sparline.c (20815B)
1 // Usage: sparline spartan://hostname:port/path --infile filename
2 // Prompt: Enter link number, b, - or + for back and q for quit
3
4 #define _GNU_SOURCE
5 #include <stdio.h>
6 #include <sys/types.h>
7 #include <sys/socket.h>
8 #include <netdb.h>
9 #include <netinet/in.h>
10 #include <string.h>
11 #include <stdlib.h>
12 #include <ctype.h>
13 #include <unistd.h>
14 #include <sys/stat.h>
15 #include <sys/time.h>
16 #include <stdbool.h>
17
18 // links in page
19 // elements are offsets in page
20 // [][0] is link start
21 // [][1] is link end
22 // linkCount is the number of links on the page
23 size_t links[100000][2] = {0};
24 size_t linkCount = 0;
25 char *page = NULL;
26 size_t pageBufSize = 0;
27
28 // history is a dynamic vector of historyEt element
29 // it is used as a stack
30 // pop is free string and decrease vector count
31 typedef struct {
32 char *hostname;
33 char *port;
34 char *path;
35 char *link;
36 } historyEt;
37
38 #define sliceT(typeName, elementType)\
39 typedef struct {\
40 size_t count;\
41 elementType *array;\
42 } typeName
43
44 sliceT(historyt, historyEt);
45
46 #define var __auto_type
47 #define TOKENPASTE2(a, b) a ## b
48 #define TOKENPASTE(a, b) TOKENPASTE2(a, b)
49 #define UNIQVAR(name) TOKENPASTE(name, __LINE__)
50
51 #define sliceInitCount(name, countInt) do{\
52 var UNIQVAR(c) = countInt;\
53 (name)->array = malloc(UNIQVAR(c) * sizeof (name)->array[0]);\
54 (name)->count = 0;\
55 } while(0)
56
57 #define sliceSz 1
58
59 #define sliceAlloc(name) do{\
60 if (!(name)->array) {\
61 (name)->array = malloc(sliceSz * sizeof (name)->array[0]);\
62 }\
63 else {\
64 (name)->array = realloc((name)->array, ((name)->count + sliceSz) * sizeof (name)->array[0]);\
65 }\
66 } while(0)
67
68 #define slicePush(name) do {\
69 sliceAlloc(name);\
70 (name)->count++;\
71 } while(0)
72
73 #define sliceAt(name, index) ((name)->array[index])
74 #define sliceLast(name) ((name)->array[(name)->count-1])
75
76 #define sliceAppend(name, v) do{\
77 slicePush(name);\
78 sliceLast(name) = v;\
79 } while(0)
80
81 /**
82 * convert string to decimal integer
83 *
84 * \param
85 * string
86 * \return
87 * int64_t
88 * 0 when string represents 0 or doesnt represent a number or the input is NULL
89 */
90 int64_t parseInt(const char *string) {
91 while (!isdigit(*string) && *string != '-' && *string != 0) {
92 string++;
93 }
94 int64_t r = strtoll(string, NULL, 10);
95 return(r);
96 }
97
98 #define startMax 20
99
100 /**
101 * read String
102 * read user input (one line) as a string
103 *
104 * there is no size limit and the buffer expands as needed
105 *
106 * \return
107 * line from the user (you must free the pointer)
108 * NULL when buffer allocation failed
109 */
110 char *readS(void) {
111 int max = startMax;
112
113 char *s = malloc((size_t)max);
114 if (!s) {
115 return(NULL);
116 }
117
118 int i = 0;
119 while (1) {
120 int c = getchar();
121 if (c == '\n') {
122 s[i] = 0;
123 break;
124 }
125 s[i] = (char)c;
126 if (i == max-1) {
127 // buffer full
128 max += max;
129 char *tmp = realloc(s, (size_t)max);
130 if (!tmp) {
131 free(s);
132 return(NULL);
133 }
134 s = tmp;
135 }
136 i++;
137 }
138 return(s);
139 }
140
141 // makeRoom is dynamic memory allocation algorithm
142 // given a length, an allocated size and the additionnal length,
143 // makeRoom returns the new allocated size for realloc
144 // when the new allocated size equals alloc value, there is no need to realloc the memory, enough space is already available
145 #define prealloc (1024*1024)
146 #define funcbegin ({
147 #define funcend })
148 #define makeRoom(length, alloc, addlength) funcbegin\
149 typeof(alloc) r;\
150 typeof(alloc) newlen = (length) + (addlength);\
151 if (newlen < (alloc)) {\
152 r = alloc;\
153 } \
154 else {\
155 if (newlen < prealloc) {\
156 r = newlen * 2;\
157 }\
158 else {\
159 r = newlen + prealloc;\
160 }\
161 }\
162 r;\
163 funcend
164
165 bool getPage(char *hostname, char *ports, char *path, size_t content_length, void *senddata) {
166 int sock;
167 struct sockaddr_in server;
168 struct hostent *hp;
169 int mysock;
170 char buf[4096] = {0};
171 char redirectPath[4096] = {0};
172 int rval;
173 int i;
174 bool r = true;
175
176 openSocket:
177 sock = socket(AF_INET, SOCK_STREAM, 0);
178 if (sock < 0){
179 perror("Failed to create socket");
180 r = false;
181 goto showPage;
182 }
183
184 // Set 10s timeouts for receive and send (SO_RCVTIMEO and SO_SNDTIMEO)
185 struct timeval timeout;
186 timeout.tv_sec = 10;
187 timeout.tv_usec = 0;
188 if (setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO, (void *) &timeout, sizeof(timeout)) < 0) {
189 perror("receive timeout failed");
190 close(sock);
191 r = false;
192 goto showPage;
193 }
194 if (setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO, (void *) &timeout, sizeof(timeout)) < 0) {
195 perror("send timeout failed");
196 close(sock);
197 r = false;
198 goto showPage;
199 }
200
201 server.sin_family = AF_INET;
202
203 hp = gethostbyname(hostname);
204 if (hp==0) {
205 perror("gethostbyname failed");
206 close(sock);
207 r = false;
208 goto showPage;
209 }
210
211 memcpy(&server.sin_addr, hp->h_addr, hp->h_length);
212
213 int64_t port = parseInt(ports);
214
215 if (port < 1 || port > 65000) {
216 close(sock);
217 printf("Invalid port %d.\n", port);
218 r = false;
219 goto showPage;
220 }
221
222 server.sin_port = htons(port);
223
224 if (connect(sock,(struct sockaddr *) &server, sizeof(server))){
225 perror("connect failed");
226 close(sock);
227 r = false;
228 goto showPage;
229 }
230
231 // build request
232 // hostname SPC path SPC content_length\r\n
233 size_t len = strlen(hostname);
234 memcpy(buf, hostname, len);
235 buf[len] = ' ';
236 char *cursor = buf + len + 1;
237 len = strlen(path);
238 memcpy(cursor, path, len);
239 char lenstr[50];
240 int ln = sprintf(lenstr, " %d\r\n", content_length);
241 memcpy(cursor + len, lenstr, ln);
242 cursor += len + ln;
243
244 // show request
245 *cursor = 0;
246 puts(buf);
247 printf("Page URL: spartan://%s:%s%s\n", hostname, ports, path);
248
249 // send request
250 if(send(sock, buf, cursor - buf, 0) < 0){
251 perror("send failed");
252 close(sock);
253 r = false;
254 goto showPage;
255 }
256
257 if (content_length > 0) {
258 // send data with request
259 size_t offset = 0;
260 while (content_length) {
261 size_t tosend = content_length > 2048 ? 2048 : content_length;
262 if(send(sock, senddata + offset, tosend, 0) < 0){
263 perror("send failed");
264 close(sock);
265 r = false;
266 goto showPage;
267 }
268 content_length -= tosend;
269 offset += tosend;
270 }
271 }
272
273 // receive server response
274 size_t pageSize = sizeof(buf);
275 size_t offset = 0;
276 free(page);
277 page = malloc(pageSize);
278 page[0] = 0;
279 do {
280 rval = recv(sock, buf, sizeof(buf), 0);
281 if (rval != -1 && rval != 0) {
282 size_t newSize = makeRoom(offset, pageSize, rval);
283 if (newSize > pageSize) {
284 char *tmp = realloc(page, newSize);
285 if (!tmp) {
286 // memory allocation error
287 // crash
288 exit(1);
289 }
290 page = tmp;
291 pageSize = newSize;
292 }
293 memcpy(page+offset, buf, rval);
294 offset += rval;
295 }
296 } while (rval != -1 && rval != 0);
297 close(sock);
298
299 if (buf[0] == '3' && buf[1] == ' ' && buf[2] != ' ' && offset <= sizeof(redirectPath)) {
300 // redirect
301 // scan path and reopen socket
302 cursor = buf + 2;
303 while (cursor < buf + offset &&
304 !isspace(*cursor)) {
305 cursor++;
306 }
307 *cursor = 0;
308 puts(buf);
309 memcpy(redirectPath, buf+2, cursor - buf - 2);
310 redirectPath[cursor - buf - 2] = 0;
311 path = redirectPath;
312 goto openSocket;
313 }
314
315 // save page size for reuse
316 // when a link fails to load, the current page is reused
317 pageBufSize = offset;
318
319 // show page
320 // highlight headers, lists, blockquote, link and fixedwidth blocks
321 // collect links in links array
322 showPage:
323 if (page == NULL) return false;
324 linkCount = 0;
325 enum {normal, header, link, list, fixedwidth, blockquote};
326 int state = normal;
327 #define RST "\x1B[0m"
328 #define BLD "\x1B[1m"
329 #define RED "\x1B[31m"
330 #define GRN "\x1B[32m"
331 #define YLW "\x1B[33m"
332 #define BLU "\x1B[34m"
333 #define MGT "\x1B[35m"
334 #define CYN "\x1B[36m"
335 #define WHT "\x1B[37m"
336 puts(RED"──────────────────────────────────────────────────────────────────────"RST);
337 for (int i = 0; i < pageBufSize; i++) {
338 if (state != fixedwidth) {
339 if (page[i] == '#' && (i == 0 || page[i-1] == '\n')) {
340 state = header;
341 printf(BLD YLW);
342 }
343 if (page[i] == '*' && (i == 0 || page[i-1] == '\n')) {
344 state = list;
345 printf(CYN);
346 }
347 if (page[i] == '>' && (i == 0 || page[i-1] == '\n')) {
348 state = blockquote;
349 printf(BLD WHT);
350 }
351 if (page[i] == '=' && page[i+1] == '>' && (i == 0 || page[i-1] == '\n')) {
352 state = link;
353 links[linkCount][0] = i;
354 links[linkCount][1] = 0; // invalid
355 printf(BLD GRN "%d " RST BLD BLU, linkCount);
356 }
357 }
358 if (page[i] == '`' && page[i+1] == '`' && page[i+2] == '`' && (i == 0 || page[i-1] == '\n')) {
359 if (state == normal) {
360 state = fixedwidth;
361 printf(BLD MGT);
362 }
363 else if (state == fixedwidth) {
364 state = normal;
365 printf(RST);
366 }
367 }
368 if (state != normal && state != fixedwidth && page[i] == '\n') {
369 if (state == link) {
370 links[linkCount++][1] = i;
371 }
372 state = normal;
373 printf(RST);
374 }
375 putchar(page[i]);
376 }
377 return true;
378 }
379
380 historyEt parseURL(char *url) {
381 size_t len = strlen(url);
382
383 char *cursor = strstr(url, "spartan://");
384 if (!cursor) return (historyEt){0};
385
386 char *s = cursor + strlen("spartan://");
387 cursor = s;
388 // scan hostname
389 while(*cursor != '/' &&
390 *cursor != ' ' &&
391 *cursor != '\t' &&
392 *cursor != ':' &&
393 *cursor != '\n' &&
394 *cursor != 0 &&
395 cursor < url + len) {
396 cursor++;
397 }
398 char *hostname = malloc(cursor-s+1);
399 memcpy(hostname, s, cursor-s);
400 hostname[cursor-s] = 0;
401
402 char *port;
403 if (*cursor == ':') {
404 // port is specified
405 cursor++;
406 port = cursor;
407 while(isdigit(*cursor)) {
408 cursor++;
409 }
410 char c = *cursor;
411 *cursor = 0;
412 port = strdup(port);
413 *cursor = c;
414 }
415 else {
416 port = strdup("300");
417 }
418
419 char *path;
420 if (*cursor == ' ' ||
421 *cursor == '\t' ||
422 *cursor == '\n' ||
423 *cursor == 0) {
424 // path is empty
425 path = strdup("/");
426 }
427 else {
428 // *cursor == '/'
429 // scan path
430 s = cursor;
431 while(!isspace(*cursor)) {
432 cursor++;
433 }
434
435 if (cursor == s+1) {
436 // path is /
437 path = strdup("/");
438 }
439 else {
440 path = malloc(cursor-s+1);
441 memcpy(path, s, cursor-s);
442 path[cursor-s] = 0;
443 }
444 }
445
446 return (historyEt){.hostname = hostname, .port = port, .path = path};
447 }
448
449 /**
450 * get file size
451 *
452 * \param
453 * filePath: path to file
454 * \return
455 * ssize_t >= 0 size
456 * -1 an error occured or filePath is NULL or empty string
457 */
458 ssize_t fileSize(const char *filePath) {
459 struct stat st;
460
461 int r = stat(filePath, &st);
462 if (r) {
463 printf("Error, the path was: \"%s\"\n", filePath);
464 return(-1);
465 }
466
467 // macOS returns a varying number a number above the constant below
468 // when the file doesnt exists
469 if ((uint64_t)(st.st_size) > 140734000000000) {
470 return(-1);//LCOV_EXCL_LINE
471 }
472 return(st.st_size);
473 }
474
475 int main(int ac, char **av){
476 if (ac < 2) {
477 puts("Usage: sparline spartan://hostname:port/path --infile filename\n"
478 "Default port is 300\n"
479 "Prompt: Enter link number, b, - or + for back and q for quit");
480 return 0;
481 }
482
483 historyt history;
484 sliceInitCount(&history, 16);
485
486 historyEt e = parseURL(av[1]);
487 if (!e.hostname) {
488 puts("Error: Could not parse the url in argument 1");
489 puts(av[1]);
490 return 1;
491 }
492 sliceAppend(&history, e);
493
494 if (ac > 3) {
495 // check for --infile
496 if (strcmp(av[2], "--infile") == 0) {
497 puts(av[3]);
498 ssize_t size = fileSize(av[3]);
499 if (size < 0) return 1;
500 char *data = malloc(size);
501 FILE *f = fopen(av[3], "r");
502 size_t sz = fread(data, 1, size, f);
503 fclose(f);
504 if (sz < size) {
505 puts("Could not read complete file.");
506 return 1;
507 }
508 getPage(e.hostname, e.port, e.path, size /*content length*/, data /*send data*/);
509 free(data);
510 free(page);
511 return 0;
512 }
513 }
514
515 bool r = getPage(e.hostname, e.port, e.path, 0 /*content length*/, NULL /*send data*/);
516 if (!r) {
517 // get page failed
518 free(page);
519 return 1;
520 }
521
522 char *userInput = strdup("");
523 while(strcmp(userInput, "q") != 0) {
524 // print prompt with link count on current page
525 if (!linkCount)
526 printf("> ");
527 else
528 printf("%d links > ", linkCount);
529 free(userInput);
530 // userInput is allocated in readS
531 userInput = readS();
532 puts(userInput);
533 if (isdigit(userInput[0])) {
534 // go to link
535 int link = parseInt(userInput);
536 if (link >= linkCount) {
537 printf("Link number too high: %d, link count: %d\n", link, linkCount);
538 continue;
539 }
540 // build a valid link string
541 page[links[link][1]] = 0;
542 char *s = &page[links[link][0]];
543 char *hostname;
544 char *port;
545 char *path;
546
547 char *cursor = strstr(s, "spartan://");
548 if (!cursor) {
549 if (strstr(s, "://")) {
550 puts("Only spartan links are supported.");
551 continue;
552 }
553 // this is a path
554 // sometimes there is no space after =>
555 cursor = isspace(*(s+2)) ? s+3 : s+2;
556 // p is link start
557 char *p = cursor;
558 // search for link end and skip the title
559 while (!isspace(*cursor) && cursor < &page[links[link][1]]) {
560 cursor++;
561 }
562 *cursor = 0;
563 // there are no hostname and port in link, use the previous ones
564 hostname = strdup(sliceLast(&history).hostname);
565 port = strdup(sliceLast(&history).port);
566 if (*p != '/') {
567 // relative path
568 // check if current path is a file like index.gmi
569 // to avoid creating a path like:
570 // /index.gmi/p
571 size_t len = strlen(sliceLast(&history).path);
572 if (len > 4 && memcmp(sliceLast(&history).path + len - 4, ".gmi", 4) == 0) {
573 // when path is .. remove filename only
574 if (strcmp(p, "..") == 0)
575 p = "";
576 else if (strcmp(p, ".") == 0) {
577 // path is ., reload same page as previous one
578 path = strdup(sliceLast(&history).path);
579 goto downloadPage;
580 }
581 // previous has a gmi extension
582 // find dirname, can be empty string
583 char *hp = sliceLast(&history).path + len;
584 while(hp > sliceLast(&history).path) {
585 hp--;
586 if (*hp == '/') {
587 *hp = 0;
588 asprintf(&path, "%s/%s", sliceLast(&history).path, p);
589 *hp = '/';
590 goto downloadPage;
591 }
592 }
593 // no slash found in history path
594 asprintf(&path, "/%s", p);
595 }
596 else
597 // add path to previous path
598 asprintf(&path, "%s/%s", sliceLast(&history).path, p);
599 }
600 else {
601 // absolute path starting with /
602 path = strdup(p);
603 }
604 goto downloadPage;
605 }
606
607 // this part is the same as the parseURL function
608 s = cursor + strlen("spartan://");
609 cursor = s;
610 // scan hostname
611 while(*cursor != '/' &&
612 *cursor != ' ' &&
613 *cursor != '\t' &&
614 *cursor != ':' &&
615 *cursor != '\n' &&
616 *cursor != 0 &&
617 cursor < &page[links[link][1]]) {
618 cursor++;
619 }
620 hostname = malloc(cursor-s+1);
621 memcpy(hostname, s, cursor-s);
622 hostname[cursor-s] = 0;
623
624 if (*cursor == ':') {
625 // port is specified
626 cursor++;
627 port = cursor;
628 while(isdigit(*cursor)) {
629 cursor++;
630 }
631 char c = *cursor;
632 *cursor = 0;
633 port = strdup(port);
634 *cursor = c;
635 }
636 else {
637 // default port is 300
638 port = strdup("300");
639 }
640
641 if (*cursor == ' ' ||
642 *cursor == '\t' ||
643 *cursor == '\n' ||
644 *cursor == 0) {
645 // path is empty
646 path = strdup("/");
647 }
648 else {
649 // *cursor == '/'
650 // scan path
651 s = cursor;
652 while(!isspace(*cursor)) {
653 cursor++;
654 }
655
656 if (cursor == s+1) {
657 // path is /
658 path = strdup("/");
659 }
660 else {
661 path = malloc(cursor-s+1);
662 memcpy(path, s, cursor-s);
663 path[cursor-s] = 0;
664 }
665 }
666 // ^^ this part is the same as the parseURL function ^^
667
668 downloadPage:
669 printf(BLD GRN "%s" RST "\n", &page[links[link][0]]);
670
671 // store hostname, port, path and link in history
672 e = (historyEt){.hostname = hostname, .port = port, .path = path, .link = strdup(&page[links[link][0]])};
673 if (getPage(hostname, port, path, 0 /*content length*/, NULL /*send data*/)) {
674 // success
675 sliceAppend(&history, e);
676 }
677 }
678 else if (userInput[0] == 'b' || userInput[0] == '-' || userInput[0] == '+') {
679 // go back in history
680 if (history.count < 2) {
681 puts("First page in history.");
682 continue;
683 }
684 else {
685 free(sliceLast(&history).hostname);
686 free(sliceLast(&history).port);
687 free(sliceLast(&history).path);
688 free(sliceLast(&history).link);
689 history.count--;
690 if (sliceLast(&history).link)
691 puts(sliceLast(&history).link);
692 getPage(sliceLast(&history).hostname, sliceLast(&history).port, sliceLast(&history).path, 0 /*content length*/, NULL /*send data*/);
693 }
694 }
695 else if (userInput[0] == 'z') {
696 // used for debug
697 puts("break");
698 }
699 }
700
701 free(page);
702 for (int i = 0; i < history.count ; i++) {
703 free(sliceAt(&history, i).hostname);
704 free(sliceAt(&history, i).port);
705 free(sliceAt(&history, i).path);
706 free(sliceAt(&history, i).link);
707 }
708 }
709