💾 Archived View for gmi.noulin.net › gitRepositories › md2html › file › md2html.c.gmi captured on 2023-01-29 at 13:24:44. Gemini links have been rewritten to link to archived content
-=-=-=-=-=-=-
md2html.c (11657B)
1 /*
2 * MD4C: Markdown parser for C
3 * (http://github.com/mity/md4c)
4 *
5 * Copyright (c) 2016-2017 Martin Mitas
6 *
7 * Permission is hereby granted, free of charge, to any person obtaining a
8 * copy of this software and associated documentation files (the "Software"),
9 * to deal in the Software without restriction, including without limitation
10 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
11 * and/or sell copies of the Software, and to permit persons to whom the
12 * Software is furnished to do so, subject to the following conditions:
13 *
14 * The above copyright notice and this permission notice shall be included in
15 * all copies or substantial portions of the Software.
16 *
17 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
18 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
19 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
20 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
21 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
22 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
23 * IN THE SOFTWARE.
24 */
25
26 #include <stdio.h>
27 #include <stdlib.h>
28 #include <string.h>
29 #include <time.h>
30
31 #include "render_html.h"
32 #include "cmdline.h"
33
34
35
36 /* Global options. */
37 static unsigned parser_flags = 0;
38 static unsigned renderer_flags = MD_RENDER_FLAG_DEBUG;
39 static int want_fullhtml = 0;
40 static int want_stat = 0;
41
42
43 /*********************************
44 *** Simple grow-able buffer ***
45 *********************************/
46
47 /* We render to a memory buffer instead of directly outputting the rendered
48 * documents, as this allows using this utility for evaluating performance
49 * of MD4C (--stat option). This allows us to measure just time of the parser,
50 * without the I/O.
51 */
52
53 struct membuffer {
54 char* data;
55 MD_SIZE asize;
56 MD_SIZE size;
57 };
58
59 static void
60 membuf_init(struct membuffer* buf, MD_SIZE new_asize)
61 {
62 buf->size = 0;
63 buf->asize = new_asize;
64 buf->data = malloc(buf->asize);
65 if(buf->data == NULL) {
66 fprintf(stderr, "membuf_init: malloc() failed.");
67 exit(1);
68 }
69 }
70
71 static void
72 membuf_fini(struct membuffer* buf)
73 {
74 if(buf->data)
75 free(buf->data);
76 }
77
78 static void
79 membuf_grow(struct membuffer* buf, MD_SIZE new_asize)
80 {
81 buf->data = realloc(buf->data, new_asize);
82 if(buf->data == NULL) {
83 fprintf(stderr, "membuf_grow: realloc() failed.");
84 exit(1);
85 }
86 buf->asize = new_asize;
87 }
88
89 static void
90 membuf_append(struct membuffer* buf, const char* data, MD_SIZE size)
91 {
92 if(buf->asize < buf->size + size)
93 membuf_grow(buf, (buf->size + size) * 2);
94 memcpy(buf->data + buf->size, data, size);
95 buf->size += size;
96 }
97
98
99 /**********************
100 *** Main program ***
101 **********************/
102
103 static void
104 process_output(const MD_CHAR* text, MD_SIZE size, void* userdata)
105 {
106 membuf_append((struct membuffer*) userdata, text, size);
107 }
108
109 static int
110 process_file(FILE* in, FILE* out)
111 {
112 MD_SIZE n;
113 struct membuffer buf_in = {0};
114 struct membuffer buf_out = {0};
115 int ret = -1;
116 clock_t t0, t1;
117
118 membuf_init(&buf_in, 32 * 1024);
119
120 /* Read the input file into a buffer. */
121 while(1) {
122 if(buf_in.size >= buf_in.asize)
123 membuf_grow(&buf_in, 2 * buf_in.asize);
124
125 n = fread(buf_in.data + buf_in.size, 1, buf_in.asize - buf_in.size, in);
126 if(n == 0)
127 break;
128 buf_in.size += n;
129 }
130
131 /* Input size is good estimation of output size. Add some more reserve to
132 * deal with the HTML header/footer and tags. */
133 membuf_init(&buf_out, buf_in.size + buf_in.size/8 + 64);
134
135 /* Parse the document. This shall call our callbacks provided via the
136 * md_renderer_t structure. */
137 t0 = clock();
138
139 ret = md_render_html(buf_in.data, buf_in.size, process_output,
140 (void*) &buf_out, parser_flags, renderer_flags);
141
142 t1 = clock();
143 if(ret != 0) {
144 fprintf(stderr, "Parsing failed.\n");
145 goto out;
146 }
147
148 /* Write down the document in the HTML format. */
149 if(want_fullhtml) {
150 fprintf(out, "<html>\n");
151 fprintf(out, "<head>\n");
152 fprintf(out, "<title></title>\n");
153 fprintf(out, "<meta name=\"generator\" content=\"md2html\">\n");
154 fprintf(out, "</head>\n");
155 fprintf(out, "<body>\n");
156 }
157
158 fwrite(buf_out.data, 1, buf_out.size, out);
159
160 if(want_fullhtml) {
161 fprintf(out, "</body>\n");
162 fprintf(out, "</html>\n");
163 }
164
165 if(want_stat) {
166 if(t0 != (clock_t)-1 && t1 != (clock_t)-1) {
167 double elapsed = (double)(t1 - t0) / CLOCKS_PER_SEC;
168 if (elapsed < 1)
169 fprintf(stderr, "Time spent on parsing: %7.2f ms.\n", elapsed*1e3);
170 else
171 fprintf(stderr, "Time spent on parsing: %6.3f s.\n", elapsed);
172 }
173 }
174
175 /* Success if we have reached here. */
176 ret = 0;
177
178 out:
179 membuf_fini(&buf_in);
180 membuf_fini(&buf_out);
181
182 return ret;
183 }
184
185
186 #define OPTION_ARG_NONE 0
187 #define OPTION_ARG_REQUIRED 1
188 #define OPTION_ARG_OPTIONAL 2
189
190 static const option cmdline_options[] = {
191 { "output", 'o', 'o', OPTION_ARG_REQUIRED },
192 { "full-html", 'f', 'f', OPTION_ARG_NONE },
193 { "stat", 's', 's', OPTION_ARG_NONE },
194 { "help", 'h', 'h', OPTION_ARG_NONE },
195 { "version", 'v', 'v', OPTION_ARG_NONE },
196 { "commonmark", 0, 'c', OPTION_ARG_NONE },
197 { "github", 0, 'g', OPTION_ARG_NONE },
198 { "fverbatim-entities", 0, 'E', OPTION_ARG_NONE },
199 { "fpermissive-atx-headers", 0, 'A', OPTION_ARG_NONE },
200 { "fpermissive-url-autolinks", 0, 'U', OPTION_ARG_NONE },
201 { "fpermissive-www-autolinks", 0, '.', OPTION_ARG_NONE },
202 { "fpermissive-email-autolinks", 0, '@', OPTION_ARG_NONE },
203 { "fpermissive-autolinks", 0, 'V', OPTION_ARG_NONE },
204 { "fno-indented-code", 0, 'I', OPTION_ARG_NONE },
205 { "fno-html-blocks", 0, 'F', OPTION_ARG_NONE },
206 { "fno-html-spans", 0, 'G', OPTION_ARG_NONE },
207 { "fno-html", 0, 'H', OPTION_ARG_NONE },
208 { "fcollapse-whitespace", 0, 'W', OPTION_ARG_NONE },
209 { "ftables", 0, 'T', OPTION_ARG_NONE },
210 { "fstrikethrough", 0, 'S', OPTION_ARG_NONE },
211 { 0 }
212 };
213
214 static void
215 usage(void)
216 {
217 printf(
218 "Usage: md2html [OPTION]... [FILE]\n"
219 "Convert input FILE (or standard input) in Markdown format to HTML.\n"
220 "\n"
221 "General options:\n"
222 " -o --output=FILE Output file (default is standard output)\n"
223 " -f, --full-html Generate full HTML document, including header\n"
224 " -s, --stat Measure time of input parsing\n"
225 " -h, --help Display this help and exit\n"
226 " -v, --version Display version and exit\n"
227 "\n"
228 "Markdown dialect options:\n"
229 "(note these are equivalent to some combinations of flags below)\n"
230 " --commonmark CommonMark (this is default)\n"
231 " --github Github Flavored Markdown\n"
232 "\n"
233 "Markdown extension options:\n"
234 " --fcollapse-whitespace\n"
235 " Collapse non-trivial whitespace\n"
236 " --fverbatim-entities\n"
237 " Do not translate entities\n"
238 " --fpermissive-atx-headers\n"
239 " Allow ATX headers without delimiting space\n"
240 " --fpermissive-url-autolinks\n"
241 " Allow URL autolinks without '<', '>'\n"
242 " --fpermissive-www-autolinks\n"
243 " Allow WWW autolinks without any scheme (e.g. 'www.example.com')\n"
244 " --fpermissive-email-autolinks \n"
245 " Allow e-mail autolinks without '<', '>' and 'mailto:'\n"
246 " --fpermissive-autolinks\n"
247 " Same as --fpermissive-url-autolinks --fpermissive-www-autolinks\n"
248 " --fpermissive-email-autolinks\n"
249 " --fno-indented-code\n"
250 " Disable indented code blocks\n"
251 " --fno-html-blocks\n"
252 " Disable raw HTML blocks\n"
253 " --fno-html-spans\n"
254 " Disable raw HTML spans\n"
255 " --fno-html Same as --fno-html-blocks --fno-html-spans\n"
256 " --ftables Enable tables\n"
257 " --fstrikethrough Enable strikethrough spans\n"
258 );
259 }
260
261 static void
262 version(void)
263 {
264 printf("%d.%d.%d\n", MD_VERSION_MAJOR, MD_VERSION_MINOR, MD_VERSION_RELEASE);
265 }
266
267 static const char* input_path = NULL;
268 static const char* output_path = NULL;
269
270 static int
271 cmdline_callback(int opt, char const* value, void* data)
272 {
273 switch(opt) {
274 case 0:
275 if(input_path) {
276 fprintf(stderr, "Too many arguments. Only one input file can be specified.\n");
277 fprintf(stderr, "Use --help for more info.\n");
278 exit(1);
279 }
280 input_path = value;
281 break;
282
283 case 'o': output_path = value; break;
284 case 'f': want_fullhtml = 1; break;
285 case 's': want_stat = 1; break;
286 case 'h': usage(); exit(0); break;
287 case 'v': version(); exit(0); break;
288
289 case 'c': parser_flags = MD_DIALECT_COMMONMARK; break;
290 case 'g': parser_flags = MD_DIALECT_GITHUB; break;
291
292 case 'E': renderer_flags |= MD_RENDER_FLAG_VERBATIM_ENTITIES; break;
293 case 'A': parser_flags |= MD_FLAG_PERMISSIVEATXHEADERS; break;
294 case 'I': parser_flags |= MD_FLAG_NOINDENTEDCODEBLOCKS; break;
295 case 'F': parser_flags |= MD_FLAG_NOHTMLBLOCKS; break;
296 case 'G': parser_flags |= MD_FLAG_NOHTMLSPANS; break;
297 case 'H': parser_flags |= MD_FLAG_NOHTML; break;
298 case 'W': parser_flags |= MD_FLAG_COLLAPSEWHITESPACE; break;
299 case 'U': parser_flags |= MD_FLAG_PERMISSIVEURLAUTOLINKS; break;
300 case '.': parser_flags |= MD_FLAG_PERMISSIVEWWWAUTOLINKS; break;
301 case '@': parser_flags |= MD_FLAG_PERMISSIVEEMAILAUTOLINKS; break;
302 case 'V': parser_flags |= MD_FLAG_PERMISSIVEAUTOLINKS; break;
303 case 'T': parser_flags |= MD_FLAG_TABLES; break;
304 case 'S': parser_flags |= MD_FLAG_STRIKETHROUGH; break;
305
306 default:
307 fprintf(stderr, "Illegal option: %s\n", value);
308 fprintf(stderr, "Use --help for more info.\n");
309 exit(1);
310 break;
311 }
312
313 return 0;
314 }
315
316 int
317 main(int argc, char** argv)
318 {
319 FILE* in = stdin;
320 FILE* out = stdout;
321 int ret = 0;
322
323 if(readoptions(cmdline_options, argc, argv, cmdline_callback, NULL) < 0) {
324 usage();
325 exit(1);
326 }
327
328 if(input_path != NULL && strcmp(input_path, "-") != 0) {
329 in = fopen(input_path, "rb");
330 if(in == NULL) {
331 fprintf(stderr, "Cannot open %s.\n", input_path);
332 exit(1);
333 }
334 }
335 if(output_path != NULL && strcmp(output_path, "-") != 0) {
336 out = fopen(output_path, "wt");
337 if(out == NULL) {
338 fprintf(stderr, "Cannot open %s.\n", input_path);
339 exit(1);
340 }
341 }
342
343 ret = process_file(in, out);
344 if(in != stdin)
345 fclose(in);
346 if(out != stdout)
347 fclose(out);
348
349 return ret;
350 }