/* * Decode chunked data */ static Dstr *Decode_chunked(Decode *dc, const char *instr, int inlen) { char *inputPtr, *eol; int inputRemaining; int chunkRemaining = *((int *)dc->state); Dstr *output = dStr_sized_new(inlen); dStr_append_l(dc->leftover, instr, inlen); inputPtr = dc->leftover->str; inputRemaining = dc->leftover->len; while (inputRemaining > 0) { if (chunkRemaining > 2) { /* chunk body to copy */ int copylen = MIN(chunkRemaining - 2, inputRemaining); dStr_append_l(output, inputPtr, copylen); chunkRemaining -= copylen; inputRemaining -= copylen; inputPtr += copylen; } if ((chunkRemaining == 2) && (inputRemaining > 0)) { /* CR to discard */ chunkRemaining--; inputRemaining--; inputPtr++; } if ((chunkRemaining == 1) && (inputRemaining > 0)) { /* LF to discard */ chunkRemaining--; inputRemaining--; inputPtr++; } /* * A chunk has a one-line header that begins with the chunk length * in hexadecimal. */ if (!(eol = (char *)memchr(inputPtr, '\n', inputRemaining))) { break; /* We don't have the whole line yet. */ } if (!(chunkRemaining = strtol(inputPtr, NULL, 0x10))) { break; /* A chunk length of 0 means we're done! */ } inputRemaining -= (eol - inputPtr) + 1; inputPtr = eol + 1; chunkRemaining += 2; /* CRLF at the end of every chunk */ } /* If we have a partial chunk header, save it for next time. */ dStr_erase(dc->leftover, 0, inputPtr - dc->leftover->str); *(int *)dc->state = chunkRemaining; return output; }
/* * Decode gzipped data */ static Dstr *Decode_gzip(Decode *dc, const char *instr, int inlen) { int rc = Z_OK; z_stream *zs = (z_stream *)dc->state; int inputConsumed = 0; Dstr *output = dStr_new(""); while ((rc == Z_OK) && (inputConsumed < inlen)) { zs->next_in = (Bytef *)instr + inputConsumed; zs->avail_in = inlen - inputConsumed; zs->next_out = (Bytef *)dc->buffer; zs->avail_out = bufsize; rc = inflate(zs, Z_SYNC_FLUSH); dStr_append_l(output, dc->buffer, zs->total_out); if ((rc == Z_OK) || (rc == Z_STREAM_END)) { // Z_STREAM_END at end of file inputConsumed += zs->total_in; zs->total_out = 0; zs->total_in = 0; } else if (rc == Z_DATA_ERROR) { MSG_ERR("gzip decompression error\n"); } } return output; }
/* * Read all the available data from a filedescriptor. * This is intended for short answers, i.e. when we know the server * will write it all before being preempted. For answers that may come * as an stream with delays, non-blocking is better. * Return value: read data, or NULL on error and no data. */ static char *Dpi_blocking_read(int fd) { int st; const int buf_sz = 8*1024; char buf[buf_sz], *msg = NULL; Dstr *dstr = dStr_sized_new(buf_sz); do { st = read(fd, buf, buf_sz); if (st < 0) { if (errno == EINTR) { continue; } else { MSG_ERR("[Dpi_blocking_read] %s\n", dStrerror(errno)); break; } } else if (st > 0) { dStr_append_l(dstr, buf, st); } } while (st == buf_sz); msg = (dstr->len > 0) ? dstr->str : NULL; dStr_free(dstr, (dstr->len > 0) ? FALSE : TRUE); return msg; }
/* * If the url belongs to a dpi server, return its name. */ static int Capi_url_uses_dpi(DilloUrl *url, char **server_ptr) { char *p, *server = NULL, *url_str = URL_STR(url); Dstr *tmp; if ((dStrncasecmp(url_str, "http:", 5) == 0) || (dStrncasecmp(url_str, "about:", 6) == 0)) { /* URL doesn't use dpi (server = NULL) */ } else if (dStrncasecmp(url_str, "dpi:/", 5) == 0) { /* dpi prefix, get this server's name */ if ((p = strchr(url_str + 5, '/')) != NULL) { server = dStrndup(url_str + 5, (uint_t)(p - url_str - 5)); } else { server = dStrdup("?"); } if (strcmp(server, "bm") == 0) { dFree(server); server = dStrdup("bookmarks"); } } else if ((p = strchr(url_str, ':')) != NULL) { tmp = dStr_new("proto."); dStr_append_l(tmp, url_str, p - url_str); server = tmp->str; dStr_free(tmp, 0); } return ((*server_ptr = server) ? 1 : 0); }
/* * Return value: 1..DataSize sent, -1 eagain, or -3 on big Error */ int a_Dpip_dsh_trywrite(Dsh *dsh, const char *Data, int DataSize) { int st; if ((st = Dpip_dsh_write(dsh, 1, Data, DataSize)) > 0) { /* update internal buffer */ if (st < DataSize) dStr_append_l(dsh->wrbuf, Data + st, DataSize - st); } return st; }
/* * Translate to desired character set (UTF-8) */ static Dstr *Decode_charset(Decode *dc, const char *instr, int inlen) { inbuf_t *inPtr; char *outPtr; size_t inLeft, outRoom; Dstr *output = dStr_new(""); int rc = 0; dStr_append_l(dc->leftover, instr, inlen); inPtr = dc->leftover->str; inLeft = dc->leftover->len; while ((rc != EINVAL) && (inLeft > 0)) { outPtr = dc->buffer; outRoom = bufsize; rc = iconv((iconv_t)dc->state, &inPtr, &inLeft, &outPtr, &outRoom); // iconv() on success, number of bytes converted // -1, errno == EILSEQ illegal byte sequence found // EINVAL partial character ends source buffer // E2BIG destination buffer is full dStr_append_l(output, dc->buffer, bufsize - outRoom); if (rc == -1) rc = errno; if (rc == EILSEQ){ inPtr++; inLeft--; dStr_append_l(output, utf8_replacement_char, sizeof(utf8_replacement_char) - 1); } } dStr_erase(dc->leftover, 0, dc->leftover->len - inLeft); return output; }
/* * Inject full page content directly into the cache. * Used for "about:splash". May be used for "about:cache" too. */ static void Cache_entry_inject(const DilloUrl *Url, Dstr *data_ds) { CacheEntry_t *entry; if (!(entry = Cache_entry_search(Url))) entry = Cache_entry_add(Url); entry->Flags |= CA_GotData + CA_GotHeader + CA_GotLength + CA_InternalUrl; if (data_ds->len) entry->Flags &= ~CA_IsEmpty; dStr_truncate(entry->Data, 0); dStr_append_l(entry->Data, data_ds->str, data_ds->len); dStr_fit(entry->Data); entry->ExpectedSize = entry->TransferSize = entry->Data->len; }
/* * Read raw data from the socket into our buffer in * either BLOCKING or NONBLOCKING mode. */ static void Dpip_dsh_read(Dsh *dsh, int blocking) { char buf[RBUF_SZ]; int req_mode, old_flags = 0, st, ret = -3, nb = !blocking; dReturn_if (dsh->status == DPIP_ERROR || dsh->status == DPIP_EOF); req_mode = (nb) ? DPIP_NONBLOCK : 0; if ((dsh->mode & DPIP_NONBLOCK) != req_mode) { /* change mode temporarily... */ old_flags = fcntl(dsh->fd_in, F_GETFL); fcntl(dsh->fd_in, F_SETFL, (nb) ? O_NONBLOCK | old_flags : old_flags & ~O_NONBLOCK); } while (1) { st = read(dsh->fd_in, buf, RBUF_SZ); if (st < 0) { if (errno == EINTR) { continue; } else if (errno == EAGAIN) { dsh->status = DPIP_EAGAIN; ret = -1; break; } else { MSG_ERR("[Dpip_dsh_read] %s\n", dStrerror(errno)); dsh->status = DPIP_ERROR; break; } } else if (st == 0) { dsh->status = DPIP_EOF; break; } else { /* append to buf */ dStr_append_l(dsh->rdbuf, buf, st); if (blocking) break; } } if ((dsh->mode & DPIP_NONBLOCK) != req_mode) { /* restore old mode */ fcntl(dsh->fd_out, F_SETFL, old_flags); } /* assert there's no more data in the wire... * (st < buf upon interrupt || st == buf and no more data) */ if (blocking) Dpip_dsh_read(dsh, 0); }
/* * Streamed write to socket * Return: 0 on success, 1 on error. */ int a_Dpip_dsh_write(Dsh *dsh, int flush, const char *Data, int DataSize) { int ret = 1; /* append to buf */ dStr_append_l(dsh->wrbuf, Data, DataSize); if (!flush || dsh->wrbuf->len == 0) return 0; ret = Dpip_dsh_write(dsh, 0, dsh->wrbuf->str, dsh->wrbuf->len); if (ret == dsh->wrbuf->len) { dStr_truncate(dsh->wrbuf, 0); ret = 0; } return ret; }
/* * Generate Content-Type header value for a POST query. */ static Dstr *Http_make_content_type(const DilloUrl *url) { Dstr *dstr; if (URL_FLAGS(url) & URL_MultipartEnc) { MSG("submitting multipart/form-data!\n"); dstr = dStr_new("multipart/form-data; boundary=\""); if (URL_DATA(url)->len > 2) { /* boundary lines have "--" prepended. Skip that. */ const char *start = URL_DATA(url)->str + 2; char *eol = strchr(start, '\r'); if (eol) dStr_append_l(dstr, start, eol - start); } else { /* Zero parts; arbitrary boundary */ dStr_append_c(dstr, '0'); } dStr_append_c(dstr,'"'); } else { dstr = dStr_new("application/x-www-form-urlencoded"); } return dstr; }
/* * Make the http query string */ Dstr *a_Http_make_query_str(const DilloUrl *url, const DilloUrl *requester, bool_t use_proxy) { char *ptr, *cookies, *referer, *auth; Dstr *query = dStr_new(""), *request_uri = dStr_new(""), *proxy_auth = dStr_new(""); if (use_proxy) { dStr_sprintfa(request_uri, "%s%s", URL_STR(url), (URL_PATH_(url) || URL_QUERY_(url)) ? "" : "/"); if ((ptr = strrchr(request_uri->str, '#'))) dStr_truncate(request_uri, ptr - request_uri->str); if (HTTP_Proxy_Auth_base64) dStr_sprintf(proxy_auth, "Proxy-Authorization: Basic %s\r\n", HTTP_Proxy_Auth_base64); } else { dStr_sprintfa(request_uri, "%s%s%s%s", URL_PATH(url), URL_QUERY_(url) ? "?" : "", URL_QUERY(url), (URL_PATH_(url) || URL_QUERY_(url)) ? "" : "/"); } cookies = a_Cookies_get_query(url, requester); auth = a_Auth_get_auth_str(url, request_uri->str); referer = Http_get_referer(url); if (URL_FLAGS(url) & URL_Post) { Dstr *content_type = Http_make_content_type(url); dStr_sprintfa( query, "POST %s HTTP/1.1\r\n" "Connection: close\r\n" "Accept: text/*,image/*,*/*;q=0.2\r\n" "Accept-Charset: utf-8,*;q=0.8\r\n" "Accept-Encoding: gzip\r\n" "%s" /* language */ "%s" /* auth */ "Host: %s\r\n" "%s" "%s" "User-Agent: %s\r\n" "Content-Length: %ld\r\n" "Content-Type: %s\r\n" "%s" /* cookies */ "\r\n", request_uri->str, HTTP_Language_hdr, auth ? auth : "", URL_AUTHORITY(url), proxy_auth->str, referer, prefs.http_user_agent, (long)URL_DATA(url)->len, content_type->str, cookies); dStr_append_l(query, URL_DATA(url)->str, URL_DATA(url)->len); dStr_free(content_type, TRUE); } else { dStr_sprintfa( query, "GET %s HTTP/1.1\r\n" "%s" "Connection: close\r\n" "Accept: text/*,image/*,*/*;q=0.2\r\n" "Accept-Charset: utf-8,*;q=0.8\r\n" "Accept-Encoding: gzip\r\n" "%s" /* language */ "%s" /* auth */ "Host: %s\r\n" "%s" "%s" "User-Agent: %s\r\n" "%s" /* cookies */ "\r\n", request_uri->str, (URL_FLAGS(url) & URL_E2EQuery) ? "Cache-Control: no-cache\r\nPragma: no-cache\r\n" : "", HTTP_Language_hdr, auth ? auth : "", URL_AUTHORITY(url), proxy_auth->str, referer, prefs.http_user_agent, cookies); } dFree(referer); dFree(cookies); dFree(auth); dStr_free(request_uri, TRUE); dStr_free(proxy_auth, TRUE); _MSG("Query: {%s}\n", dStr_printable(query, 8192)); return query; }
/* * Receive new data, update the reception buffer (for next read), update the * cache, and service the client queue. * * This function gets called whenever the IO has new data. * 'Op' is the operation to perform * 'VPtr' is a (void) pointer to the IO control structure */ void a_Cache_process_dbuf(int Op, const char *buf, size_t buf_size, const DilloUrl *Url) { int offset, len; const char *str; Dstr *dstr1, *dstr2, *dstr3; CacheEntry_t *entry = Cache_entry_search(Url); /* Assert a valid entry (not aborted) */ dReturn_if_fail (entry != NULL); _MSG("__a_Cache_process_dbuf__\n"); if (Op == IORead) { /* * Cache_get_header() will set CA_GotHeader if it has a full header, and * Cache_parse_header() will unset it if the header ends being * merely an informational response from the server (i.e., 100 Continue) */ for (offset = 0; !(entry->Flags & CA_GotHeader) && (len = Cache_get_header(entry, buf + offset, buf_size - offset)); Cache_parse_header(entry) ) { offset += len; } if (entry->Flags & CA_GotHeader) { str = buf + offset; len = buf_size - offset; entry->TransferSize += len; dstr1 = dstr2 = dstr3 = NULL; /* Decode arrived data (<= 3 stages) */ if (entry->TransferDecoder) { dstr1 = a_Decode_process(entry->TransferDecoder, str, len); str = dstr1->str; len = dstr1->len; } if (entry->ContentDecoder) { dstr2 = a_Decode_process(entry->ContentDecoder, str, len); str = dstr2->str; len = dstr2->len; } dStr_append_l(entry->Data, str, len); if (entry->CharsetDecoder && entry->UTF8Data) { dstr3 = a_Decode_process(entry->CharsetDecoder, str, len); dStr_append_l(entry->UTF8Data, dstr3->str, dstr3->len); } dStr_free(dstr1, 1); dStr_free(dstr2, 1); dStr_free(dstr3, 1); if (entry->Data->len) entry->Flags &= ~CA_IsEmpty; entry = Cache_process_queue(entry); } } else if (Op == IOClose) { if ((entry->ExpectedSize || entry->TransferSize) && entry->TypeHdr == NULL) { MSG_HTTP("Message with a body lacked Content-Type header.\n"); } if ((entry->Flags & CA_GotLength) && (entry->ExpectedSize != entry->TransferSize)) { MSG_HTTP("Content-Length does NOT match message body,\n" " at: %s\n", URL_STR_(entry->Url)); MSG("entry->ExpectedSize = %d, entry->TransferSize = %d\n", entry->ExpectedSize, entry->TransferSize); } if (!entry->TransferSize && !(entry->Flags & CA_Redirect) && (entry->Flags & WEB_RootUrl)) { char *eol = strchr(entry->Header->str, '\n'); if (eol) { char *status_line = dStrndup(entry->Header->str, eol - entry->Header->str); MSG_HTTP("Body was empty. Server sent status: %s\n", status_line); dFree(status_line); } } entry->Flags |= CA_GotData; entry->Flags &= ~CA_Stopped; /* it may catch up! */ if (entry->TransferDecoder) { a_Decode_free(entry->TransferDecoder); entry->TransferDecoder = NULL; } if (entry->ContentDecoder) { a_Decode_free(entry->ContentDecoder); entry->ContentDecoder = NULL; } dStr_fit(entry->Data); /* fit buffer size! */ if ((entry = Cache_process_queue(entry))) { if (entry->Flags & CA_GotHeader) { Cache_unref_data(entry); } } } else if (Op == IOAbort) { /* unused */ MSG("a_Cache_process_dbuf Op = IOAbort; not implemented!\n"); } }
/* * Receive new data, update the reception buffer (for next read), update the * cache, and service the client queue. * * This function gets called whenever the IO has new data. * 'Op' is the operation to perform * 'VPtr' is a (void) pointer to the IO control structure */ bool_t a_Cache_process_dbuf(int Op, const char *buf, size_t buf_size, const DilloUrl *Url) { int offset, len; const char *str; Dstr *dstr1, *dstr2, *dstr3; bool_t done = FALSE; CacheEntry_t *entry = Cache_entry_search(Url); /* Assert a valid entry (not aborted) */ dReturn_val_if_fail (entry != NULL, FALSE); _MSG("__a_Cache_process_dbuf__\n"); if (Op == IORead) { /* * Cache_get_header() will set CA_GotHeader if it has a full header, and * Cache_parse_header() will unset it if the header ends being * merely an informational response from the server (i.e., 100 Continue) */ for (offset = 0; !(entry->Flags & CA_GotHeader) && (len = Cache_get_header(entry, buf + offset, buf_size - offset)); Cache_parse_header(entry) ) { offset += len; } if (entry->Flags & CA_GotHeader) { str = buf + offset; len = buf_size - offset; entry->TransferSize += len; dstr1 = dstr2 = dstr3 = NULL; /* Decode arrived data (<= 3 stages) */ if (entry->TransferDecoder) { dstr1 = a_Decode_transfer_process(entry->TransferDecoder, str,len); done = a_Decode_transfer_finished(entry->TransferDecoder); str = dstr1->str; len = dstr1->len; } if (entry->ContentDecoder) { dstr2 = a_Decode_process(entry->ContentDecoder, str, len); str = dstr2->str; len = dstr2->len; } dStr_append_l(entry->Data, str, len); if (entry->CharsetDecoder && entry->UTF8Data) { dstr3 = a_Decode_process(entry->CharsetDecoder, str, len); dStr_append_l(entry->UTF8Data, dstr3->str, dstr3->len); } dStr_free(dstr1, 1); dStr_free(dstr2, 1); dStr_free(dstr3, 1); if (entry->Data->len) entry->Flags &= ~CA_IsEmpty; if ((entry->Flags & CA_GotLength) && (entry->TransferSize >= entry->ExpectedSize)) { done = TRUE; } if (!(entry->Flags & CA_KeepAlive)) { /* Let IOClose finish it later */ done = FALSE; } entry = Cache_process_queue(entry); if (entry && done) Cache_finish_msg(entry); } } else if (Op == IOClose) { Cache_finish_msg(entry); } else if (Op == IOAbort) { int i; CacheClient_t *Client; for (i = 0; (Client = dList_nth_data(ClientQueue, i)); ++i) { if (Client->Url == entry->Url) { DilloWeb *web = (DilloWeb *)Client->Web; a_Bw_remove_client(web->bw, Client->Key); Cache_client_dequeue(Client); --i; /* Keep the index value in the next iteration */ } } } return done; }