_node_append_child(parser->root, footnote->node);
       footnote->node = NULL;
     }
   }
 
+  cmark_unlink_footnotes_map(map);
   cmark_map_free(map);
 }
 
@@ -748,6 +776,40 @@ static void chop_trailing_hashtags(cmark_chunk *ch) {
   }
 }
 
+// Check for thematic break.  On failure, return 0 and update
+// thematic_break_kill_pos with the index at which the
+// parse fails.  On success, return length of match.
+// "...three or more hyphens, asterisks,
+// or underscores on a line by themselves. If you wish, you may use
+// spaces between the hyphens or asterisks."
+static int S_scan_thematic_break(cmark_parser *parser, cmark_chunk *input,
+                                 bufsize_t offset) {
+  bufsize_t i;
+  char c;
+  char nextc = '\0';
+  int count;
+  i = offset;
+  c = peek_at(input, i);
+  if (!(c == '*' || c == '_' || c == '-')) {
+    parser->thematic_break_kill_pos = i;
+    return 0;
+  }
+  count = 1;
+  while ((nextc = peek_at(input, ++i))) {
+    if (nextc == c) {
+      count++;
+    } else if (nextc != ' ' && nextc != '\t') {
+      break;
+    }
+  }
+  if (count >= 3 && (nextc == '\r' || nextc == '\n')) {
+    return (i - offset) + 1;
+  } else {
+    parser->thematic_break_kill_pos = i;
+    return 0;
+  }
+}
+
 // Find first nonspace character from current offset, setting
 // parser->first_nonspace, parser->first_nonspace_column,
 // parser->indent, and parser->blank. Does not advance parser->offset.
@@ -1040,6 +1102,7 @@ static void open_new_blocks(cmark_parser *parser, cmark_node **container,
   bufsize_t matched = 0;
   int lev = 0;
   bool save_partially_consumed_tab;
+  bool has_content;
   int save_offset;
   int save_column;
 
@@ -1112,13 +1175,20 @@ static void open_new_blocks(cmark_parser *parser, cmark_node **container,
     } else if (!indented && cont_type == CMARK_NODE_PARAGRAPH &&
                (lev =
                     scan_setext_heading_line(input, parser->first_nonspace))) {
-      (*container)->type = (uint16_t)CMARK_NODE_HEADING;
-      (*container)->as.heading.level = lev;
-      (*container)->as.heading.setext = true;
-      S_advance_offset(parser, input, input->len - 1 - parser->offset, false);
+      // finalize paragraph, resolving reference links
+      has_content = resolve_reference_link_definitions(parser, *container);
+
+      if (has_content) {
+
+        (*container)->type = (uint16_t)CMARK_NODE_HEADING;
+        (*container)->as.heading.level = lev;
+        (*container)->as.heading.setext = true;
+        S_advance_offset(parser, input, input->len - 1 - parser->offset, false);
+      }
     } else if (!indented &&
                !(cont_type == CMARK_NODE_PARAGRAPH && !all_matched) &&
-               (matched = scan_thematic_break(input, parser->first_nonspace))) {
+	       (parser->thematic_break_kill_pos <= parser->first_nonspace) &&
+               (matched = S_scan_thematic_break(parser, input, parser->first_nonspace))) {
       // it's only now that we know the line is not part of a setext heading:
       *container = add_child(parser, *container, CMARK_NODE_THEMATIC_BREAK,
                              parser->first_nonspace + 1);
@@ -1377,6 +1447,7 @@ static void S_process_line(cmark_parser *parser, const unsigned char *buffer,
   parser->column = 0;
   parser->first_nonspace = 0;
   parser->first_nonspace_column = 0;
+  parser->thematic_break_kill_pos = 0;
   parser->indent = 0;
   parser->blank = false;
   parser->partially_consumed_tab = false;

data/ext/commonmarker/buffer.c

@@ -10,7 +10,6 @@
 #include "config.h"
 #include "cmark_ctype.h"
 #include "buffer.h"
-#include "memory.h"
 
 /* Used as default value for cmark_strbuf->ptr so that people can always
  * assume ptr is non-NULL and zero terminated even for new cmark_strbufs.

data/ext/commonmarker/chunk.h

@@ -6,7 +6,6 @@
 #include <assert.h>
 #include "cmark-gfm.h"
 #include "buffer.h"
-#include "memory.h"
 #include "cmark_ctype.h"
 
 #define CMARK_CHUNK_EMPTY                                                      \

data/ext/commonmarker/cmark-gfm-core-extensions.h

@@ -7,6 +7,7 @@ extern "C" {
 
 #include "cmark-gfm-extension_api.h"
 #include "cmark-gfm-extensions_export.h"
+#include "config.h" // for bool
 #include <stdint.h>
 
 CMARK_GFM_EXTENSIONS_EXPORT
@@ -15,9 +16,37 @@ void cmark_gfm_core_extensions_ensure_registered(void);
 CMARK_GFM_EXTENSIONS_EXPORT
 uint16_t cmark_gfm_extensions_get_table_columns(cmark_node *node);
 
+/** Sets the number of columns for the table, returning 1 on success and 0 on error.
+ */
+CMARK_GFM_EXTENSIONS_EXPORT
+int cmark_gfm_extensions_set_table_columns(cmark_node *node, uint16_t n_columns);
+
 CMARK_GFM_EXTENSIONS_EXPORT
 uint8_t *cmark_gfm_extensions_get_table_alignments(cmark_node *node);
 
+/** Sets the alignments for the table, returning 1 on success and 0 on error.
+ */
+CMARK_GFM_EXTENSIONS_EXPORT
+int cmark_gfm_extensions_set_table_alignments(cmark_node *node, uint16_t ncols, uint8_t *alignments);
+
+CMARK_GFM_EXTENSIONS_EXPORT
+int cmark_gfm_extensions_get_table_row_is_header(cmark_node *node);
+
+/** Sets whether the node is a table header row, returning 1 on success and 0 on error.
+ */
+CMARK_GFM_EXTENSIONS_EXPORT
+int cmark_gfm_extensions_set_table_row_is_header(cmark_node *node, int is_header);
+
+CMARK_GFM_EXTENSIONS_EXPORT
+bool cmark_gfm_extensions_get_tasklist_item_checked(cmark_node *node);
+/* For backwards compatibility */
+#define cmark_gfm_extensions_tasklist_is_checked cmark_gfm_extensions_get_tasklist_item_checked
+
+/** Sets whether a tasklist item is "checked" (completed), returning 1 on success and 0 on error.
+ */
+CMARK_GFM_EXTENSIONS_EXPORT
+int cmark_gfm_extensions_set_tasklist_item_checked(cmark_node *node, bool is_checked);
+
 #ifdef __cplusplus
 }
 #endif

data/ext/commonmarker/cmark-gfm-extension_api.h

@@ -106,8 +106,6 @@ typedef struct cmark_plugin cmark_plugin;
  * with 'cmark_syntax_extension_set_private',
  * and optionally define a free function for this data.
  */
-typedef struct cmark_syntax_extension cmark_syntax_extension;
-
 typedef struct subject cmark_inline_parser;
 
 /** Exposed raw for now */
@@ -238,6 +236,9 @@ typedef int (*cmark_commonmark_escape_func) (cmark_syntax_extension *extension,
                                               cmark_node *node,
                                               int c);
 
+typedef const char* (*cmark_xml_attr_func) (cmark_syntax_extension *extension,
+                                            cmark_node *node);
+
 typedef void (*cmark_html_render_func) (cmark_syntax_extension *extension,
                                         struct cmark_html_renderer *renderer,
                                         cmark_node *node,
@@ -254,6 +255,10 @@ typedef cmark_node *(*cmark_postprocess_func) (cmark_syntax_extension *extension
 
 typedef int (*cmark_ispunct_func) (char c);
 
+typedef void (*cmark_opaque_alloc_func) (cmark_syntax_extension *extension,
+                                         cmark_mem *mem,
+                                         cmark_node *node);
+
 typedef void (*cmark_opaque_free_func) (cmark_syntax_extension *extension,
                                         cmark_mem *mem,
                                         cmark_node *node);
@@ -343,6 +348,12 @@ void cmark_syntax_extension_set_latex_render_func(cmark_syntax_extension *extens
 /** See the documentation for 'cmark_syntax_extension'
  */
 CMARK_GFM_EXPORT
+void cmark_syntax_extension_set_xml_attr_func(cmark_syntax_extension *extension,
+                                              cmark_xml_attr_func func);
+
+  /** See the documentation for 'cmark_syntax_extension'
+ */
+CMARK_GFM_EXPORT
 void cmark_syntax_extension_set_man_render_func(cmark_syntax_extension *extension,
                                                 cmark_common_render_func func);
 
@@ -382,6 +393,12 @@ CMARK_GFM_EXPORT
 void cmark_syntax_extension_set_postprocess_func(cmark_syntax_extension *extension,
                                                  cmark_postprocess_func func);
 
+/** See the documentation for 'cmark_syntax_extension'
+ */
+CMARK_GFM_EXPORT
+void cmark_syntax_extension_set_opaque_alloc_func(cmark_syntax_extension *extension,
+                                                  cmark_opaque_alloc_func func);
+
 /** See the documentation for 'cmark_syntax_extension'
  */
 CMARK_GFM_EXPORT

data/ext/commonmarker/cmark-gfm.h

@@ -92,6 +92,7 @@ typedef enum {
 typedef struct cmark_node cmark_node;
 typedef struct cmark_parser cmark_parser;
 typedef struct cmark_iter cmark_iter;
+typedef struct cmark_syntax_extension cmark_syntax_extension;
 
 /**
  * ## Custom memory allocator support
@@ -187,6 +188,13 @@ CMARK_GFM_EXPORT cmark_node *cmark_node_new(cmark_node_type type);
 CMARK_GFM_EXPORT cmark_node *cmark_node_new_with_mem(cmark_node_type type,
                                                  cmark_mem *mem);
 
+CMARK_GFM_EXPORT cmark_node *cmark_node_new_with_ext(cmark_node_type type,
+                                                cmark_syntax_extension *extension);
+
+CMARK_GFM_EXPORT cmark_node *cmark_node_new_with_mem_and_ext(cmark_node_type type,
+                                                cmark_mem *mem,
+                                                cmark_syntax_extension *extension);
+
 /** Frees the memory allocated for a node and any children.
  */
 CMARK_GFM_EXPORT void cmark_node_free(cmark_node *node);
@@ -682,14 +690,20 @@ char *cmark_render_latex_with_mem(cmark_node *root, int options, int width, cmar
  */
 #define CMARK_OPT_HARDBREAKS (1 << 2)
 
-/** Suppress raw HTML and unsafe links (`javascript:`, `vbscript:`,
- * `file:`, and `data:`, except for `image/png`, `image/gif`,
- * `image/jpeg`, or `image/webp` mime types).  Raw HTML is replaced
- * by a placeholder HTML comment. Unsafe links are replaced by
- * empty strings.
+/** `CMARK_OPT_SAFE` is defined here for API compatibility,
+    but it no longer has any effect. "Safe" mode is now the default:
+    set `CMARK_OPT_UNSAFE` to disable it.
  */
 #define CMARK_OPT_SAFE (1 << 3)
 
+/** Render raw HTML and unsafe links (`javascript:`, `vbscript:`,
+ * `file:`, and `data:`, except for `image/png`, `image/gif`,
+ * `image/jpeg`, or `image/webp` mime types).  By default,
+ * raw HTML is replaced by a placeholder HTML comment. Unsafe
+ * links are replaced by empty strings.
+ */
+#define CMARK_OPT_UNSAFE (1 << 17)
+
 /** Render `softbreak` elements as spaces.
  */
 #define CMARK_OPT_NOBREAKS (1 << 4)

data/ext/commonmarker/cmark-gfm_version.h

@@ -1,7 +1,7 @@
 #ifndef CMARK_GFM_VERSION_H
 #define CMARK_GFM_VERSION_H
 
-#define CMARK_GFM_VERSION ((0 << 24) | (28 << 16) | (3 << 8) | 16)
-#define CMARK_GFM_VERSION_STRING "0.28.3.gfm.16"
+#define CMARK_GFM_VERSION ((0 << 24) | (29 << 16) | (0 << 8) | 3)
+#define CMARK_GFM_VERSION_STRING "0.29.0.gfm.3"
 
 #endif

data/ext/commonmarker/commonmark.c

@@ -34,7 +34,8 @@ static CMARK_INLINE void outc(cmark_renderer *renderer, cmark_node *node,
   needs_escaping =
       c < 0x80 && escape != LITERAL &&
       ((escape == NORMAL &&
-        (c == '*' || c == '_' || c == '[' || c == ']' || c == '#' || c == '<' ||
+        (c < 0x20 ||
+	 c == '*' || c == '_' || c == '[' || c == ']' || c == '#' || c == '<' ||
          c == '>' || c == '\\' || c == '`' || c == '~' || c == '!' ||
          (c == '&' && cmark_isalpha(nextc)) || (c == '!' && nextc == '[') ||
          (renderer->begin_content && (c == '-' || c == '+' || c == '=') &&
@@ -50,14 +51,18 @@ static CMARK_INLINE void outc(cmark_renderer *renderer, cmark_node *node,
         (c == '`' || c == '<' || c == '>' || c == '"' || c == '\\')));
 
   if (needs_escaping) {
-    if (cmark_isspace((char)c)) {
+    if (escape == URL && cmark_isspace((char)c)) {
       // use percent encoding for spaces
-      snprintf(encoded, ENCODED_SIZE, "%%%2x", c);
+      snprintf(encoded, ENCODED_SIZE, "%%%2X", c);
       cmark_strbuf_puts(renderer->buffer, encoded);
       renderer->column += 3;
-    } else {
+    } else if (cmark_ispunct((char)c)) {
       cmark_render_ascii(renderer, "\\");
       cmark_render_code_point(renderer, c);
+    } else { // render as entity
+      snprintf(encoded, ENCODED_SIZE, "&#%d;", c);
+      cmark_strbuf_puts(renderer->buffer, encoded);
+      renderer->column += (int)strlen(encoded);
     }
   } else {
     cmark_render_code_point(renderer, c);
@@ -168,9 +173,11 @@ static int S_render_node(cmark_renderer *renderer, cmark_node *node,
   int list_number;
   cmark_delim_type list_delim;
   int numticks;
+  bool extra_spaces;
   int i;
   bool entering = (ev_type == CMARK_EVENT_ENTER);
   const char *info, *code, *title;
+  char fencechar[2] = {'\0', '\0'};
   size_t info_len, code_len;
   char listmarker[LISTMARKER_SIZE];
   char *emph_delim;
@@ -283,6 +290,7 @@ static int S_render_node(cmark_renderer *renderer, cmark_node *node,
     }
     info = cmark_node_get_fence_info(node);
     info_len = strlen(info);
+    fencechar[0] = strchr(info, '`') == NULL ? '`' : '~';
     code = cmark_node_get_literal(node);
     code_len = strlen(code);
     // use indented form if no info, and code doesn't
@@ -302,7 +310,7 @@ static int S_render_node(cmark_renderer *renderer, cmark_node *node,
         numticks = 3;
       }
       for (i = 0; i < numticks; i++) {
-        LIT("`");
+        LIT(fencechar);
       }
       LIT(" ");
       OUT(info, false, LITERAL);
@@ -310,7 +318,7 @@ static int S_render_node(cmark_renderer *renderer, cmark_node *node,
       OUT(cmark_node_get_literal(node), false, LITERAL);
       CR();
       for (i = 0; i < numticks; i++) {
-        LIT("`");
+        LIT(fencechar);
       }
     }
     BLANKLINE();
@@ -369,14 +377,17 @@ static int S_render_node(cmark_renderer *renderer, cmark_node *node,
     code = cmark_node_get_literal(node);
     code_len = strlen(code);
     numticks = shortest_unused_backtick_sequence(code);
+    extra_spaces = code_len == 0 ||
+	    code[0] == '`' || code[code_len - 1] == '`' ||
+	    code[0] == ' ' || code[code_len - 1] == ' ';
     for (i = 0; i < numticks; i++) {
       LIT("`");
     }
-    if (code_len == 0 || code[0] == '`') {
+    if (extra_spaces) {
       LIT(" ");
     }
     OUT(cmark_node_get_literal(node), allow_wrap, LITERAL);
-    if (code_len == 0 || code[code_len - 1] == '`') {
+    if (extra_spaces) {
       LIT(" ");
     }
     for (i = 0; i < numticks; i++) {
@@ -466,7 +477,13 @@ static int S_render_node(cmark_renderer *renderer, cmark_node *node,
   case CMARK_NODE_FOOTNOTE_REFERENCE:
     if (entering) {
       LIT("[^");
-      OUT(cmark_chunk_to_cstr(renderer->mem, &node->as.literal), false, LITERAL);
+
+      char *footnote_label = renderer->mem->calloc(node->parent_footnote_def->as.literal.len + 1, sizeof(char));
+      memmove(footnote_label, node->parent_footnote_def->as.literal.data, node->parent_footnote_def->as.literal.len);
+
+      OUT(footnote_label, false, LITERAL);
+      renderer->mem->free(footnote_label);
+
       LIT("]");
     }
     break;
@@ -475,9 +492,13 @@ static int S_render_node(cmark_renderer *renderer, cmark_node *node,
     if (entering) {
       renderer->footnote_ix += 1;
       LIT("[^");
-      char n[32];
-      snprintf(n, sizeof(n), "%d", renderer->footnote_ix);
-      OUT(n, false, LITERAL);
+
+      char *footnote_label = renderer->mem->calloc(node->as.literal.len + 1, sizeof(char));
+      memmove(footnote_label, node->as.literal.data, node->as.literal.len);
+
+      OUT(footnote_label, false, LITERAL);
+      renderer->mem->free(footnote_label);
+
       LIT("]:\n");
 
       cmark_strbuf_puts(renderer->prefix, "    ");