Add test changes from cmark

test/cmark.py

@@ -1,40 +1,87 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 
-from ctypes import CDLL, c_char_p, c_long, c_int
-from subprocess import *
+from ctypes import *
+from subprocess import Popen, PIPE
 import platform
 import os
 
+class cmark_mem(Structure):
+    _fields_ = [("calloc", c_void_p),
+                ("realloc", c_void_p),
+                ("free", CFUNCTYPE(None, c_void_p))]
+
 def pipe_through_prog(prog, text):
     p1 = Popen(prog.split(), stdout=PIPE, stdin=PIPE, stderr=PIPE)
     [result, err] = p1.communicate(input=text.encode('utf-8'))
-    return [p1.returncode, result, err]
+    return [p1.returncode, result.decode('utf-8'), err]
+
+def to_html(lib, text):
+    get_alloc = lib.cmark_get_default_mem_allocator
+    get_alloc.restype = POINTER(cmark_mem)
+    free_func = get_alloc().contents.free
+
+    markdown = lib.cmark_markdown_to_html
+    markdown.restype = POINTER(c_char)
+    markdown.argtypes = [c_char_p, c_size_t, c_int]
 
-def use_library(lib, text):
     textbytes = text.encode('utf-8')
     textlen = len(textbytes)
-    return [0, lib(textbytes, textlen, 0), '']
+    # 1 << 17 == CMARK_OPT_UNSAFE
+    cstring = markdown(textbytes, textlen, 1 << 17)
+    result = string_at(cstring).decode('utf-8')
+    free_func(cstring)
+
+    return [0, result, '']
+
+def to_commonmark(lib, text):
+    get_alloc = lib.cmark_get_default_mem_allocator
+    get_alloc.restype = POINTER(cmark_mem)
+    free_func = get_alloc().contents.free
+
+    parse_document = lib.cmark_parse_document
+    parse_document.restype = c_void_p
+    parse_document.argtypes = [c_char_p, c_size_t, c_int]
+
+    render_commonmark = lib.cmark_render_commonmark
+    render_commonmark.restype = POINTER(c_char)
+    render_commonmark.argtypes = [c_void_p, c_int, c_int]
+
+    free_node = lib.cmark_node_free
+    free_node.argtypes = [c_void_p]
+
+    textbytes = text.encode('utf-8')
+    textlen = len(textbytes)
+    node = parse_document(textbytes, textlen, 0)
+    cstring = render_commonmark(node, 0, 0)
+    result = string_at(cstring).decode('utf-8')
+    free_func(cstring)
+    free_node(node)
+
+    return [0, result, '']
 
 class CMark:
     def __init__(self, prog=None, library_dir=None):
         self.prog = prog
         if prog:
+            prog += ' --unsafe'
             self.to_html = lambda x: pipe_through_prog(prog, x)
+            self.to_commonmark = lambda x: pipe_through_prog(prog + ' -t commonmark', x)
         else:
             sysname = platform.system()
             if sysname == 'Darwin':
-                libname = "libcmark.dylib"
+                libnames = [ "libcmark.dylib" ]
             elif sysname == 'Windows':
-                libname = "cmark.dll"
-            else:
-                libname = "libcmark.so"
-            if library_dir:
-                libpath = os.path.join(library_dir, libname)
+                libnames = [ "cmark.dll", "libcmark.dll" ]
             else:
-                libpath = os.path.join("build", "src", libname)
+                libnames = [ "libcmark.so" ]
+            if not library_dir:
+                library_dir = os.path.join("build", "src")
+            for libname in libnames:
+                candidate = os.path.join(library_dir, libname)
+                if os.path.isfile(candidate):
+                    libpath = candidate
+                    break
             cmark = CDLL(libpath)
-            markdown = cmark.cmark_markdown_to_html
-            markdown.restype = c_char_p
-            markdown.argtypes = [c_char_p, c_long, c_int]
-            self.to_html = lambda x: use_library(markdown, x)
+            self.to_html = lambda x: to_html(cmark, x)
+            self.to_commonmark = lambda x: to_commonmark(cmark, x)

test/spec_tests.py

@@ -6,42 +6,45 @@
 import argparse
 import re
 import json
+import os
 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='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')
-    parser.add_argument('--track', metavar='path',
-            help='track which test cases pass/fail in the given JSON file and only report changes')
-    args = parser.parse_args(sys.argv[1:])
+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='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')
+parser.add_argument('--track', metavar='path',
+        help='track which test cases pass/fail in the given JSON file and only report changes')
+parser.add_argument('--fuzz-corpus',
+        help='convert test cases to fuzz corpus')
+args = parser.parse_args(sys.argv[1:])
 
 def out(str):
     sys.stdout.buffer.write(str.encode('utf-8')) 
 
 def print_test_header(test):
-    out("Example %d (lines %d-%d) %s\n" % (test['example'], test['start_line'], test['end_line'], test['section']))
+    out("Example %d (lines %d-%d) %s\n"
+        % (test['example'], test['start_line'], test['end_line'], test['section']))
 
-def do_test(test, normalize, prev_result):
-    [retcode, actual_html_bytes, err] = cmark.to_html(test['markdown'])
+def do_test(converter, test, normalize, prev_result):
+    [retcode, actual_html_bytes, err] = converter(test['markdown'])
     if retcode != 0:
         if prev_result != 'error':
             print_test_header(test)
@@ -52,7 +55,7 @@ def do_test(test, normalize, prev_result):
     expected_html = test['html']
 
     try:
-        actual_html = actual_html_bytes.decode('utf-8')
+        actual_html = actual_html_bytes
     except UnicodeDecodeError as e:
         if prev_result != 'fail':
             print_test_header(test)
@@ -136,17 +139,31 @@ def get_tests(specfile):
         exit(0)
 
     all_tests = get_tests(args.spec)
+
+    if args.fuzz_corpus:
+        i = 1
+        base = os.path.basename(args.spec)
+        (name, ext) = os.path.splitext(base)
+        for test in all_tests:
+            filename = os.path.join(args.fuzz_corpus, '%s.%d' % (name, i))
+            with open(filename, 'wb') as f:
+                f.write(b'\0' * 8) # options header
+                f.write(test['markdown'].encode())
+            i += 1
+        exit(0)
+
     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) ]
+    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:
         out(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)
+        converter = CMark(prog=args.program, library_dir=args.library_dir).to_html
         result_counts = {'pass': 0, 'fail': 0, 'error': 0, 'skip': skipped}
 
         previous = {}
@@ -161,7 +178,7 @@ def get_tests(specfile):
         results = {}
 
         for test in tests:
-            result = do_test(test, args.normalize, previous.get(str(test['example'])))
+            result = do_test(converter, test, args.normalize, previous.get(str(test['example'])))
             result_counts[result] += 1
             results[test['example']] = result