X-Git-Url: https://git.librecmc.org/?a=blobdiff_plain;f=networking%2Fwget.c;h=4eafebe40119c704c39d9f8e64d0ef3c6e2c265c;hb=a38f9faa9fa230eb3753381c4f626acf029379fb;hp=3a4be98783380e5790788302c0d9e80df9f23a8d;hpb=ca18311d0ae16a96e988df15a9009095f93df85e;p=oweals%2Fbusybox.git diff --git a/networking/wget.c b/networking/wget.c index 3a4be9878..4eafebe40 100644 --- a/networking/wget.c +++ b/networking/wget.c @@ -13,8 +13,9 @@ //usage: IF_FEATURE_WGET_LONG_OPTIONS( //usage: "[-c|--continue] [-s|--spider] [-q|--quiet] [-O|--output-document FILE]\n" //usage: " [--header 'header: value'] [-Y|--proxy on/off] [-P DIR]\n" -//usage: " [--no-check-certificate] [-U|--user-agent AGENT]" -//usage: IF_FEATURE_WGET_TIMEOUT(" [-T SEC]") " URL..." +/* Since we ignore these opts, we don't show them in --help */ +/* //usage: " [--no-check-certificate] [--no-cache]" */ +//usage: " [-U|--user-agent AGENT]" IF_FEATURE_WGET_TIMEOUT(" [-T SEC]") " URL..." //usage: ) //usage: IF_NOT_FEATURE_WGET_LONG_OPTIONS( //usage: "[-csq] [-O FILE] [-Y on/off] [-P DIR] [-U AGENT]" @@ -22,7 +23,6 @@ //usage: ) //usage:#define wget_full_usage "\n\n" //usage: "Retrieve files via HTTP or FTP\n" -//usage: "\nOptions:" //usage: "\n -s Spider mode - only check file existence" //usage: "\n -c Continue retrieval of aborted transfer" //usage: "\n -q Quiet" @@ -36,8 +36,11 @@ #include "libbb.h" -//#define log_io(...) bb_error_msg(__VA_ARGS__) -#define log_io(...) ((void)0) +#if 0 +# define log_io(...) bb_error_msg(__VA_ARGS__) +#else +# define log_io(...) ((void)0) +#endif struct host_info { @@ -299,8 +302,13 @@ static void parse_url(const char *src_url, struct host_info *h) sp = strrchr(h->host, '@'); if (sp != NULL) { - h->user = h->host; + // URL-decode "user:password" string before base64-encoding: + // wget http://test:my%20pass@example.com should send + // Authorization: Basic dGVzdDpteSBwYXNz + // which decodes to "test:my pass". + // Standard wget and curl do this too. *sp = '\0'; + h->user = percent_decode_in_place(h->host, /*strict:*/ 0); h->host = sp + 1; } @@ -312,8 +320,6 @@ static char *gethdr(FILE *fp) char *s, *hdrval; int c; - /* *istrunc = 0; */ - /* retrieve header line */ c = fgets_and_trim(fp); @@ -344,6 +350,15 @@ static char *gethdr(FILE *fp) return hdrval; } +static void reset_beg_range_to_zero(void) +{ + bb_error_msg("restart failed"); + G.beg_range = 0; + xlseek(G.output_fd, 0, SEEK_SET); + /* Done at the end instead: */ + /* ftruncate(G.output_fd, 0); */ +} + static FILE* prepare_ftp_session(FILE **dfpp, struct host_info *target, len_and_sockaddr *lsa) { FILE *sfp; @@ -411,10 +426,12 @@ static FILE* prepare_ftp_session(FILE **dfpp, struct host_info *target, len_and_ *dfpp = open_socket(lsa); - if (G.beg_range) { + if (G.beg_range != 0) { sprintf(G.wget_buf, "REST %"OFF_FMT"u", G.beg_range); if (ftpcmd(G.wget_buf, NULL, sfp) == 350) G.content_len -= G.beg_range; + else + reset_beg_range_to_zero(); } if (ftpcmd("RETR ", target->path, sfp) > 150) @@ -427,7 +444,7 @@ static void NOINLINE retrieve_file_data(FILE *dfp) { #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT # if ENABLE_FEATURE_WGET_TIMEOUT - unsigned second_cnt; + unsigned second_cnt = G.timeout_seconds; # endif struct pollfd polldata; @@ -448,7 +465,7 @@ static void NOINLINE retrieve_file_data(FILE *dfp) * which messes up progress bar and/or timeout logic. * Because of nonblocking I/O, we need to dance * very carefully around EAGAIN. See explanation at - * clearerr() call. + * clearerr() calls. */ ndelay_on(polldata.fd); #endif @@ -456,32 +473,7 @@ static void NOINLINE retrieve_file_data(FILE *dfp) int n; unsigned rdsz; - rdsz = sizeof(G.wget_buf); - if (G.got_clen) { - if (G.content_len < (off_t)sizeof(G.wget_buf)) { - if ((int)G.content_len <= 0) - break; - rdsz = (unsigned)G.content_len; - } - } - #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT -# if ENABLE_FEATURE_WGET_TIMEOUT - second_cnt = G.timeout_seconds; -# endif - while (1) { - if (safe_poll(&polldata, 1, 1000) != 0) - break; /* error, EOF, or data is available */ -# if ENABLE_FEATURE_WGET_TIMEOUT - if (second_cnt != 0 && --second_cnt == 0) { - progress_meter(PROGRESS_END); - bb_error_msg_and_die("download timed out"); - } -# endif - /* Needed for "stalled" indicator */ - progress_meter(PROGRESS_BUMP); - } - /* fread internally uses read loop, which in our case * is usually exited when we get EAGAIN. * In this case, libc sets error marker on the stream. @@ -491,36 +483,71 @@ static void NOINLINE retrieve_file_data(FILE *dfp) * into if (n <= 0) ... */ clearerr(dfp); - errno = 0; #endif + errno = 0; + rdsz = sizeof(G.wget_buf); + if (G.got_clen) { + if (G.content_len < (off_t)sizeof(G.wget_buf)) { + if ((int)G.content_len <= 0) + break; + rdsz = (unsigned)G.content_len; + } + } n = fread(G.wget_buf, 1, rdsz, dfp); - /* man fread: + + if (n > 0) { + xwrite(G.output_fd, G.wget_buf, n); +#if ENABLE_FEATURE_WGET_STATUSBAR + G.transferred += n; +#endif + if (G.got_clen) { + G.content_len -= n; + if (G.content_len == 0) + break; + } +#if ENABLE_FEATURE_WGET_TIMEOUT + second_cnt = G.timeout_seconds; +#endif + continue; + } + + /* n <= 0. + * man fread: * If error occurs, or EOF is reached, the return value * is a short item count (or zero). * fread does not distinguish between EOF and error. */ - if (n <= 0) { -#if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT - if (errno == EAGAIN) /* poll lied, there is no data? */ - continue; /* yes */ -#endif - if (ferror(dfp)) + if (errno != EAGAIN) { + if (ferror(dfp)) { + progress_meter(PROGRESS_END); bb_perror_msg_and_die(bb_msg_read_error); + } break; /* EOF, not error */ } - xwrite(G.output_fd, G.wget_buf, n); - -#if ENABLE_FEATURE_WGET_STATUSBAR - G.transferred += n; +#if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT + /* It was EAGAIN. There is no data. Wait up to one second + * then abort if timed out, or update the bar and try reading again. + */ + if (safe_poll(&polldata, 1, 1000) == 0) { +# if ENABLE_FEATURE_WGET_TIMEOUT + if (second_cnt != 0 && --second_cnt == 0) { + progress_meter(PROGRESS_END); + bb_error_msg_and_die("download timed out"); + } +# endif + /* We used to loop back to poll here, + * but there is no great harm in letting fread + * to try reading anyway. + */ + } + /* Need to do it _every_ second for "stalled" indicator + * to be shown properly. + */ progress_meter(PROGRESS_BUMP); #endif - if (G.got_clen) { - G.content_len -= n; - if (G.content_len == 0) - break; - } - } + } /* while (reading data) */ + #if ENABLE_FEATURE_WGET_STATUSBAR || ENABLE_FEATURE_WGET_TIMEOUT clearerr(dfp); ndelay_off(polldata.fd); /* else fgets can get very unhappy */ @@ -536,6 +563,24 @@ static void NOINLINE retrieve_file_data(FILE *dfp) if (G.content_len == 0) break; /* all done! */ G.got_clen = 1; + /* + * Note that fgets may result in some data being buffered in dfp. + * We loop back to fread, which will retrieve this data. + * Also note that code has to be arranged so that fread + * is done _before_ one-second poll wait - poll doesn't know + * about stdio buffering and can result in spurious one second waits! + */ + } + + /* If -c failed, we restart from the beginning, + * but we do not truncate file then, we do it only now, at the end. + * This lets user to ^C if his 99% complete 10 GB file download + * failed to restart *without* losing the almost complete file. + */ + { + off_t pos = lseek(G.output_fd, 0, SEEK_CUR); + if (pos != (off_t)-1) + ftruncate(G.output_fd, pos); } /* Draw full bar and free its resources */ @@ -553,6 +598,7 @@ static void download_one_url(const char *url) FILE *dfp; /* socket to ftp server (data) */ char *proxy = NULL; char *fname_out_alloc; + char *redirected_path = NULL; struct host_info server; struct host_info target; @@ -592,13 +638,11 @@ static void download_one_url(const char *url) if (G.fname_out[0] == '/' || !G.fname_out[0]) G.fname_out = (char*)"index.html"; /* -P DIR is considered only if there was no -O FILE */ + if (G.dir_prefix) + G.fname_out = fname_out_alloc = concat_path_file(G.dir_prefix, G.fname_out); else { - if (G.dir_prefix) - G.fname_out = fname_out_alloc = concat_path_file(G.dir_prefix, G.fname_out); - else { - /* redirects may free target.path later, need to make a copy */ - G.fname_out = fname_out_alloc = xstrdup(G.fname_out); - } + /* redirects may free target.path later, need to make a copy */ + G.fname_out = fname_out_alloc = xstrdup(G.fname_out); } } #if ENABLE_FEATURE_WGET_STATUSBAR @@ -670,7 +714,7 @@ static void download_one_url(const char *url) } #endif - if (G.beg_range) + if (G.beg_range != 0) fprintf(sfp, "Range: bytes=%"OFF_FMT"u-\r\n", G.beg_range); #if ENABLE_FEATURE_WGET_LONG_OPTIONS @@ -737,15 +781,23 @@ However, in real world it was observed that some web servers (e.g. Boa/0.94.14rc21) simply use code 204 when file size is zero. */ case 204: + if (G.beg_range != 0) { + /* "Range:..." was not honored by the server. + * Restart download from the beginning. + */ + reset_beg_range_to_zero(); + } break; case 300: /* redirection */ case 301: case 302: case 303: break; - case 206: - if (G.beg_range) + case 206: /* Partial Content */ + if (G.beg_range != 0) + /* "Range:..." worked. Good. */ break; + /* Partial Content even though we did not ask for it??? */ /* fall through */ default: bb_error_msg_and_die("server returned error: %s", sanitize_string(G.wget_buf)); @@ -789,8 +841,8 @@ However, in real world it was observed that some web servers bb_error_msg_and_die("too many redirections"); fclose(sfp); if (str[0] == '/') { - free(target.allocated); - target.path = target.allocated = xstrdup(str+1); + free(redirected_path); + target.path = redirected_path = xstrdup(str+1); /* lsa stays the same: it's on the same server */ } else { parse_url(str, &target); @@ -845,6 +897,7 @@ However, in real world it was observed that some web servers free(server.allocated); free(target.allocated); free(fname_out_alloc); + free(redirected_path); } int wget_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE; @@ -872,6 +925,8 @@ int wget_main(int argc UNUSED_PARAM, char **argv) "post-data\0" Required_argument "\xfd" /* Ignored (we don't do ssl) */ "no-check-certificate\0" No_argument "\xfc" + /* Ignored (we don't support caching) */ + "no-cache\0" No_argument "\xfb" ; #endif