commonmarker 0.0.1

data/ext/commonmarker/cmark/test/spec_tests.py

@@ -0,0 +1,139 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+import sys
+from difflib import unified_diff
+import argparse
+import re
+import json
+from cmark import CMark
+from normalize import normalize_html
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description='Run cmark tests.')
+    parser.add_argument('-p', '--program', dest='program', nargs='?', default=None,
+            help='program to test')
+    parser.add_argument('-s', '--spec', dest='spec', nargs='?', default='test/spec.txt',
+            help='path to spec')
+    parser.add_argument('-P', '--pattern', dest='pattern', nargs='?',
+            default=None, help='limit to sections matching regex pattern')
+    parser.add_argument('--library-dir', dest='library_dir', nargs='?',
+            default=None, help='directory containing dynamic library')
+    parser.add_argument('--no-normalize', dest='normalize',
+            action='store_const', const=False, default=True,
+            help='do not normalize HTML')
+    parser.add_argument('-d', '--dump-tests', dest='dump_tests',
+            action='store_const', const=True, default=False,
+            help='dump tests in JSON format')
+    parser.add_argument('--debug-normalization', dest='debug_normalization',
+            action='store_const', const=True,
+            default=False, help='filter stdin through normalizer for testing')
+    parser.add_argument('-n', '--number', type=int, default=None,
+            help='only consider the test with the given number')
+    args = parser.parse_args(sys.argv[1:])
+
+def print_test_header(headertext, example_number, start_line, end_line):
+    print("Example %d (lines %d-%d) %s" % (example_number,start_line,end_line,headertext))
+
+def do_test(test, normalize, result_counts):
+    [retcode, actual_html, err] = cmark.to_html(test['markdown'])
+    if retcode == 0:
+        expected_html = test['html']
+        unicode_error = None
+        if normalize:
+            try:
+                passed = normalize_html(actual_html) == normalize_html(expected_html)
+            except UnicodeDecodeError as e:
+                unicode_error = e
+                passed = False
+        else:
+            passed = actual_html == expected_html
+        if passed:
+            result_counts['pass'] += 1
+        else:
+            print_test_header(test['section'], test['example'], test['start_line'], test['end_line'])
+            sys.stdout.write(test['markdown'])
+            if unicode_error:
+                print("Unicode error: " + str(unicode_error))
+                print("Expected: " + repr(expected_html))
+                print("Got:      " + repr(actual_html))
+            else:
+                expected_html_lines = expected_html.splitlines(True)
+                actual_html_lines = actual_html.splitlines(True)
+                for diffline in unified_diff(expected_html_lines, actual_html_lines,
+                                "expected HTML", "actual HTML"):
+                    sys.stdout.write(diffline)
+            sys.stdout.write('\n')
+            result_counts['fail'] += 1
+    else:
+        print_test_header(test['section'], test['example'], test['start_line'], test['end_line'])
+        print("program returned error code %d" % retcode)
+        print(err)
+        result_counts['error'] += 1
+
+def get_tests(specfile):
+    line_number = 0
+    start_line = 0
+    end_line = 0
+    example_number = 0
+    markdown_lines = []
+    html_lines = []
+    state = 0  # 0 regular text, 1 markdown example, 2 html output
+    headertext = ''
+    tests = []
+
+    header_re = re.compile('#+ ')
+
+    with open(specfile, 'r', encoding='utf-8') as specf:
+        for line in specf:
+            line_number = line_number + 1
+            if state == 0 and re.match(header_re, line):
+                headertext = header_re.sub('', line).strip()
+            if line.strip() == ".":
+                state = (state + 1) % 3
+                if state == 0:
+                    example_number = example_number + 1
+                    end_line = line_number
+                    tests.append({
+                        "markdown":''.join(markdown_lines).replace('→',"\t"),
+                        "html":''.join(html_lines),
+                        "example": example_number,
+                        "start_line": start_line,
+                        "end_line": end_line,
+                        "section": headertext})
+                    start_line = 0
+                    markdown_lines = []
+                    html_lines = []
+            elif state == 1:
+                if start_line == 0:
+                    start_line = line_number - 1
+                markdown_lines.append(line)
+            elif state == 2:
+                html_lines.append(line)
+    return tests
+
+if __name__ == "__main__":
+    if args.debug_normalization:
+        print(normalize_html(sys.stdin.read()))
+        exit(0)
+
+    all_tests = get_tests(args.spec)
+    if args.pattern:
+        pattern_re = re.compile(args.pattern, re.IGNORECASE)
+    else:
+        pattern_re = re.compile('.')
+    tests = [ test for test in all_tests if re.search(pattern_re, test['section']) and (not args.number or test['example'] == args.number) ]
+    if args.dump_tests:
+        print(json.dumps(tests, ensure_ascii=False, indent=2))
+        exit(0)
+    else:
+        skipped = len(all_tests) - len(tests)
+        cmark = CMark(prog=args.program, library_dir=args.library_dir)
+        result_counts = {'pass': 0, 'fail': 0, 'error': 0, 'skip': skipped}
+        for test in tests:
+            do_test(test, args.normalize, result_counts)
+        print("{pass} passed, {fail} failed, {error} errored, {skip} skipped".format(**result_counts))
+        if result_counts['fail'] == 0 and result_counts['error'] == 0:
+            exit(0)
+        else:
+            exit(1)

data/ext/commonmarker/cmark/toolchain-mingw32.cmake

@@ -0,0 +1,17 @@
+# the name of the target operating system
+SET(CMAKE_SYSTEM_NAME Windows)
+
+# which compilers to use for C and C++
+SET(CMAKE_C_COMPILER i586-mingw32msvc-gcc)
+SET(CMAKE_CXX_COMPILER i586-mingw32msvc-g++)
+SET(CMAKE_RC_COMPILER i586-mingw32msvc-windres)
+
+# here is the target environment located
+SET(CMAKE_FIND_ROOT_PATH  /usr/i586-mingw32msvc "${CMAKE_SOURCE_DIR}/windows")
+
+# adjust the default behaviour of the FIND_XXX() commands:
+# search headers and libraries in the target environment, search 
+# programs in the host environment
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE ONLY)

data/ext/commonmarker/cmark/tools/mkcasefold.pl

@@ -0,0 +1,22 @@
+binmode STDOUT;
+print("    switch (c) {\n");
+my $lastchar = "";
+while (<STDIN>) {
+  if (/^[A-F0-9]/ and / [CF]; /) {
+    my ($char, $type, $subst) = m/([A-F0-9]+); ([CF]); ([^;]+)/;
+    if ($char eq $lastchar) {
+      break;
+    }
+    my @subst = $subst =~ m/(\w+)/g;
+    printf("      case 0x%s:\n", $char);
+    foreach (@subst) {
+      printf("        bufpush(0x%s);\n", $_);
+    }
+    printf("        break;\n");
+    $lastchar = $char;
+  }
+}
+printf("      default:\n");
+printf("        bufpush(c);\n");
+print("    }\n");
+

data/ext/commonmarker/cmark/wrappers/wrapper.lua

@@ -0,0 +1,239 @@
+#!/usr/bin/env luajit
+
+local ffi = require("ffi")
+
+cmark = ffi.load("libcmark")
+
+ffi.cdef[[
+char *cmark_markdown_to_html(const char *text, int len, int options);
+
+typedef enum {
+	/* Error status */
+	CMARK_NODE_NONE,
+
+	/* Block */
+	CMARK_NODE_DOCUMENT,
+	CMARK_NODE_BLOCK_QUOTE,
+	CMARK_NODE_LIST,
+	CMARK_NODE_ITEM,
+	CMARK_NODE_CODE_BLOCK,
+	CMARK_NODE_HTML,
+	CMARK_NODE_PARAGRAPH,
+	CMARK_NODE_HEADER,
+	CMARK_NODE_HRULE,
+
+	CMARK_NODE_FIRST_BLOCK = CMARK_NODE_DOCUMENT,
+	CMARK_NODE_LAST_BLOCK  = CMARK_NODE_HRULE,
+
+	/* Inline */
+	CMARK_NODE_TEXT,
+	CMARK_NODE_SOFTBREAK,
+	CMARK_NODE_LINEBREAK,
+	CMARK_NODE_CODE,
+	CMARK_NODE_INLINE_HTML,
+	CMARK_NODE_EMPH,
+	CMARK_NODE_STRONG,
+	CMARK_NODE_LINK,
+	CMARK_NODE_IMAGE,
+
+	CMARK_NODE_FIRST_INLINE = CMARK_NODE_TEXT,
+	CMARK_NODE_LAST_INLINE  = CMARK_NODE_IMAGE,
+} cmark_node_type;
+
+
+typedef enum {
+	CMARK_NO_LIST,
+	CMARK_BULLET_LIST,
+	CMARK_ORDERED_LIST
+}  cmark_list_type;
+
+typedef enum {
+	CMARK_NO_DELIM,
+	CMARK_PERIOD_DELIM,
+	CMARK_PAREN_DELIM
+} cmark_delim_type;
+
+typedef struct cmark_node cmark_node;
+typedef struct cmark_parser cmark_parser;
+typedef struct cmark_iter cmark_iter;
+
+typedef enum {
+	CMARK_EVENT_NONE,
+	CMARK_EVENT_DONE,
+	CMARK_EVENT_ENTER,
+	CMARK_EVENT_EXIT
+} cmark_event_type;
+
+cmark_node*
+cmark_node_new(cmark_node_type type);
+
+void
+cmark_node_free(cmark_node *node);
+
+cmark_node*
+cmark_node_next(cmark_node *node);
+
+cmark_node*
+cmark_node_previous(cmark_node *node);
+
+cmark_node*
+cmark_node_parent(cmark_node *node);
+
+cmark_node*
+cmark_node_first_child(cmark_node *node);
+
+cmark_node*
+cmark_node_last_child(cmark_node *node);
+
+cmark_iter*
+cmark_iter_new(cmark_node *root);
+
+void
+cmark_iter_free(cmark_iter *iter);
+
+cmark_event_type
+cmark_iter_next(cmark_iter *iter);
+
+cmark_node*
+cmark_iter_get_node(cmark_iter *iter);
+
+cmark_event_type
+cmark_iter_get_event_type(cmark_iter *iter);
+
+cmark_node*
+cmark_iter_get_root(cmark_iter *iter);
+
+void
+cmark_iter_reset(cmark_iter *iter, cmark_node *current,
+                 cmark_event_type event_type);
+
+void*
+cmark_node_get_user_data(cmark_node *node);
+
+int
+cmark_node_set_user_data(cmark_node *node, void *user_data);
+
+cmark_node_type
+cmark_node_get_type(cmark_node *node);
+
+const char*
+cmark_node_get_type_string(cmark_node *node);
+
+const char*
+cmark_node_get_literal(cmark_node *node);
+
+int
+cmark_node_set_literal(cmark_node *node, const char *content);
+
+int
+cmark_node_get_header_level(cmark_node *node);
+
+int
+cmark_node_set_header_level(cmark_node *node, int level);
+
+cmark_list_type
+cmark_node_get_list_type(cmark_node *node);
+
+int
+cmark_node_set_list_type(cmark_node *node, cmark_list_type type);
+
+cmark_delim_type
+cmark_node_get_list_delim(cmark_node *node);
+
+int
+cmark_node_set_list_delim(cmark_node *node, cmark_delim_type delim);
+
+int
+cmark_node_get_list_start(cmark_node *node);
+
+int
+cmark_node_set_list_start(cmark_node *node, int start);
+
+int
+cmark_node_get_list_tight(cmark_node *node);
+
+int
+cmark_node_set_list_tight(cmark_node *node, int tight);
+
+const char*
+cmark_node_get_fence_info(cmark_node *node);
+
+int
+cmark_node_set_fence_info(cmark_node *node, const char *info);
+
+const char*
+cmark_node_get_url(cmark_node *node);
+
+int
+cmark_node_set_url(cmark_node *node, const char *url);
+
+const char*
+cmark_node_get_title(cmark_node *node);
+
+int
+cmark_node_set_title(cmark_node *node, const char *title);
+
+int
+cmark_node_get_start_line(cmark_node *node);
+
+int
+cmark_node_get_start_column(cmark_node *node);
+
+int
+cmark_node_get_end_line(cmark_node *node);
+
+int
+cmark_node_get_end_column(cmark_node *node);
+
+void
+cmark_node_unlink(cmark_node *node);
+
+int
+cmark_node_insert_before(cmark_node *node, cmark_node *sibling);
+
+int
+cmark_node_insert_after(cmark_node *node, cmark_node *sibling);
+
+int
+cmark_node_prepend_child(cmark_node *node, cmark_node *child);
+
+int
+cmark_node_append_child(cmark_node *node, cmark_node *child);
+
+void
+cmark_consolidate_text_nodes(cmark_node *root);
+
+cmark_parser *cmark_parser_new(int options);
+
+void cmark_parser_free(cmark_parser *parser);
+
+void cmark_parser_feed(cmark_parser *parser, const char *buffer, size_t len);
+
+cmark_node *cmark_parser_finish(cmark_parser *parser);
+
+cmark_node *cmark_parse_document(const char *buffer, size_t len, int options);
+
+char *cmark_render_xml(cmark_node *root, int options);
+
+char *cmark_render_html(cmark_node *root, int options);
+
+char *cmark_render_man(cmark_node *root, int options);
+
+char *cmark_render_commonmark(cmark_node *root, int options, int width);
+
+extern const int cmark_version;
+
+extern const char cmark_version_string[];
+
+        ]]
+
+CMARK_OPT_DEFAULT = 0
+CMARK_OPT_SOURCEPOS = 1
+CMARK_OPT_HARDBREAKS = 2
+CMARK_OPT_NORMALIZE = 4
+CMARK_OPT_SMART = 8
+
+local inp = io.read("*all")
+local doc = cmark.cmark_parse_document(inp, string.len(inp), CMARK_OPT_SMART)
+local html = ffi.string(cmark.cmark_render_html(doc, CMARK_OPT_DEFAULT))
+print(html)

data/ext/commonmarker/cmark/wrappers/wrapper.py

@@ -0,0 +1,37 @@
+#!/usr/bin/env python
+
+# Example for using the shared library from python
+# Will work with either python 2 or python 3
+# Requires cmark library to be installed
+
+from ctypes import CDLL, c_char_p, c_long
+import sys
+import platform
+
+sysname = platform.system()
+
+if sysname == 'Darwin':
+    libname = "libcmark.dylib"
+elif sysname == 'Windows':
+    libname = "cmark.dll"
+else:
+    libname = "libcmark.so"
+cmark = CDLL(libname)
+
+markdown = cmark.cmark_markdown_to_html
+markdown.restype = c_char_p
+markdown.argtypes = [c_char_p, c_long, c_long]
+
+opts = 8 # CMARK_OPT_PRETTY
+
+def md2html(text):
+    if sys.version_info >= (3,0):
+        textbytes = text.encode('utf-8')
+        textlen = len(textbytes)
+        return markdown(textbytes, textlen, opts).decode('utf-8')
+    else:
+        textbytes = text
+        textlen = len(text)
+        return markdown(textbytes, textlen, opts)
+
+sys.stdout.write(md2html(sys.stdin.read()))

data/ext/commonmarker/cmark/wrappers/wrapper.rb

@@ -0,0 +1,15 @@
+#!/usr/bin/env ruby
+require 'ffi'
+
+module CMark
+  extend FFI::Library
+  ffi_lib ['libcmark', 'cmark']
+  attach_function :cmark_markdown_to_html, [:string, :int, :int], :string
+end
+
+def markdown_to_html(s)
+  len = s.bytesize
+  CMark::cmark_markdown_to_html(s, len, 0)
+end
+
+STDOUT.write(markdown_to_html(ARGF.read()))

data/ext/commonmarker/cmark/wrappers/wrapper.rkt

@@ -0,0 +1,190 @@
+#lang racket/base
+
+;; requires racket >= 5.3 because of submodules
+
+;; Lowlevel interface
+
+(module low-level racket/base
+
+  (require ffi/unsafe ffi/unsafe/define)
+
+  (provide (all-defined-out))
+
+  (define-ffi-definer defcmark (ffi-lib "libcmark"))
+
+  (define _cmark_node_type
+    (_enum '(none
+             ;; Block
+             document block-quote list item code-block
+             html paragraph header hrule
+             ;; Inline
+             text softbreak linebreak code inline-html
+             emph strong link image)))
+  (define _cmark_list_type
+    (_enum '(no_list bullet_list ordered_list)))
+  (define _cmark_delim_type
+    (_enum '(no_delim period_delim paren_delim)))
+  (define _cmark_opts
+    (_bitmask '(sourcepos = 1 hardbreaks = 2 normalize = 4 smart = 8)))
+
+  (define-cpointer-type _node)
+
+  (defcmark cmark_markdown_to_html
+    (_fun [bs : _bytes] [_int = (bytes-length bs)] _cmark_opts
+          -> [r : _bytes] -> (begin0 (bytes->string/utf-8 r) (free r))))
+
+  (defcmark cmark_parse_document
+    (_fun [bs : _bytes] [_int = (bytes-length bs)] _cmark_opts
+          -> _node))
+
+  (defcmark cmark_render_html
+    (_fun _node _cmark_opts
+          -> [r : _bytes] -> (begin0 (bytes->string/utf-8 r) (free r))))
+
+  (defcmark cmark_node_new              (_fun _cmark_node_type -> _node))
+  (defcmark cmark_node_free             (_fun _node -> _void))
+
+  (defcmark cmark_node_next             (_fun _node -> _node/null))
+  (defcmark cmark_node_previous         (_fun _node -> _node/null))
+  (defcmark cmark_node_parent           (_fun _node -> _node/null))
+  (defcmark cmark_node_first_child      (_fun _node -> _node/null))
+  (defcmark cmark_node_last_child       (_fun _node -> _node/null))
+
+  (defcmark cmark_node_get_user_data    (_fun _node -> _racket))
+  (defcmark cmark_node_set_user_data    (_fun _node _racket -> _bool))
+  (defcmark cmark_node_get_type         (_fun _node -> _cmark_node_type))
+  (defcmark cmark_node_get_type_string  (_fun _node -> _bytes))
+  (defcmark cmark_node_get_literal      (_fun _node -> _string))
+  (defcmark cmark_node_set_literal      (_fun _node _string -> _bool))
+  (defcmark cmark_node_get_header_level (_fun _node -> _int))
+  (defcmark cmark_node_set_header_level (_fun _node _int -> _bool))
+  (defcmark cmark_node_get_list_type    (_fun _node -> _cmark_list_type))
+  (defcmark cmark_node_set_list_type    (_fun _node _cmark_list_type -> _bool))
+  (defcmark cmark_node_get_list_delim   (_fun _node -> _cmark_delim_type))
+  (defcmark cmark_node_set_list_delim   (_fun _node _cmark_delim_type -> _bool))
+  (defcmark cmark_node_get_list_start   (_fun _node -> _int))
+  (defcmark cmark_node_set_list_start   (_fun _node _int -> _bool))
+  (defcmark cmark_node_get_list_tight   (_fun _node -> _bool))
+  (defcmark cmark_node_set_list_tight   (_fun _node _bool -> _bool))
+  (defcmark cmark_node_get_fence_info   (_fun _node -> _string))
+  (defcmark cmark_node_set_fence_info   (_fun _node _string -> _bool))
+  (defcmark cmark_node_get_url          (_fun _node -> _string))
+  (defcmark cmark_node_set_url          (_fun _node _string -> _bool))
+  (defcmark cmark_node_get_title        (_fun _node -> _string))
+  (defcmark cmark_node_set_title        (_fun _node _string -> _bool))
+  (defcmark cmark_node_get_start_line   (_fun _node -> _int))
+  (defcmark cmark_node_get_start_column (_fun _node -> _int))
+  (defcmark cmark_node_get_end_line     (_fun _node -> _int))
+  (defcmark cmark_node_get_end_column   (_fun _node -> _int))
+
+  (defcmark cmark_node_unlink           (_fun _node -> _void))
+  (defcmark cmark_node_insert_before    (_fun _node _node -> _bool))
+  (defcmark cmark_node_insert_after     (_fun _node _node -> _bool))
+  (defcmark cmark_node_prepend_child    (_fun _node _node -> _bool))
+  (defcmark cmark_node_append_child     (_fun _node _node -> _bool))
+  (defcmark cmark_consolidate_text_nodes (_fun _node -> _void))
+
+  )
+
+;; Rackety interface
+
+(module high-level racket/base
+
+  (require (submod ".." low-level) ffi/unsafe)
+
+  (provide cmark-markdown-to-html)
+  (define (cmark-markdown-to-html str [options '(normalize smart)])
+    (cmark_markdown_to_html (if (bytes? str) str (string->bytes/utf-8 str))
+                            options))
+
+  (require (for-syntax racket/base racket/syntax))
+  (define-syntax (make-getter+setter stx)
+    (syntax-case stx ()
+      [(_ name) (with-syntax ([(getter setter)
+                               (map (λ(op) (format-id #'name "cmark_node_~a_~a"
+                                                      op #'name))
+                                    '(get set))])
+                  #'(cons getter setter))]))
+  (define-syntax-rule (define-getters+setters name [type field ...] ...)
+    (define name (list (list 'type (make-getter+setter field) ...) ...)))
+  (define-getters+setters getters+setters
+    [header header_level] [code-block fence_info]
+    [link url title] [image url title]
+    [list list_type list_delim list_start list_tight])
+
+  (provide cmark->sexpr)
+  (define (cmark->sexpr node)
+    (define text (cmark_node_get_literal node))
+    (define type (cmark_node_get_type node))
+    (define children
+      (let loop ([node (cmark_node_first_child node)])
+        (if (not node) '()
+            (cons (cmark->sexpr node) (loop (cmark_node_next node))))))
+    (define info
+      (cond [(assq type getters+setters)
+             => (λ(gss) (map (λ(gs) ((car gs) node)) (cdr gss)))]
+            [else '()]))
+    (define (assert-no what-not b)
+      (when b (error 'cmark->sexpr "unexpected ~a in ~s" what-not type)))
+    (cond [(memq type '(document paragraph header block-quote list item
+                        emph strong link image))
+           (assert-no 'text text)
+           (list type info children)]
+          [(memq type '(text code code-block html inline-html
+                        softbreak linebreak hrule))
+           (assert-no 'children (pair? children))
+           (list type info text)]
+          [else (error 'cmark->sexpr "unknown type: ~s" type)]))
+
+  (provide sexpr->cmark)
+  (define (sexpr->cmark sexpr) ; assumes valid input, as generated by the above
+    (define (loop sexpr)
+      (define type (car sexpr))
+      (define info (cadr sexpr))
+      (define data (caddr sexpr))
+      (define node (cmark_node_new type))
+      (let ([gss (assq type getters+setters)])
+        (when gss
+          (unless (= (length (cdr gss)) (length info))
+            (error 'sexpr->cmark "bad number of info values in ~s" sexpr))
+          (for-each (λ(gs x) ((cdr gs) node x)) (cdr gss) info)))
+      (cond [(string? data) (cmark_node_set_literal node data)]
+            [(not data) (void)]
+            [(list? data)
+             (for ([child (in-list data)])
+               (cmark_node_append_child node (sexpr->cmark child)))]
+            [else (error 'sexpr->cmark "bad data in ~s" sexpr)])
+      node)
+    (define root (loop sexpr))
+    (register-finalizer root cmark_node_free)
+    root)
+
+  ;; Registers a `cmark_node_free` finalizer
+  (provide cmark-parse-document)
+  (define (cmark-parse-document str [options '(normalize smart)])
+    (define root (cmark_parse_document
+                  (if (bytes? str) str (string->bytes/utf-8 str))
+                  options))
+    (register-finalizer root cmark_node_free)
+    root)
+
+  (provide cmark-render-html)
+  (define (cmark-render-html root [options '(normalize smart)])
+    (cmark_render_html root options)))
+
+#; ;; sample use
+(begin
+  (require 'high-level racket/string)
+  (cmark-render-html
+   (cmark-parse-document
+    (string-join '("foo"
+                   "==="
+                   ""
+                   "> blah"
+                   ">"
+                   "> blah *blah* `bar()` blah:"
+                   ">"
+                   ">     function foo() {"
+                   ">       bar();"
+                   ">     }")
+                 "\n"))))