From b1e647baf2e1724838e34d25339a3672caf09e14 Mon Sep 17 00:00:00 2001 From: Andreas Baumann Date: Thu, 12 Jul 2012 19:18:06 +0200 Subject: added basic structure --- libfetch/GNUmakefile | 36 + libfetch/Makefile | 83 +++ libfetch/common.c | 907 +++++++++++++++++++++++ libfetch/common.h | 132 ++++ libfetch/fetch.3 | 725 ++++++++++++++++++ libfetch/fetch.c | 469 ++++++++++++ libfetch/fetch.h | 151 ++++ libfetch/file.c | 149 ++++ libfetch/ftp.c | 1206 ++++++++++++++++++++++++++++++ libfetch/ftp.errors | 47 ++ libfetch/http.c | 2005 ++++++++++++++++++++++++++++++++++++++++++++++++++ libfetch/http.errors | 45 ++ 12 files changed, 5955 insertions(+) create mode 100644 libfetch/GNUmakefile create mode 100644 libfetch/Makefile create mode 100644 libfetch/common.c create mode 100644 libfetch/common.h create mode 100644 libfetch/fetch.3 create mode 100644 libfetch/fetch.c create mode 100644 libfetch/fetch.h create mode 100644 libfetch/file.c create mode 100644 libfetch/ftp.c create mode 100644 libfetch/ftp.errors create mode 100644 libfetch/http.c create mode 100644 libfetch/http.errors (limited to 'libfetch') diff --git a/libfetch/GNUmakefile b/libfetch/GNUmakefile new file mode 100644 index 0000000..8088aca --- /dev/null +++ b/libfetch/GNUmakefile @@ -0,0 +1,36 @@ +TOPDIR = .. + +SUBDIRS = + +-include $(TOPDIR)/makefiles/gmake/platform.mk + +INCLUDE_CXXFLAGS = + +INCLUDE_LDFLAGS = \ + +INCLUDE_DIRS = \ + +INCLUDE_LIBS = \ + +OBJS = \ + common.o \ + fetch.o \ + file.o \ + ftp.o \ + http.o + +CPP_BINS = + +-include $(TOPDIR)/makefiles/gmake/sub.mk + +local_all: + +local_clean: + +local_distclean: + +local_install: + +local_uninstall: + +local_test: diff --git a/libfetch/Makefile b/libfetch/Makefile new file mode 100644 index 0000000..7b29aa2 --- /dev/null +++ b/libfetch/Makefile @@ -0,0 +1,83 @@ +# $FreeBSD$ + +.include + +LIB= fetch +CFLAGS+= -I. +SRCS= fetch.c common.c ftp.c http.c file.c \ + ftperr.h httperr.h +INCS= fetch.h +MAN= fetch.3 +CLEANFILES= ftperr.h httperr.h + +.if ${MK_INET6_SUPPORT} != "no" +CFLAGS+= -DINET6 +.endif + +.if ${MK_OPENSSL} != "no" +CFLAGS+= -DWITH_SSL +DPADD= ${LIBSSL} ${LIBCRYPTO} ${LIBMD} +LDADD= -lssl -lcrypto -lmd +.else +DPADD= ${LIBMD} +LDADD= -lmd +.endif + +CFLAGS+= -DFTP_COMBINE_CWDS + +CSTD?= c99 + +SHLIB_MAJOR= 6 + +ftperr.h: ftp.errors ${.CURDIR}/Makefile + @echo "static struct fetcherr ftp_errlist[] = {" > ${.TARGET} + @cat ${.CURDIR}/ftp.errors \ + | grep -v ^# \ + | sort \ + | while read NUM CAT STRING; do \ + echo " { $${NUM}, FETCH_$${CAT}, \"$${STRING}\" },"; \ + done >> ${.TARGET} + @echo " { -1, FETCH_UNKNOWN, \"Unknown FTP error\" }" >> ${.TARGET} + @echo "};" >> ${.TARGET} + +httperr.h: http.errors ${.CURDIR}/Makefile + @echo "static struct fetcherr http_errlist[] = {" > ${.TARGET} + @cat ${.CURDIR}/http.errors \ + | grep -v ^# \ + | sort \ + | while read NUM CAT STRING; do \ + echo " { $${NUM}, FETCH_$${CAT}, \"$${STRING}\" },"; \ + done >> ${.TARGET} + @echo " { -1, FETCH_UNKNOWN, \"Unknown HTTP error\" }" >> ${.TARGET} + @echo "};" >> ${.TARGET} + +MLINKS+= fetch.3 fetchFreeURL.3 +MLINKS+= fetch.3 fetchGet.3 +MLINKS+= fetch.3 fetchGetFTP.3 +MLINKS+= fetch.3 fetchGetFile.3 +MLINKS+= fetch.3 fetchGetHTTP.3 +MLINKS+= fetch.3 fetchGetURL.3 +MLINKS+= fetch.3 fetchList.3 +MLINKS+= fetch.3 fetchListFTP.3 +MLINKS+= fetch.3 fetchListFile.3 +MLINKS+= fetch.3 fetchListHTTP.3 +MLINKS+= fetch.3 fetchListURL.3 +MLINKS+= fetch.3 fetchMakeURL.3 +MLINKS+= fetch.3 fetchParseURL.3 +MLINKS+= fetch.3 fetchPut.3 +MLINKS+= fetch.3 fetchPutFTP.3 +MLINKS+= fetch.3 fetchPutFile.3 +MLINKS+= fetch.3 fetchPutHTTP.3 +MLINKS+= fetch.3 fetchPutURL.3 +MLINKS+= fetch.3 fetchStat.3 +MLINKS+= fetch.3 fetchStatFTP.3 +MLINKS+= fetch.3 fetchStatFile.3 +MLINKS+= fetch.3 fetchStatHTTP.3 +MLINKS+= fetch.3 fetchStatURL.3 +MLINKS+= fetch.3 fetchXGet.3 +MLINKS+= fetch.3 fetchXGetFTP.3 +MLINKS+= fetch.3 fetchXGetFile.3 +MLINKS+= fetch.3 fetchXGetHTTP.3 +MLINKS+= fetch.3 fetchXGetURL.3 + +.include diff --git a/libfetch/common.c b/libfetch/common.c new file mode 100644 index 0000000..425b6de --- /dev/null +++ b/libfetch/common.c @@ -0,0 +1,907 @@ +/*- + * Copyright (c) 1998-2011 Dag-Erling Smørgrav + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. The name of the author may not be used to endorse or promote products + * derived from this software without specific prior written permission + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +#include +#include +#include +#include + +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "fetch.h" +#include "common.h" + + +/*** Local data **************************************************************/ + +/* + * Error messages for resolver errors + */ +static struct fetcherr netdb_errlist[] = { +#ifdef EAI_NODATA + { EAI_NODATA, FETCH_RESOLV, "Host not found" }, +#endif + { EAI_AGAIN, FETCH_TEMP, "Transient resolver failure" }, + { EAI_FAIL, FETCH_RESOLV, "Non-recoverable resolver failure" }, + { EAI_NONAME, FETCH_RESOLV, "No address record" }, + { -1, FETCH_UNKNOWN, "Unknown resolver error" } +}; + +/* End-of-Line */ +static const char ENDL[2] = "\r\n"; + + +/*** Error-reporting functions ***********************************************/ + +/* + * Map error code to string + */ +static struct fetcherr * +fetch_finderr(struct fetcherr *p, int e) +{ + while (p->num != -1 && p->num != e) + p++; + return (p); +} + +/* + * Set error code + */ +void +fetch_seterr(struct fetcherr *p, int e) +{ + p = fetch_finderr(p, e); + fetchLastErrCode = p->cat; + snprintf(fetchLastErrString, MAXERRSTRING, "%s", p->string); +} + +/* + * Set error code according to errno + */ +void +fetch_syserr(void) +{ + switch (errno) { + case 0: + fetchLastErrCode = FETCH_OK; + break; + case EPERM: + case EACCES: + case EROFS: + case EAUTH: + case ENEEDAUTH: + fetchLastErrCode = FETCH_AUTH; + break; + case ENOENT: + case EISDIR: /* XXX */ + fetchLastErrCode = FETCH_UNAVAIL; + break; + case ENOMEM: + fetchLastErrCode = FETCH_MEMORY; + break; + case EBUSY: + case EAGAIN: + fetchLastErrCode = FETCH_TEMP; + break; + case EEXIST: + fetchLastErrCode = FETCH_EXISTS; + break; + case ENOSPC: + fetchLastErrCode = FETCH_FULL; + break; + case EADDRINUSE: + case EADDRNOTAVAIL: + case ENETDOWN: + case ENETUNREACH: + case ENETRESET: + case EHOSTUNREACH: + fetchLastErrCode = FETCH_NETWORK; + break; + case ECONNABORTED: + case ECONNRESET: + fetchLastErrCode = FETCH_ABORT; + break; + case ETIMEDOUT: + fetchLastErrCode = FETCH_TIMEOUT; + break; + case ECONNREFUSED: + case EHOSTDOWN: + fetchLastErrCode = FETCH_DOWN; + break; +default: + fetchLastErrCode = FETCH_UNKNOWN; + } + snprintf(fetchLastErrString, MAXERRSTRING, "%s", strerror(errno)); +} + + +/* + * Emit status message + */ +void +fetch_info(const char *fmt, ...) +{ + va_list ap; + + va_start(ap, fmt); + vfprintf(stderr, fmt, ap); + va_end(ap); + fputc('\n', stderr); +} + + +/*** Network-related utility functions ***************************************/ + +/* + * Return the default port for a scheme + */ +int +fetch_default_port(const char *scheme) +{ + struct servent *se; + + if ((se = getservbyname(scheme, "tcp")) != NULL) + return (ntohs(se->s_port)); + if (strcasecmp(scheme, SCHEME_FTP) == 0) + return (FTP_DEFAULT_PORT); + if (strcasecmp(scheme, SCHEME_HTTP) == 0) + return (HTTP_DEFAULT_PORT); + return (0); +} + +/* + * Return the default proxy port for a scheme + */ +int +fetch_default_proxy_port(const char *scheme) +{ + if (strcasecmp(scheme, SCHEME_FTP) == 0) + return (FTP_DEFAULT_PROXY_PORT); + if (strcasecmp(scheme, SCHEME_HTTP) == 0) + return (HTTP_DEFAULT_PROXY_PORT); + return (0); +} + + +/* + * Create a connection for an existing descriptor. + */ +conn_t * +fetch_reopen(int sd) +{ + conn_t *conn; + int opt = 1; + + /* allocate and fill connection structure */ + if ((conn = calloc(1, sizeof(*conn))) == NULL) + return (NULL); + fcntl(sd, F_SETFD, FD_CLOEXEC); + setsockopt(sd, SOL_SOCKET, SO_NOSIGPIPE, &opt, sizeof opt); + conn->sd = sd; + ++conn->ref; + return (conn); +} + + +/* + * Bump a connection's reference count. + */ +conn_t * +fetch_ref(conn_t *conn) +{ + + ++conn->ref; + return (conn); +} + + +/* + * Bind a socket to a specific local address + */ +int +fetch_bind(int sd, int af, const char *addr) +{ + struct addrinfo hints, *res, *res0; + int err; + + memset(&hints, 0, sizeof(hints)); + hints.ai_family = af; + hints.ai_socktype = SOCK_STREAM; + hints.ai_protocol = 0; + if ((err = getaddrinfo(addr, NULL, &hints, &res0)) != 0) + return (-1); + for (res = res0; res; res = res->ai_next) + if (bind(sd, res->ai_addr, res->ai_addrlen) == 0) + return (0); + return (-1); +} + + +/* + * Establish a TCP connection to the specified port on the specified host. + */ +conn_t * +fetch_connect(const char *host, int port, int af, int verbose) +{ + conn_t *conn; + char pbuf[10]; + const char *bindaddr; + struct addrinfo hints, *res, *res0; + int sd, err; + + DEBUG(fprintf(stderr, "---> %s:%d\n", host, port)); + + if (verbose) + fetch_info("looking up %s", host); + + /* look up host name and set up socket address structure */ + snprintf(pbuf, sizeof(pbuf), "%d", port); + memset(&hints, 0, sizeof(hints)); + hints.ai_family = af; + hints.ai_socktype = SOCK_STREAM; + hints.ai_protocol = 0; + if ((err = getaddrinfo(host, pbuf, &hints, &res0)) != 0) { + netdb_seterr(err); + return (NULL); + } + bindaddr = getenv("FETCH_BIND_ADDRESS"); + + if (verbose) + fetch_info("connecting to %s:%d", host, port); + + /* try to connect */ + for (sd = -1, res = res0; res; sd = -1, res = res->ai_next) { + if ((sd = socket(res->ai_family, res->ai_socktype, + res->ai_protocol)) == -1) + continue; + if (bindaddr != NULL && *bindaddr != '\0' && + fetch_bind(sd, res->ai_family, bindaddr) != 0) { + fetch_info("failed to bind to '%s'", bindaddr); + close(sd); + continue; + } + if (connect(sd, res->ai_addr, res->ai_addrlen) == 0 && + fcntl(sd, F_SETFL, O_NONBLOCK) == 0) + break; + close(sd); + } + freeaddrinfo(res0); + if (sd == -1) { + fetch_syserr(); + return (NULL); + } + + if ((conn = fetch_reopen(sd)) == NULL) { + fetch_syserr(); + close(sd); + } + return (conn); +} + + +/* + * Enable SSL on a connection. + */ +int +fetch_ssl(conn_t *conn, int verbose) +{ +#ifdef WITH_SSL + int ret, ssl_err; + + /* Init the SSL library and context */ + if (!SSL_library_init()){ + fprintf(stderr, "SSL library init failed\n"); + return (-1); + } + + SSL_load_error_strings(); + + conn->ssl_meth = SSLv23_client_method(); + conn->ssl_ctx = SSL_CTX_new(conn->ssl_meth); + SSL_CTX_set_mode(conn->ssl_ctx, SSL_MODE_AUTO_RETRY); + + conn->ssl = SSL_new(conn->ssl_ctx); + if (conn->ssl == NULL){ + fprintf(stderr, "SSL context creation failed\n"); + return (-1); + } + SSL_set_fd(conn->ssl, conn->sd); + while ((ret = SSL_connect(conn->ssl)) == -1) { + ssl_err = SSL_get_error(conn->ssl, ret); + if (ssl_err != SSL_ERROR_WANT_READ && + ssl_err != SSL_ERROR_WANT_WRITE) { + ERR_print_errors_fp(stderr); + return (-1); + } + } + + if (verbose) { + X509_NAME *name; + char *str; + + fprintf(stderr, "SSL connection established using %s\n", + SSL_get_cipher(conn->ssl)); + conn->ssl_cert = SSL_get_peer_certificate(conn->ssl); + name = X509_get_subject_name(conn->ssl_cert); + str = X509_NAME_oneline(name, 0, 0); + printf("Certificate subject: %s\n", str); + free(str); + name = X509_get_issuer_name(conn->ssl_cert); + str = X509_NAME_oneline(name, 0, 0); + printf("Certificate issuer: %s\n", str); + free(str); + } + + return (0); +#else + (void)conn; + (void)verbose; + fprintf(stderr, "SSL support disabled\n"); + return (-1); +#endif +} + +#define FETCH_READ_WAIT -2 +#define FETCH_READ_ERROR -1 +#define FETCH_READ_DONE 0 + +#ifdef WITH_SSL +static ssize_t +fetch_ssl_read(SSL *ssl, char *buf, size_t len) +{ + ssize_t rlen; + int ssl_err; + + rlen = SSL_read(ssl, buf, len); + if (rlen < 0) { + ssl_err = SSL_get_error(ssl, rlen); + if (ssl_err == SSL_ERROR_WANT_READ || + ssl_err == SSL_ERROR_WANT_WRITE) { + return (FETCH_READ_WAIT); + } else { + ERR_print_errors_fp(stderr); + return (FETCH_READ_ERROR); + } + } + return (rlen); +} +#endif + +/* + * Cache some data that was read from a socket but cannot be immediately + * returned because of an interrupted system call. + */ +static int +fetch_cache_data(conn_t *conn, char *src, size_t nbytes) +{ + char *tmp; + + if (conn->cache.size < nbytes) { + tmp = realloc(conn->cache.buf, nbytes); + if (tmp == NULL) { + fetch_syserr(); + return (-1); + } + conn->cache.buf = tmp; + conn->cache.size = nbytes; + } + + memcpy(conn->cache.buf, src, nbytes); + conn->cache.len = nbytes; + conn->cache.pos = 0; + + return (0); +} + + +static ssize_t +fetch_socket_read(int sd, char *buf, size_t len) +{ + ssize_t rlen; + + rlen = read(sd, buf, len); + if (rlen < 0) { + if (errno == EAGAIN || (errno == EINTR && fetchRestartCalls)) + return (FETCH_READ_WAIT); + else + return (FETCH_READ_ERROR); + } + return (rlen); +} + +/* + * Read a character from a connection w/ timeout + */ +ssize_t +fetch_read(conn_t *conn, char *buf, size_t len) +{ + struct timeval now, timeout, delta; + fd_set readfds; + ssize_t rlen, total; + char *start; + + if (fetchTimeout > 0) { + gettimeofday(&timeout, NULL); + timeout.tv_sec += fetchTimeout; + } + + total = 0; + start = buf; + + if (conn->cache.len > 0) { + /* + * The last invocation of fetch_read was interrupted by a + * signal after some data had been read from the socket. Copy + * the cached data into the supplied buffer before trying to + * read from the socket again. + */ + total = (conn->cache.len < len) ? conn->cache.len : len; + memcpy(buf, conn->cache.buf, total); + + conn->cache.len -= total; + conn->cache.pos += total; + len -= total; + buf += total; + } + + while (len > 0) { + /* + * The socket is non-blocking. Instead of the canonical + * select() -> read(), we do the following: + * + * 1) call read() or SSL_read(). + * 2) if an error occurred, return -1. + * 3) if we received data but we still expect more, + * update our counters and loop. + * 4) if read() or SSL_read() signaled EOF, return. + * 5) if we did not receive any data but we're not at EOF, + * call select(). + * + * In the SSL case, this is necessary because if we + * receive a close notification, we have to call + * SSL_read() one additional time after we've read + * everything we received. + * + * In the non-SSL case, it may improve performance (very + * slightly) when reading small amounts of data. + */ +#ifdef WITH_SSL + if (conn->ssl != NULL) + rlen = fetch_ssl_read(conn->ssl, buf, len); + else +#endif + rlen = fetch_socket_read(conn->sd, buf, len); + if (rlen == 0) { + break; + } else if (rlen > 0) { + len -= rlen; + buf += rlen; + total += rlen; + continue; + } else if (rlen == FETCH_READ_ERROR) { + if (errno == EINTR) + fetch_cache_data(conn, start, total); + return (-1); + } + // assert(rlen == FETCH_READ_WAIT); + FD_ZERO(&readfds); + while (!FD_ISSET(conn->sd, &readfds)) { + FD_SET(conn->sd, &readfds); + if (fetchTimeout > 0) { + gettimeofday(&now, NULL); + if (!timercmp(&timeout, &now, >)) { + errno = ETIMEDOUT; + fetch_syserr(); + return (-1); + } + timersub(&timeout, &now, &delta); + } + errno = 0; + if (select(conn->sd + 1, &readfds, NULL, NULL, + fetchTimeout > 0 ? &delta : NULL) < 0) { + if (errno == EINTR) { + if (fetchRestartCalls) + continue; + /* Save anything that was read. */ + fetch_cache_data(conn, start, total); + } + fetch_syserr(); + return (-1); + } + } + } + return (total); +} + + +/* + * Read a line of text from a connection w/ timeout + */ +#define MIN_BUF_SIZE 1024 + +int +fetch_getln(conn_t *conn) +{ + char *tmp; + size_t tmpsize; + ssize_t len; + char c; + + if (conn->buf == NULL) { + if ((conn->buf = malloc(MIN_BUF_SIZE)) == NULL) { + errno = ENOMEM; + return (-1); + } + conn->bufsize = MIN_BUF_SIZE; + } + + conn->buf[0] = '\0'; + conn->buflen = 0; + + do { + len = fetch_read(conn, &c, 1); + if (len == -1) + return (-1); + if (len == 0) + break; + conn->buf[conn->buflen++] = c; + if (conn->buflen == conn->bufsize) { + tmp = conn->buf; + tmpsize = conn->bufsize * 2 + 1; + if ((tmp = realloc(tmp, tmpsize)) == NULL) { + errno = ENOMEM; + return (-1); + } + conn->buf = tmp; + conn->bufsize = tmpsize; + } + } while (c != '\n'); + + conn->buf[conn->buflen] = '\0'; + DEBUG(fprintf(stderr, "<<< %s", conn->buf)); + return (0); +} + + +/* + * Write to a connection w/ timeout + */ +ssize_t +fetch_write(conn_t *conn, const char *buf, size_t len) +{ + struct iovec iov; + + iov.iov_base = __DECONST(char *, buf); + iov.iov_len = len; + return fetch_writev(conn, &iov, 1); +} + +/* + * Write a vector to a connection w/ timeout + * Note: can modify the iovec. + */ +ssize_t +fetch_writev(conn_t *conn, struct iovec *iov, int iovcnt) +{ + struct timeval now, timeout, delta; + fd_set writefds; + ssize_t wlen, total; + int r; + + if (fetchTimeout) { + FD_ZERO(&writefds); + gettimeofday(&timeout, NULL); + timeout.tv_sec += fetchTimeout; + } + + total = 0; + while (iovcnt > 0) { + while (fetchTimeout && !FD_ISSET(conn->sd, &writefds)) { + FD_SET(conn->sd, &writefds); + gettimeofday(&now, NULL); + delta.tv_sec = timeout.tv_sec - now.tv_sec; + delta.tv_usec = timeout.tv_usec - now.tv_usec; + if (delta.tv_usec < 0) { + delta.tv_usec += 1000000; + delta.tv_sec--; + } + if (delta.tv_sec < 0) { + errno = ETIMEDOUT; + fetch_syserr(); + return (-1); + } + errno = 0; + r = select(conn->sd + 1, NULL, &writefds, NULL, &delta); + if (r == -1) { + if (errno == EINTR && fetchRestartCalls) + continue; + return (-1); + } + } + errno = 0; +#ifdef WITH_SSL + if (conn->ssl != NULL) + wlen = SSL_write(conn->ssl, + iov->iov_base, iov->iov_len); + else +#endif + wlen = writev(conn->sd, iov, iovcnt); + if (wlen == 0) { + /* we consider a short write a failure */ + /* XXX perhaps we shouldn't in the SSL case */ + errno = EPIPE; + fetch_syserr(); + return (-1); + } + if (wlen < 0) { + if (errno == EINTR && fetchRestartCalls) + continue; + return (-1); + } + total += wlen; + while (iovcnt > 0 && wlen >= (ssize_t)iov->iov_len) { + wlen -= iov->iov_len; + iov++; + iovcnt--; + } + if (iovcnt > 0) { + iov->iov_len -= wlen; + iov->iov_base = __DECONST(char *, iov->iov_base) + wlen; + } + } + return (total); +} + + +/* + * Write a line of text to a connection w/ timeout + */ +int +fetch_putln(conn_t *conn, const char *str, size_t len) +{ + struct iovec iov[2]; + int ret; + + DEBUG(fprintf(stderr, ">>> %s\n", str)); + iov[0].iov_base = __DECONST(char *, str); + iov[0].iov_len = len; + iov[1].iov_base = __DECONST(char *, ENDL); + iov[1].iov_len = sizeof(ENDL); + if (len == 0) + ret = fetch_writev(conn, &iov[1], 1); + else + ret = fetch_writev(conn, iov, 2); + if (ret == -1) + return (-1); + return (0); +} + + +/* + * Close connection + */ +int +fetch_close(conn_t *conn) +{ + int ret; + + if (--conn->ref > 0) + return (0); + ret = close(conn->sd); + free(conn->cache.buf); + free(conn->buf); + free(conn); + return (ret); +} + + +/*** Directory-related utility functions *************************************/ + +int +fetch_add_entry(struct url_ent **p, int *size, int *len, + const char *name, struct url_stat *us) +{ + struct url_ent *tmp; + + if (*p == NULL) { + *size = 0; + *len = 0; + } + + if (*len >= *size - 1) { + tmp = realloc(*p, (*size * 2 + 1) * sizeof(**p)); + if (tmp == NULL) { + errno = ENOMEM; + fetch_syserr(); + return (-1); + } + *size = (*size * 2 + 1); + *p = tmp; + } + + tmp = *p + *len; + snprintf(tmp->name, PATH_MAX, "%s", name); + memcpy(&tmp->stat, us, sizeof(*us)); + + (*len)++; + (++tmp)->name[0] = 0; + + return (0); +} + + +/*** Authentication-related utility functions ********************************/ + +static const char * +fetch_read_word(FILE *f) +{ + static char word[1024]; + + if (fscanf(f, " %1023s ", word) != 1) + return (NULL); + return (word); +} + +/* + * Get authentication data for a URL from .netrc + */ +int +fetch_netrc_auth(struct url *url) +{ + char fn[PATH_MAX]; + const char *word; + char *p; + FILE *f; + + if ((p = getenv("NETRC")) != NULL) { + if (snprintf(fn, sizeof(fn), "%s", p) >= (int)sizeof(fn)) { + fetch_info("$NETRC specifies a file name " + "longer than PATH_MAX"); + return (-1); + } + } else { + if ((p = getenv("HOME")) != NULL) { + struct passwd *pwd; + + if ((pwd = getpwuid(getuid())) == NULL || + (p = pwd->pw_dir) == NULL) + return (-1); + } + if (snprintf(fn, sizeof(fn), "%s/.netrc", p) >= (int)sizeof(fn)) + return (-1); + } + + if ((f = fopen(fn, "r")) == NULL) + return (-1); + while ((word = fetch_read_word(f)) != NULL) { + if (strcmp(word, "default") == 0) { + DEBUG(fetch_info("Using default .netrc settings")); + break; + } + if (strcmp(word, "machine") == 0 && + (word = fetch_read_word(f)) != NULL && + strcasecmp(word, url->host) == 0) { + DEBUG(fetch_info("Using .netrc settings for %s", word)); + break; + } + } + if (word == NULL) + goto ferr; + while ((word = fetch_read_word(f)) != NULL) { + if (strcmp(word, "login") == 0) { + if ((word = fetch_read_word(f)) == NULL) + goto ferr; + if (snprintf(url->user, sizeof(url->user), + "%s", word) > (int)sizeof(url->user)) { + fetch_info("login name in .netrc is too long"); + url->user[0] = '\0'; + } + } else if (strcmp(word, "password") == 0) { + if ((word = fetch_read_word(f)) == NULL) + goto ferr; + if (snprintf(url->pwd, sizeof(url->pwd), + "%s", word) > (int)sizeof(url->pwd)) { + fetch_info("password in .netrc is too long"); + url->pwd[0] = '\0'; + } + } else if (strcmp(word, "account") == 0) { + if ((word = fetch_read_word(f)) == NULL) + goto ferr; + /* XXX not supported! */ + } else { + break; + } + } + fclose(f); + return (0); + ferr: + fclose(f); + return (-1); +} + +/* + * The no_proxy environment variable specifies a set of domains for + * which the proxy should not be consulted; the contents is a comma-, + * or space-separated list of domain names. A single asterisk will + * override all proxy variables and no transactions will be proxied + * (for compatability with lynx and curl, see the discussion at + * ). + */ +int +fetch_no_proxy_match(const char *host) +{ + const char *no_proxy, *p, *q; + size_t h_len, d_len; + + if ((no_proxy = getenv("NO_PROXY")) == NULL && + (no_proxy = getenv("no_proxy")) == NULL) + return (0); + + /* asterisk matches any hostname */ + if (strcmp(no_proxy, "*") == 0) + return (1); + + h_len = strlen(host); + p = no_proxy; + do { + /* position p at the beginning of a domain suffix */ + while (*p == ',' || isspace((unsigned char)*p)) + p++; + + /* position q at the first separator character */ + for (q = p; *q; ++q) + if (*q == ',' || isspace((unsigned char)*q)) + break; + + d_len = q - p; + if (d_len > 0 && h_len >= d_len && + strncasecmp(host + h_len - d_len, + p, d_len) == 0) { + /* domain name matches */ + return (1); + } + + p = q + 1; + } while (*q); + + return (0); +} diff --git a/libfetch/common.h b/libfetch/common.h new file mode 100644 index 0000000..241dbbf --- /dev/null +++ b/libfetch/common.h @@ -0,0 +1,132 @@ +/*- + * Copyright (c) 1998-2011 Dag-Erling Smørgrav + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. The name of the author may not be used to endorse or promote products + * derived from this software without specific prior written permission + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * $FreeBSD$ + */ + +#ifndef _COMMON_H_INCLUDED +#define _COMMON_H_INCLUDED + +#define FTP_DEFAULT_PORT 21 +#define HTTP_DEFAULT_PORT 80 +#define FTP_DEFAULT_PROXY_PORT 21 +#define HTTP_DEFAULT_PROXY_PORT 3128 + +#ifdef WITH_SSL +#include +#include +#include +#include +#include +#endif + +/* Connection */ +typedef struct fetchconn conn_t; +struct fetchconn { + int sd; /* socket descriptor */ + char *buf; /* buffer */ + size_t bufsize; /* buffer size */ + size_t buflen; /* length of buffer contents */ + int err; /* last protocol reply code */ + struct { /* data cached after an interrupted + read */ + char *buf; + size_t size; + size_t pos; + size_t len; + } cache; +#ifdef WITH_SSL + SSL *ssl; /* SSL handle */ + SSL_CTX *ssl_ctx; /* SSL context */ + X509 *ssl_cert; /* server certificate */ + SSL_METHOD *ssl_meth; /* SSL method */ +#endif + int ref; /* reference count */ +}; + +/* Structure used for error message lists */ +struct fetcherr { + const int num; + const int cat; + const char *string; +}; + +/* for fetch_writev */ +struct iovec; + +void fetch_seterr(struct fetcherr *, int); +void fetch_syserr(void); +void fetch_info(const char *, ...); +int fetch_default_port(const char *); +int fetch_default_proxy_port(const char *); +int fetch_bind(int, int, const char *); +conn_t *fetch_connect(const char *, int, int, int); +conn_t *fetch_reopen(int); +conn_t *fetch_ref(conn_t *); +int fetch_ssl(conn_t *, int); +ssize_t fetch_read(conn_t *, char *, size_t); +int fetch_getln(conn_t *); +ssize_t fetch_write(conn_t *, const char *, size_t); +ssize_t fetch_writev(conn_t *, struct iovec *, int); +int fetch_putln(conn_t *, const char *, size_t); +int fetch_close(conn_t *); +int fetch_add_entry(struct url_ent **, int *, int *, + const char *, struct url_stat *); +int fetch_netrc_auth(struct url *url); +int fetch_no_proxy_match(const char *); + +#define ftp_seterr(n) fetch_seterr(ftp_errlist, n) +#define http_seterr(n) fetch_seterr(http_errlist, n) +#define netdb_seterr(n) fetch_seterr(netdb_errlist, n) +#define url_seterr(n) fetch_seterr(url_errlist, n) + +#ifndef NDEBUG +#define DEBUG(x) do { if (fetchDebug) { x; } } while (0) +#else +#define DEBUG(x) do { } while (0) +#endif + +/* + * I don't really like exporting http_request() and ftp_request(), + * but the HTTP and FTP code occasionally needs to cross-call + * eachother, and this saves me from adding a lot of special-case code + * to handle those cases. + * + * Note that _*_request() free purl, which is way ugly but saves us a + * whole lot of trouble. + */ +FILE *http_request(struct url *, const char *, + struct url_stat *, struct url *, const char *); +FILE *ftp_request(struct url *, const char *, + struct url_stat *, struct url *, const char *); + +/* + * Check whether a particular flag is set + */ +#define CHECK_FLAG(x) (flags && strchr(flags, (x))) + +#endif diff --git a/libfetch/fetch.3 b/libfetch/fetch.3 new file mode 100644 index 0000000..30372f2 --- /dev/null +++ b/libfetch/fetch.3 @@ -0,0 +1,725 @@ +.\"- +.\" Copyright (c) 1998-2011 Dag-Erling Smørgrav +.\" All rights reserved. +.\" +.\" Redistribution and use in source and binary forms, with or without +.\" modification, are permitted provided that the following conditions +.\" are met: +.\" 1. Redistributions of source code must retain the above copyright +.\" notice, this list of conditions and the following disclaimer. +.\" 2. Redistributions in binary form must reproduce the above copyright +.\" notice, this list of conditions and the following disclaimer in the +.\" documentation and/or other materials provided with the distribution. +.\" +.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND +.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE +.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS +.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) +.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT +.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY +.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF +.\" SUCH DAMAGE. +.\" +.\" $FreeBSD$ +.\" +.Dd September 27, 2011 +.Dt FETCH 3 +.Os +.Sh NAME +.Nm fetchMakeURL , +.Nm fetchParseURL , +.Nm fetchFreeURL , +.Nm fetchXGetURL , +.Nm fetchGetURL , +.Nm fetchPutURL , +.Nm fetchStatURL , +.Nm fetchListURL , +.Nm fetchXGet , +.Nm fetchGet , +.Nm fetchPut , +.Nm fetchStat , +.Nm fetchList , +.Nm fetchXGetFile , +.Nm fetchGetFile , +.Nm fetchPutFile , +.Nm fetchStatFile , +.Nm fetchListFile , +.Nm fetchXGetHTTP , +.Nm fetchGetHTTP , +.Nm fetchPutHTTP , +.Nm fetchStatHTTP , +.Nm fetchListHTTP , +.Nm fetchXGetFTP , +.Nm fetchGetFTP , +.Nm fetchPutFTP , +.Nm fetchStatFTP , +.Nm fetchListFTP +.Nd file transfer functions +.Sh LIBRARY +.Lb libfetch +.Sh SYNOPSIS +.In sys/param.h +.In stdio.h +.In fetch.h +.Ft struct url * +.Fn fetchMakeURL "const char *scheme" "const char *host" "int port" "const char *doc" "const char *user" "const char *pwd" +.Ft struct url * +.Fn fetchParseURL "const char *URL" +.Ft void +.Fn fetchFreeURL "struct url *u" +.Ft FILE * +.Fn fetchXGetURL "const char *URL" "struct url_stat *us" "const char *flags" +.Ft FILE * +.Fn fetchGetURL "const char *URL" "const char *flags" +.Ft FILE * +.Fn fetchPutURL "const char *URL" "const char *flags" +.Ft int +.Fn fetchStatURL "const char *URL" "struct url_stat *us" "const char *flags" +.Ft struct url_ent * +.Fn fetchListURL "const char *URL" "const char *flags" +.Ft FILE * +.Fn fetchXGet "struct url *u" "struct url_stat *us" "const char *flags" +.Ft FILE * +.Fn fetchGet "struct url *u" "const char *flags" +.Ft FILE * +.Fn fetchPut "struct url *u" "const char *flags" +.Ft int +.Fn fetchStat "struct url *u" "struct url_stat *us" "const char *flags" +.Ft struct url_ent * +.Fn fetchList "struct url *u" "const char *flags" +.Ft FILE * +.Fn fetchXGetFile "struct url *u" "struct url_stat *us" "const char *flags" +.Ft FILE * +.Fn fetchGetFile "struct url *u" "const char *flags" +.Ft FILE * +.Fn fetchPutFile "struct url *u" "const char *flags" +.Ft int +.Fn fetchStatFile "struct url *u" "struct url_stat *us" "const char *flags" +.Ft struct url_ent * +.Fn fetchListFile "struct url *u" "const char *flags" +.Ft FILE * +.Fn fetchXGetHTTP "struct url *u" "struct url_stat *us" "const char *flags" +.Ft FILE * +.Fn fetchGetHTTP "struct url *u" "const char *flags" +.Ft FILE * +.Fn fetchPutHTTP "struct url *u" "const char *flags" +.Ft int +.Fn fetchStatHTTP "struct url *u" "struct url_stat *us" "const char *flags" +.Ft struct url_ent * +.Fn fetchListHTTP "struct url *u" "const char *flags" +.Ft FILE * +.Fn fetchXGetFTP "struct url *u" "struct url_stat *us" "const char *flags" +.Ft FILE * +.Fn fetchGetFTP "struct url *u" "const char *flags" +.Ft FILE * +.Fn fetchPutFTP "struct url *u" "const char *flags" +.Ft int +.Fn fetchStatFTP "struct url *u" "struct url_stat *us" "const char *flags" +.Ft struct url_ent * +.Fn fetchListFTP "struct url *u" "const char *flags" +.Sh DESCRIPTION +These functions implement a high-level library for retrieving and +uploading files using Uniform Resource Locators (URLs). +.Pp +.Fn fetchParseURL +takes a URL in the form of a null-terminated string and splits it into +its components function according to the Common Internet Scheme Syntax +detailed in RFC1738. +A regular expression which produces this syntax is: +.Bd -literal + :(//((:)?@)?(:)?)?/()? +.Ed +.Pp +If the URL does not seem to begin with a scheme name, the following +syntax is assumed: +.Bd -literal + (((:)?@)?(:)?)?/()? +.Ed +.Pp +Note that some components of the URL are not necessarily relevant to +all URL schemes. +For instance, the file scheme only needs the and +components. +.Pp +.Fn fetchMakeURL +and +.Fn fetchParseURL +return a pointer to a +.Vt url +structure, which is defined as follows in +.In fetch.h : +.Bd -literal +#define URL_SCHEMELEN 16 +#define URL_USERLEN 256 +#define URL_PWDLEN 256 + +struct url { + char scheme[URL_SCHEMELEN+1]; + char user[URL_USERLEN+1]; + char pwd[URL_PWDLEN+1]; + char host[MAXHOSTNAMELEN+1]; + int port; + char *doc; + off_t offset; + size_t length; + time_t ims_time; +}; +.Ed +.Pp +The +.Va ims_time +field stores the time value for +.Li If-Modified-Since +HTTP requests. +.Pp +The pointer returned by +.Fn fetchMakeURL +or +.Fn fetchParseURL +should be freed using +.Fn fetchFreeURL . +.Pp +.Fn fetchXGetURL , +.Fn fetchGetURL , +and +.Fn fetchPutURL +constitute the recommended interface to the +.Nm fetch +library. +They examine the URL passed to them to determine the transfer +method, and call the appropriate lower-level functions to perform the +actual transfer. +.Fn fetchXGetURL +also returns the remote document's metadata in the +.Vt url_stat +structure pointed to by the +.Fa us +argument. +.Pp +The +.Fa flags +argument is a string of characters which specify transfer options. +The +meaning of the individual flags is scheme-dependent, and is detailed +in the appropriate section below. +.Pp +.Fn fetchStatURL +attempts to obtain the requested document's metadata and fill in the +structure pointed to by its second argument. +The +.Vt url_stat +structure is defined as follows in +.In fetch.h : +.Bd -literal +struct url_stat { + off_t size; + time_t atime; + time_t mtime; +}; +.Ed +.Pp +If the size could not be obtained from the server, the +.Fa size +field is set to -1. +If the modification time could not be obtained from the server, the +.Fa mtime +field is set to the epoch. +If the access time could not be obtained from the server, the +.Fa atime +field is set to the modification time. +.Pp +.Fn fetchListURL +attempts to list the contents of the directory pointed to by the URL +provided. +If successful, it returns a malloced array of +.Vt url_ent +structures. +The +.Vt url_ent +structure is defined as follows in +.In fetch.h : +.Bd -literal +struct url_ent { + char name[PATH_MAX]; + struct url_stat stat; +}; +.Ed +.Pp +The list is terminated by an entry with an empty name. +.Pp +The pointer returned by +.Fn fetchListURL +should be freed using +.Fn free . +.Pp +.Fn fetchXGet , +.Fn fetchGet , +.Fn fetchPut +and +.Fn fetchStat +are similar to +.Fn fetchXGetURL , +.Fn fetchGetURL , +.Fn fetchPutURL +and +.Fn fetchStatURL , +except that they expect a pre-parsed URL in the form of a pointer to +a +.Vt struct url +rather than a string. +.Pp +All of the +.Fn fetchXGetXXX , +.Fn fetchGetXXX +and +.Fn fetchPutXXX +functions return a pointer to a stream which can be used to read or +write data from or to the requested document, respectively. +Note that +although the implementation details of the individual access methods +vary, it can generally be assumed that a stream returned by one of the +.Fn fetchXGetXXX +or +.Fn fetchGetXXX +functions is read-only, and that a stream returned by one of the +.Fn fetchPutXXX +functions is write-only. +.Sh FILE SCHEME +.Fn fetchXGetFile , +.Fn fetchGetFile +and +.Fn fetchPutFile +provide access to documents which are files in a locally mounted file +system. +Only the component of the URL is used. +.Pp +.Fn fetchXGetFile +and +.Fn fetchGetFile +do not accept any flags. +.Pp +.Fn fetchPutFile +accepts the +.Ql a +(append to file) flag. +If that flag is specified, the data written to +the stream returned by +.Fn fetchPutFile +will be appended to the previous contents of the file, instead of +replacing them. +.Sh FTP SCHEME +.Fn fetchXGetFTP , +.Fn fetchGetFTP +and +.Fn fetchPutFTP +implement the FTP protocol as described in RFC959. +.Pp +If the +.Ql P +(not passive) flag is specified, an active (rather than passive) +connection will be attempted. +.Pp +The +.Ql p +flag is supported for compatibility with earlier versions where active +connections were the default. +It has precedence over the +.Ql P +flag, so if both are specified, +.Nm +will use a passive connection. +.Pp +If the +.Ql l +(low) flag is specified, data sockets will be allocated in the low (or +default) port range instead of the high port range (see +.Xr ip 4 ) . +.Pp +If the +.Ql d +(direct) flag is specified, +.Fn fetchXGetFTP , +.Fn fetchGetFTP +and +.Fn fetchPutFTP +will use a direct connection even if a proxy server is defined. +.Pp +If no user name or password is given, the +.Nm fetch +library will attempt an anonymous login, with user name "anonymous" +and password "anonymous@". +.Sh HTTP SCHEME +The +.Fn fetchXGetHTTP , +.Fn fetchGetHTTP +and +.Fn fetchPutHTTP +functions implement the HTTP/1.1 protocol. +With a little luck, there is +even a chance that they comply with RFC2616 and RFC2617. +.Pp +If the +.Ql d +(direct) flag is specified, +.Fn fetchXGetHTTP , +.Fn fetchGetHTTP +and +.Fn fetchPutHTTP +will use a direct connection even if a proxy server is defined. +.Pp +If the +.Ql i +(if-modified-since) flag is specified, and +the +.Va ims_time +field is set in +.Vt "struct url" , +then +.Fn fetchXGetHTTP +and +.Fn fetchGetHTTP +will send a conditional +.Li If-Modified-Since +HTTP header to only fetch the content if it is newer than +.Va ims_time . +.Pp +Since there seems to be no good way of implementing the HTTP PUT +method in a manner consistent with the rest of the +.Nm fetch +library, +.Fn fetchPutHTTP +is currently unimplemented. +.Sh AUTHENTICATION +Apart from setting the appropriate environment variables and +specifying the user name and password in the URL or the +.Vt struct url , +the calling program has the option of defining an authentication +function with the following prototype: +.Pp +.Ft int +.Fn myAuthMethod "struct url *u" +.Pp +The callback function should fill in the +.Fa user +and +.Fa pwd +fields in the provided +.Vt struct url +and return 0 on success, or any other value to indicate failure. +.Pp +To register the authentication callback, simply set +.Va fetchAuthMethod +to point at it. +The callback will be used whenever a site requires authentication and +the appropriate environment variables are not set. +.Pp +This interface is experimental and may be subject to change. +.Sh RETURN VALUES +.Fn fetchParseURL +returns a pointer to a +.Vt struct url +containing the individual components of the URL. +If it is +unable to allocate memory, or the URL is syntactically incorrect, +.Fn fetchParseURL +returns a NULL pointer. +.Pp +The +.Fn fetchStat +functions return 0 on success and -1 on failure. +.Pp +All other functions return a stream pointer which may be used to +access the requested document, or NULL if an error occurred. +.Pp +The following error codes are defined in +.In fetch.h : +.Bl -tag -width 18n +.It Bq Er FETCH_ABORT +Operation aborted +.It Bq Er FETCH_AUTH +Authentication failed +.It Bq Er FETCH_DOWN +Service unavailable +.It Bq Er FETCH_EXISTS +File exists +.It Bq Er FETCH_FULL +File system full +.It Bq Er FETCH_INFO +Informational response +.It Bq Er FETCH_MEMORY +Insufficient memory +.It Bq Er FETCH_MOVED +File has moved +.It Bq Er FETCH_NETWORK +Network error +.It Bq Er FETCH_OK +No error +.It Bq Er FETCH_PROTO +Protocol error +.It Bq Er FETCH_RESOLV +Resolver error +.It Bq Er FETCH_SERVER +Server error +.It Bq Er FETCH_TEMP +Temporary error +.It Bq Er FETCH_TIMEOUT +Operation timed out +.It Bq Er FETCH_UNAVAIL +File is not available +.It Bq Er FETCH_UNKNOWN +Unknown error +.It Bq Er FETCH_URL +Invalid URL +.El +.Pp +The accompanying error message includes a protocol-specific error code +and message, e.g.\& "File is not available (404 Not Found)" +.Sh ENVIRONMENT +.Bl -tag -width ".Ev FETCH_BIND_ADDRESS" +.It Ev FETCH_BIND_ADDRESS +Specifies a hostname or IP address to which sockets used for outgoing +connections will be bound. +.It Ev FTP_LOGIN +Default FTP login if none was provided in the URL. +.It Ev FTP_PASSIVE_MODE +If set to +.Ql no , +forces the FTP code to use active mode. +If set to any other value, forces passive mode even if the application +requested active mode. +.It Ev FTP_PASSWORD +Default FTP password if the remote server requests one and none was +provided in the URL. +.It Ev FTP_PROXY +URL of the proxy to use for FTP requests. +The document part is ignored. +FTP and HTTP proxies are supported; if no scheme is specified, FTP is +assumed. +If the proxy is an FTP proxy, +.Nm libfetch +will send +.Ql user@host +as user name to the proxy, where +.Ql user +is the real user name, and +.Ql host +is the name of the FTP server. +.Pp +If this variable is set to an empty string, no proxy will be used for +FTP requests, even if the +.Ev HTTP_PROXY +variable is set. +.It Ev ftp_proxy +Same as +.Ev FTP_PROXY , +for compatibility. +.It Ev HTTP_AUTH +Specifies HTTP authorization parameters as a colon-separated list of +items. +The first and second item are the authorization scheme and realm +respectively; further items are scheme-dependent. +Currently, the +.Dq basic +and +.Dq digest +authorization methods are supported. +.Pp +Both methods require two parameters: the user name and +password, in that order. +.Pp +This variable is only used if the server requires authorization and +no user name or password was specified in the URL. +.It Ev HTTP_PROXY +URL of the proxy to use for HTTP requests. +The document part is ignored. +Only HTTP proxies are supported for HTTP requests. +If no port number is specified, the default is 3128. +.Pp +Note that this proxy will also be used for FTP documents, unless the +.Ev FTP_PROXY +variable is set. +.It Ev http_proxy +Same as +.Ev HTTP_PROXY , +for compatibility. +.It Ev HTTP_PROXY_AUTH +Specifies authorization parameters for the HTTP proxy in the same +format as the +.Ev HTTP_AUTH +variable. +.Pp +This variable is used if and only if connected to an HTTP proxy, and +is ignored if a user and/or a password were specified in the proxy +URL. +.It Ev HTTP_REFERER +Specifies the referrer URL to use for HTTP requests. +If set to +.Dq auto , +the document URL will be used as referrer URL. +.It Ev HTTP_USER_AGENT +Specifies the User-Agent string to use for HTTP requests. +This can be useful when working with HTTP origin or proxy servers that +differentiate between user agents. +.It Ev NETRC +Specifies a file to use instead of +.Pa ~/.netrc +to look up login names and passwords for FTP sites. +See +.Xr ftp 1 +for a description of the file format. +This feature is experimental. +.It Ev NO_PROXY +Either a single asterisk, which disables the use of proxies +altogether, or a comma- or whitespace-separated list of hosts for +which proxies should not be used. +.It Ev no_proxy +Same as +.Ev NO_PROXY , +for compatibility. +.El +.Sh EXAMPLES +To access a proxy server on +.Pa proxy.example.com +port 8080, set the +.Ev HTTP_PROXY +environment variable in a manner similar to this: +.Pp +.Dl HTTP_PROXY=http://proxy.example.com:8080 +.Pp +If the proxy server requires authentication, there are +two options available for passing the authentication data. +The first method is by using the proxy URL: +.Pp +.Dl HTTP_PROXY=http://:@proxy.example.com:8080 +.Pp +The second method is by using the +.Ev HTTP_PROXY_AUTH +environment variable: +.Bd -literal -offset indent +HTTP_PROXY=http://proxy.example.com:8080 +HTTP_PROXY_AUTH=basic:*:: +.Ed +.Pp +To disable the use of a proxy for an HTTP server running on the local +host, define +.Ev NO_PROXY +as follows: +.Bd -literal -offset indent +NO_PROXY=localhost,127.0.0.1 +.Ed +.Sh SEE ALSO +.Xr fetch 1 , +.Xr ftpio 3 , +.Xr ip 4 +.Rs +.%A J. Postel +.%A J. K. Reynolds +.%D October 1985 +.%B File Transfer Protocol +.%O RFC959 +.Re +.Rs +.%A P. Deutsch +.%A A. Emtage +.%A A. Marine. +.%D May 1994 +.%T How to Use Anonymous FTP +.%O RFC1635 +.Re +.Rs +.%A T. Berners-Lee +.%A L. Masinter +.%A M. McCahill +.%D December 1994 +.%T Uniform Resource Locators (URL) +.%O RFC1738 +.Re +.Rs +.%A R. Fielding +.%A J. Gettys +.%A J. Mogul +.%A H. Frystyk +.%A L. Masinter +.%A P. Leach +.%A T. Berners-Lee +.%D January 1999 +.%B Hypertext Transfer Protocol -- HTTP/1.1 +.%O RFC2616 +.Re +.Rs +.%A J. Franks +.%A P. Hallam-Baker +.%A J. Hostetler +.%A S. Lawrence +.%A P. Leach +.%A A. Luotonen +.%A L. Stewart +.%D June 1999 +.%B HTTP Authentication: Basic and Digest Access Authentication +.%O RFC2617 +.Re +.Sh HISTORY +The +.Nm fetch +library first appeared in +.Fx 3.0 . +.Sh AUTHORS +.An -nosplit +The +.Nm fetch +library was mostly written by +.An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org +with numerous suggestions and contributions from +.An Jordan K. Hubbard Aq jkh@FreeBSD.org , +.An Eugene Skepner Aq eu@qub.com , +.An Hajimu Umemoto Aq ume@FreeBSD.org , +.An Henry Whincup Aq henry@techiebod.com , +.An Jukka A. Ukkonen Aq jau@iki.fi , +.An Jean-Fran\(,cois Dockes Aq jf@dockes.org +and others. +It replaces the older +.Nm ftpio +library written by +.An Poul-Henning Kamp Aq phk@FreeBSD.org +and +.An Jordan K. Hubbard Aq jkh@FreeBSD.org . +.Pp +This manual page was written by +.An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org . +.Sh BUGS +Some parts of the library are not yet implemented. +The most notable +examples of this are +.Fn fetchPutHTTP , +.Fn fetchListHTTP , +.Fn fetchListFTP +and FTP proxy support. +.Pp +There is no way to select a proxy at run-time other than setting the +.Ev HTTP_PROXY +or +.Ev FTP_PROXY +environment variables as appropriate. +.Pp +.Nm libfetch +does not understand or obey 305 (Use Proxy) replies. +.Pp +Error numbers are unique only within a certain context; the error +codes used for FTP and HTTP overlap, as do those used for resolver and +system errors. +For instance, error code 202 means "Command not +implemented, superfluous at this site" in an FTP context and +"Accepted" in an HTTP context. +.Pp +.Fn fetchStatFTP +does not check that the result of an MDTM command is a valid date. +.Pp +The man page is incomplete, poorly written and produces badly +formatted text. +.Pp +The error reporting mechanism is unsatisfactory. +.Pp +Some parts of the code are not fully reentrant. diff --git a/libfetch/fetch.c b/libfetch/fetch.c new file mode 100644 index 0000000..a081520 --- /dev/null +++ b/libfetch/fetch.c @@ -0,0 +1,469 @@ +/*- + * Copyright (c) 1998-2004 Dag-Erling Smørgrav + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. The name of the author may not be used to endorse or promote products + * derived from this software without specific prior written permission + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +#include +#include + +#include +#include +#include +#include + +#include "fetch.h" +#include "common.h" + +auth_t fetchAuthMethod; +int fetchLastErrCode; +char fetchLastErrString[MAXERRSTRING]; +int fetchTimeout; +int fetchRestartCalls = 1; +int fetchDebug; + + +/*** Local data **************************************************************/ + +/* + * Error messages for parser errors + */ +#define URL_MALFORMED 1 +#define URL_BAD_SCHEME 2 +#define URL_BAD_PORT 3 +static struct fetcherr url_errlist[] = { + { URL_MALFORMED, FETCH_URL, "Malformed URL" }, + { URL_BAD_SCHEME, FETCH_URL, "Invalid URL scheme" }, + { URL_BAD_PORT, FETCH_URL, "Invalid server port" }, + { -1, FETCH_UNKNOWN, "Unknown parser error" } +}; + + +/*** Public API **************************************************************/ + +/* + * Select the appropriate protocol for the URL scheme, and return a + * read-only stream connected to the document referenced by the URL. + * Also fill out the struct url_stat. + */ +FILE * +fetchXGet(struct url *URL, struct url_stat *us, const char *flags) +{ + + if (us != NULL) { + us->size = -1; + us->atime = us->mtime = 0; + } + if (strcasecmp(URL->scheme, SCHEME_FILE) == 0) + return (fetchXGetFile(URL, us, flags)); + else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) + return (fetchXGetFTP(URL, us, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTP) == 0) + return (fetchXGetHTTP(URL, us, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0) + return (fetchXGetHTTP(URL, us, flags)); + url_seterr(URL_BAD_SCHEME); + return (NULL); +} + +/* + * Select the appropriate protocol for the URL scheme, and return a + * read-only stream connected to the document referenced by the URL. + */ +FILE * +fetchGet(struct url *URL, const char *flags) +{ + return (fetchXGet(URL, NULL, flags)); +} + +/* + * Select the appropriate protocol for the URL scheme, and return a + * write-only stream connected to the document referenced by the URL. + */ +FILE * +fetchPut(struct url *URL, const char *flags) +{ + + if (strcasecmp(URL->scheme, SCHEME_FILE) == 0) + return (fetchPutFile(URL, flags)); + else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) + return (fetchPutFTP(URL, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTP) == 0) + return (fetchPutHTTP(URL, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0) + return (fetchPutHTTP(URL, flags)); + url_seterr(URL_BAD_SCHEME); + return (NULL); +} + +/* + * Select the appropriate protocol for the URL scheme, and return the + * size of the document referenced by the URL if it exists. + */ +int +fetchStat(struct url *URL, struct url_stat *us, const char *flags) +{ + + if (us != NULL) { + us->size = -1; + us->atime = us->mtime = 0; + } + if (strcasecmp(URL->scheme, SCHEME_FILE) == 0) + return (fetchStatFile(URL, us, flags)); + else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) + return (fetchStatFTP(URL, us, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTP) == 0) + return (fetchStatHTTP(URL, us, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0) + return (fetchStatHTTP(URL, us, flags)); + url_seterr(URL_BAD_SCHEME); + return (-1); +} + +/* + * Select the appropriate protocol for the URL scheme, and return a + * list of files in the directory pointed to by the URL. + */ +struct url_ent * +fetchList(struct url *URL, const char *flags) +{ + + if (strcasecmp(URL->scheme, SCHEME_FILE) == 0) + return (fetchListFile(URL, flags)); + else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) + return (fetchListFTP(URL, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTP) == 0) + return (fetchListHTTP(URL, flags)); + else if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0) + return (fetchListHTTP(URL, flags)); + url_seterr(URL_BAD_SCHEME); + return (NULL); +} + +/* + * Attempt to parse the given URL; if successful, call fetchXGet(). + */ +FILE * +fetchXGetURL(const char *URL, struct url_stat *us, const char *flags) +{ + struct url *u; + FILE *f; + + if ((u = fetchParseURL(URL)) == NULL) + return (NULL); + + f = fetchXGet(u, us, flags); + + fetchFreeURL(u); + return (f); +} + +/* + * Attempt to parse the given URL; if successful, call fetchGet(). + */ +FILE * +fetchGetURL(const char *URL, const char *flags) +{ + return (fetchXGetURL(URL, NULL, flags)); +} + +/* + * Attempt to parse the given URL; if successful, call fetchPut(). + */ +FILE * +fetchPutURL(const char *URL, const char *flags) +{ + struct url *u; + FILE *f; + + if ((u = fetchParseURL(URL)) == NULL) + return (NULL); + + f = fetchPut(u, flags); + + fetchFreeURL(u); + return (f); +} + +/* + * Attempt to parse the given URL; if successful, call fetchStat(). + */ +int +fetchStatURL(const char *URL, struct url_stat *us, const char *flags) +{ + struct url *u; + int s; + + if ((u = fetchParseURL(URL)) == NULL) + return (-1); + + s = fetchStat(u, us, flags); + + fetchFreeURL(u); + return (s); +} + +/* + * Attempt to parse the given URL; if successful, call fetchList(). + */ +struct url_ent * +fetchListURL(const char *URL, const char *flags) +{ + struct url *u; + struct url_ent *ue; + + if ((u = fetchParseURL(URL)) == NULL) + return (NULL); + + ue = fetchList(u, flags); + + fetchFreeURL(u); + return (ue); +} + +/* + * Make a URL + */ +struct url * +fetchMakeURL(const char *scheme, const char *host, int port, const char *doc, + const char *user, const char *pwd) +{ + struct url *u; + + if (!scheme || (!host && !doc)) { + url_seterr(URL_MALFORMED); + return (NULL); + } + + if (port < 0 || port > 65535) { + url_seterr(URL_BAD_PORT); + return (NULL); + } + + /* allocate struct url */ + if ((u = calloc(1, sizeof(*u))) == NULL) { + fetch_syserr(); + return (NULL); + } + + if ((u->doc = strdup(doc ? doc : "/")) == NULL) { + fetch_syserr(); + free(u); + return (NULL); + } + +#define seturl(x) snprintf(u->x, sizeof(u->x), "%s", x) + seturl(scheme); + seturl(host); + seturl(user); + seturl(pwd); +#undef seturl + u->port = port; + + return (u); +} + +/* + * Return value of the given hex digit. + */ +static int +fetch_hexval(char ch) +{ + + if (ch >= '0' && ch <= '9') + return (ch - '0'); + else if (ch >= 'a' && ch <= 'f') + return (ch - 'a' + 10); + else if (ch >= 'A' && ch <= 'F') + return (ch - 'A' + 10); + return (-1); +} + +/* + * Decode percent-encoded URL component from src into dst, stopping at end + * of string, or at @ or : separators. Returns a pointer to the unhandled + * part of the input string (null terminator, @, or :). No terminator is + * written to dst (it is the caller's responsibility). + */ +static const char * +fetch_pctdecode(char *dst, const char *src, size_t dlen) +{ + int d1, d2; + char c; + const char *s; + + for (s = src; *s != '\0' && *s != '@' && *s != ':'; s++) { + if (s[0] == '%' && (d1 = fetch_hexval(s[1])) >= 0 && + (d2 = fetch_hexval(s[2])) >= 0 && (d1 > 0 || d2 > 0)) { + c = d1 << 4 | d2; + s += 2; + } else { + c = *s; + } + if (dlen-- > 0) + *dst++ = c; + } + return (s); +} + +/* + * Split an URL into components. URL syntax is: + * [method:/][/[user[:pwd]@]host[:port]/][document] + * This almost, but not quite, RFC1738 URL syntax. + */ +struct url * +fetchParseURL(const char *URL) +{ + char *doc; + const char *p, *q; + struct url *u; + int i; + + /* allocate struct url */ + if ((u = calloc(1, sizeof(*u))) == NULL) { + fetch_syserr(); + return (NULL); + } + + /* scheme name */ + if ((p = strstr(URL, ":/"))) { + snprintf(u->scheme, URL_SCHEMELEN+1, + "%.*s", (int)(p - URL), URL); + URL = ++p; + /* + * Only one slash: no host, leave slash as part of document + * Two slashes: host follows, strip slashes + */ + if (URL[1] == '/') + URL = (p += 2); + } else { + p = URL; + } + if (!*URL || *URL == '/' || *URL == '.' || + (u->scheme[0] == '\0' && + strchr(URL, '/') == NULL && strchr(URL, ':') == NULL)) + goto nohost; + + p = strpbrk(URL, "/@"); + if (p && *p == '@') { + /* username */ + q = fetch_pctdecode(u->user, URL, URL_USERLEN); + + /* password */ + if (*q == ':') + q = fetch_pctdecode(u->pwd, ++q, URL_PWDLEN); + + p++; + } else { + p = URL; + } + + /* hostname */ +#ifdef INET6 + if (*p == '[' && (q = strchr(p + 1, ']')) != NULL && + (*++q == '\0' || *q == '/' || *q == ':')) { + if ((i = q - p - 2) > MAXHOSTNAMELEN) + i = MAXHOSTNAMELEN; + strncpy(u->host, ++p, i); + p = q; + } else +#endif + for (i = 0; *p && (*p != '/') && (*p != ':'); p++) + if (i < MAXHOSTNAMELEN) + u->host[i++] = *p; + + /* port */ + if (*p == ':') { + for (q = ++p; *q && (*q != '/'); q++) + if (isdigit((unsigned char)*q)) + u->port = u->port * 10 + (*q - '0'); + else { + /* invalid port */ + url_seterr(URL_BAD_PORT); + goto ouch; + } + p = q; + } + +nohost: + /* document */ + if (!*p) + p = "/"; + + if (strcasecmp(u->scheme, SCHEME_HTTP) == 0 || + strcasecmp(u->scheme, SCHEME_HTTPS) == 0) { + const char hexnums[] = "0123456789abcdef"; + + /* percent-escape whitespace. */ + if ((doc = malloc(strlen(p) * 3 + 1)) == NULL) { + fetch_syserr(); + goto ouch; + } + u->doc = doc; + while (*p != '\0') { + if (!isspace((unsigned char)*p)) { + *doc++ = *p++; + } else { + *doc++ = '%'; + *doc++ = hexnums[((unsigned int)*p) >> 4]; + *doc++ = hexnums[((unsigned int)*p) & 0xf]; + p++; + } + } + *doc = '\0'; + } else if ((u->doc = strdup(p)) == NULL) { + fetch_syserr(); + goto ouch; + } + + DEBUG(fprintf(stderr, + "scheme: [%s]\n" + "user: [%s]\n" + "password: [%s]\n" + "host: [%s]\n" + "port: [%d]\n" + "document: [%s]\n", + u->scheme, u->user, u->pwd, + u->host, u->port, u->doc)); + + return (u); + +ouch: + free(u); + return (NULL); +} + +/* + * Free a URL + */ +void +fetchFreeURL(struct url *u) +{ + free(u->doc); + free(u); +} diff --git a/libfetch/fetch.h b/libfetch/fetch.h new file mode 100644 index 0000000..be49482 --- /dev/null +++ b/libfetch/fetch.h @@ -0,0 +1,151 @@ +/*- + * Copyright (c) 1998-2004 Dag-Erling Smørgrav + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. The name of the author may not be used to endorse or promote products + * derived from this software without specific prior written permission + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * $FreeBSD$ + */ + +#ifndef _FETCH_H_INCLUDED +#define _FETCH_H_INCLUDED + +#define _LIBFETCH_VER "libfetch/2.0" + +#define URL_SCHEMELEN 16 +#define URL_USERLEN 256 +#define URL_PWDLEN 256 + +struct url { + char scheme[URL_SCHEMELEN+1]; + char user[URL_USERLEN+1]; + char pwd[URL_PWDLEN+1]; + char host[MAXHOSTNAMELEN+1]; + int port; + char *doc; + off_t offset; + size_t length; + time_t ims_time; +}; + +struct url_stat { + off_t size; + time_t atime; + time_t mtime; +}; + +struct url_ent { + char name[PATH_MAX]; + struct url_stat stat; +}; + +/* Recognized schemes */ +#define SCHEME_FTP "ftp" +#define SCHEME_HTTP "http" +#define SCHEME_HTTPS "https" +#define SCHEME_FILE "file" + +/* Error codes */ +#define FETCH_ABORT 1 +#define FETCH_AUTH 2 +#define FETCH_DOWN 3 +#define FETCH_EXISTS 4 +#define FETCH_FULL 5 +#define FETCH_INFO 6 +#define FETCH_MEMORY 7 +#define FETCH_MOVED 8 +#define FETCH_NETWORK 9 +#define FETCH_OK 10 +#define FETCH_PROTO 11 +#define FETCH_RESOLV 12 +#define FETCH_SERVER 13 +#define FETCH_TEMP 14 +#define FETCH_TIMEOUT 15 +#define FETCH_UNAVAIL 16 +#define FETCH_UNKNOWN 17 +#define FETCH_URL 18 +#define FETCH_VERBOSE 19 + +__BEGIN_DECLS + +/* FILE-specific functions */ +FILE *fetchXGetFile(struct url *, struct url_stat *, const char *); +FILE *fetchGetFile(struct url *, const char *); +FILE *fetchPutFile(struct url *, const char *); +int fetchStatFile(struct url *, struct url_stat *, const char *); +struct url_ent *fetchListFile(struct url *, const char *); + +/* HTTP-specific functions */ +FILE *fetchXGetHTTP(struct url *, struct url_stat *, const char *); +FILE *fetchGetHTTP(struct url *, const char *); +FILE *fetchPutHTTP(struct url *, const char *); +int fetchStatHTTP(struct url *, struct url_stat *, const char *); +struct url_ent *fetchListHTTP(struct url *, const char *); + +/* FTP-specific functions */ +FILE *fetchXGetFTP(struct url *, struct url_stat *, const char *); +FILE *fetchGetFTP(struct url *, const char *); +FILE *fetchPutFTP(struct url *, const char *); +int fetchStatFTP(struct url *, struct url_stat *, const char *); +struct url_ent *fetchListFTP(struct url *, const char *); + +/* Generic functions */ +FILE *fetchXGetURL(const char *, struct url_stat *, const char *); +FILE *fetchGetURL(const char *, const char *); +FILE *fetchPutURL(const char *, const char *); +int fetchStatURL(const char *, struct url_stat *, const char *); +struct url_ent *fetchListURL(const char *, const char *); +FILE *fetchXGet(struct url *, struct url_stat *, const char *); +FILE *fetchGet(struct url *, const char *); +FILE *fetchPut(struct url *, const char *); +int fetchStat(struct url *, struct url_stat *, const char *); +struct url_ent *fetchList(struct url *, const char *); + +/* URL parsing */ +struct url *fetchMakeURL(const char *, const char *, int, + const char *, const char *, const char *); +struct url *fetchParseURL(const char *); +void fetchFreeURL(struct url *); + +__END_DECLS + +/* Authentication */ +typedef int (*auth_t)(struct url *); +extern auth_t fetchAuthMethod; + +/* Last error code */ +extern int fetchLastErrCode; +#define MAXERRSTRING 256 +extern char fetchLastErrString[MAXERRSTRING]; + +/* I/O timeout */ +extern int fetchTimeout; + +/* Restart interrupted syscalls */ +extern int fetchRestartCalls; + +/* Extra verbosity */ +extern int fetchDebug; + +#endif diff --git a/libfetch/file.c b/libfetch/file.c new file mode 100644 index 0000000..8569ff3 --- /dev/null +++ b/libfetch/file.c @@ -0,0 +1,149 @@ +/*- + * Copyright (c) 1998-2011 Dag-Erling Smørgrav + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. The name of the author may not be used to endorse or promote products + * derived from this software without specific prior written permission + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +#include +#include + +#include +#include +#include +#include + +#include "fetch.h" +#include "common.h" + +FILE * +fetchXGetFile(struct url *u, struct url_stat *us, const char *flags) +{ + FILE *f; + + if (us && fetchStatFile(u, us, flags) == -1) + return (NULL); + + f = fopen(u->doc, "r"); + + if (f == NULL) + fetch_syserr(); + + if (u->offset && fseeko(f, u->offset, SEEK_SET) == -1) { + fclose(f); + fetch_syserr(); + } + + fcntl(fileno(f), F_SETFD, FD_CLOEXEC); + return (f); +} + +FILE * +fetchGetFile(struct url *u, const char *flags) +{ + return (fetchXGetFile(u, NULL, flags)); +} + +FILE * +fetchPutFile(struct url *u, const char *flags) +{ + FILE *f; + + if (CHECK_FLAG('a')) + f = fopen(u->doc, "a"); + else + f = fopen(u->doc, "w+"); + + if (f == NULL) + fetch_syserr(); + + if (u->offset && fseeko(f, u->offset, SEEK_SET) == -1) { + fclose(f); + fetch_syserr(); + } + + fcntl(fileno(f), F_SETFD, FD_CLOEXEC); + return (f); +} + +static int +fetch_stat_file(const char *fn, struct url_stat *us) +{ + struct stat sb; + + us->size = -1; + us->atime = us->mtime = 0; + if (stat(fn, &sb) == -1) { + fetch_syserr(); + return (-1); + } + us->size = sb.st_size; + us->atime = sb.st_atime; + us->mtime = sb.st_mtime; + return (0); +} + +int +fetchStatFile(struct url *u, struct url_stat *us, const char *flags __unused) +{ + return (fetch_stat_file(u->doc, us)); +} + +struct url_ent * +fetchListFile(struct url *u, const char *flags __unused) +{ + struct dirent *de; + struct url_stat us; + struct url_ent *ue; + int size, len; + char fn[PATH_MAX], *p; + DIR *dir; + int l; + + if ((dir = opendir(u->doc)) == NULL) { + fetch_syserr(); + return (NULL); + } + + ue = NULL; + strncpy(fn, u->doc, sizeof(fn) - 2); + fn[sizeof(fn) - 2] = 0; + strcat(fn, "/"); + p = strchr(fn, 0); + l = sizeof(fn) - strlen(fn) - 1; + + while ((de = readdir(dir)) != NULL) { + strncpy(p, de->d_name, l - 1); + p[l - 1] = 0; + if (fetch_stat_file(fn, &us) == -1) + /* should I return a partial result, or abort? */ + break; + fetch_add_entry(&ue, &size, &len, de->d_name, &us); + } + + return (ue); +} diff --git a/libfetch/ftp.c b/libfetch/ftp.c new file mode 100644 index 0000000..4e650b9 --- /dev/null +++ b/libfetch/ftp.c @@ -0,0 +1,1206 @@ +/*- + * Copyright (c) 1998-2011 Dag-Erling Smørgrav + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. The name of the author may not be used to endorse or promote products + * derived from this software without specific prior written permission + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +/* + * Portions of this code were taken from or based on ftpio.c: + * + * ---------------------------------------------------------------------------- + * "THE BEER-WARE LICENSE" (Revision 42): + * wrote this file. As long as you retain this notice you + * can do whatever you want with this stuff. If we meet some day, and you think + * this stuff is worth it, you can buy me a beer in return. Poul-Henning Kamp + * ---------------------------------------------------------------------------- + * + * Major Changelog: + * + * Dag-Erling Smørgrav + * 9 Jun 1998 + * + * Incorporated into libfetch + * + * Jordan K. Hubbard + * 17 Jan 1996 + * + * Turned inside out. Now returns xfers as new file ids, not as a special + * `state' of FTP_t + * + * $ftpioId: ftpio.c,v 1.30 1998/04/11 07:28:53 phk Exp $ + * + */ + +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "fetch.h" +#include "common.h" +#include "ftperr.h" + +#define FTP_ANONYMOUS_USER "anonymous" + +#define FTP_CONNECTION_ALREADY_OPEN 125 +#define FTP_OPEN_DATA_CONNECTION 150 +#define FTP_OK 200 +#define FTP_FILE_STATUS 213 +#define FTP_SERVICE_READY 220 +#define FTP_TRANSFER_COMPLETE 226 +#define FTP_PASSIVE_MODE 227 +#define FTP_LPASSIVE_MODE 228 +#define FTP_EPASSIVE_MODE 229 +#define FTP_LOGGED_IN 230 +#define FTP_FILE_ACTION_OK 250 +#define FTP_DIRECTORY_CREATED 257 /* multiple meanings */ +#define FTP_FILE_CREATED 257 /* multiple meanings */ +#define FTP_WORKING_DIRECTORY 257 /* multiple meanings */ +#define FTP_NEED_PASSWORD 331 +#define FTP_NEED_ACCOUNT 332 +#define FTP_FILE_OK 350 +#define FTP_SYNTAX_ERROR 500 +#define FTP_PROTOCOL_ERROR 999 + +static struct url cached_host; +static conn_t *cached_connection; + +#define isftpreply(foo) \ + (isdigit((unsigned char)foo[0]) && \ + isdigit((unsigned char)foo[1]) && \ + isdigit((unsigned char)foo[2]) && \ + (foo[3] == ' ' || foo[3] == '\0')) +#define isftpinfo(foo) \ + (isdigit((unsigned char)foo[0]) && \ + isdigit((unsigned char)foo[1]) && \ + isdigit((unsigned char)foo[2]) && \ + foo[3] == '-') + +/* + * Translate IPv4 mapped IPv6 address to IPv4 address + */ +static void +unmappedaddr(struct sockaddr_in6 *sin6) +{ + struct sockaddr_in *sin4; + u_int32_t addr; + int port; + + if (sin6->sin6_family != AF_INET6 || + !IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr)) + return; + sin4 = (struct sockaddr_in *)sin6; + addr = *(u_int32_t *)(uintptr_t)&sin6->sin6_addr.s6_addr[12]; + port = sin6->sin6_port; + memset(sin4, 0, sizeof(struct sockaddr_in)); + sin4->sin_addr.s_addr = addr; + sin4->sin_port = port; + sin4->sin_family = AF_INET; + sin4->sin_len = sizeof(struct sockaddr_in); +} + +/* + * Get server response + */ +static int +ftp_chkerr(conn_t *conn) +{ + if (fetch_getln(conn) == -1) { + fetch_syserr(); + return (-1); + } + if (isftpinfo(conn->buf)) { + while (conn->buflen && !isftpreply(conn->buf)) { + if (fetch_getln(conn) == -1) { + fetch_syserr(); + return (-1); + } + } + } + + while (conn->buflen && + isspace((unsigned char)conn->buf[conn->buflen - 1])) + conn->buflen--; + conn->buf[conn->buflen] = '\0'; + + if (!isftpreply(conn->buf)) { + ftp_seterr(FTP_PROTOCOL_ERROR); + return (-1); + } + + conn->err = (conn->buf[0] - '0') * 100 + + (conn->buf[1] - '0') * 10 + + (conn->buf[2] - '0'); + + return (conn->err); +} + +/* + * Send a command and check reply + */ +static int +ftp_cmd(conn_t *conn, const char *fmt, ...) +{ + va_list ap; + size_t len; + char *msg; + int r; + + va_start(ap, fmt); + len = vasprintf(&msg, fmt, ap); + va_end(ap); + + if (msg == NULL) { + errno = ENOMEM; + fetch_syserr(); + return (-1); + } + + r = fetch_putln(conn, msg, len); + free(msg); + + if (r == -1) { + fetch_syserr(); + return (-1); + } + + return (ftp_chkerr(conn)); +} + +/* + * Return a pointer to the filename part of a path + */ +static const char * +ftp_filename(const char *file, int *len, int *type) +{ + const char *s; + + if ((s = strrchr(file, '/')) == NULL) + s = file; + else + s = s + 1; + *len = strlen(s); + if (*len > 7 && strncmp(s + *len - 7, ";type=", 6) == 0) { + *type = s[*len - 1]; + *len -= 7; + } else { + *type = '\0'; + } + return (s); +} + +/* + * Get current working directory from the reply to a CWD, PWD or CDUP + * command. + */ +static int +ftp_pwd(conn_t *conn, char *pwd, size_t pwdlen) +{ + char *src, *dst, *end; + int q; + + if (conn->err != FTP_WORKING_DIRECTORY && + conn->err != FTP_FILE_ACTION_OK) + return (FTP_PROTOCOL_ERROR); + end = conn->buf + conn->buflen; + src = conn->buf + 4; + if (src >= end || *src++ != '"') + return (FTP_PROTOCOL_ERROR); + for (q = 0, dst = pwd; src < end && pwdlen--; ++src) { + if (!q && *src == '"') + q = 1; + else if (q && *src != '"') + break; + else if (q) + *dst++ = '"', q = 0; + else + *dst++ = *src; + } + if (!pwdlen) + return (FTP_PROTOCOL_ERROR); + *dst = '\0'; +#if 0 + DEBUG(fprintf(stderr, "pwd: [%s]\n", pwd)); +#endif + return (FTP_OK); +} + +/* + * Change working directory to the directory that contains the specified + * file. + */ +static int +ftp_cwd(conn_t *conn, const char *file) +{ + const char *beg, *end; + char pwd[PATH_MAX]; + int e, i, len; + + /* If no slashes in name, no need to change dirs. */ + if ((end = strrchr(file, '/')) == NULL) + return (0); + if ((e = ftp_cmd(conn, "PWD")) != FTP_WORKING_DIRECTORY || + (e = ftp_pwd(conn, pwd, sizeof(pwd))) != FTP_OK) { + ftp_seterr(e); + return (-1); + } + for (;;) { + len = strlen(pwd); + + /* Look for a common prefix between PWD and dir to fetch. */ + for (i = 0; i <= len && i <= end - file; ++i) + if (pwd[i] != file[i]) + break; +#if 0 + DEBUG(fprintf(stderr, "have: [%.*s|%s]\n", i, pwd, pwd + i)); + DEBUG(fprintf(stderr, "want: [%.*s|%s]\n", i, file, file + i)); +#endif + /* Keep going up a dir until we have a matching prefix. */ + if (pwd[i] == '\0' && (file[i - 1] == '/' || file[i] == '/')) + break; + if ((e = ftp_cmd(conn, "CDUP")) != FTP_FILE_ACTION_OK || + (e = ftp_cmd(conn, "PWD")) != FTP_WORKING_DIRECTORY || + (e = ftp_pwd(conn, pwd, sizeof(pwd))) != FTP_OK) { + ftp_seterr(e); + return (-1); + } + } + +#ifdef FTP_COMBINE_CWDS + /* Skip leading slashes, even "////". */ + for (beg = file + i; beg < end && *beg == '/'; ++beg, ++i) + /* nothing */ ; + + /* If there is no trailing dir, we're already there. */ + if (beg >= end) + return (0); + + /* Change to the directory all in one chunk (e.g., foo/bar/baz). */ + e = ftp_cmd(conn, "CWD %.*s", (int)(end - beg), beg); + if (e == FTP_FILE_ACTION_OK) + return (0); +#endif /* FTP_COMBINE_CWDS */ + + /* That didn't work so go back to legacy behavior (multiple CWDs). */ + for (beg = file + i; beg < end; beg = file + i + 1) { + while (*beg == '/') + ++beg, ++i; + for (++i; file + i < end && file[i] != '/'; ++i) + /* nothing */ ; + e = ftp_cmd(conn, "CWD %.*s", file + i - beg, beg); + if (e != FTP_FILE_ACTION_OK) { + ftp_seterr(e); + return (-1); + } + } + return (0); +} + +/* + * Set transfer mode and data type + */ +static int +ftp_mode_type(conn_t *conn, int mode, int type) +{ + int e; + + switch (mode) { + case 0: + case 's': + mode = 'S'; + case 'S': + break; + default: + return (FTP_PROTOCOL_ERROR); + } + if ((e = ftp_cmd(conn, "MODE %c", mode)) != FTP_OK) { + if (mode == 'S') { + /* + * Stream mode is supposed to be the default - so + * much so that some servers not only do not + * support any other mode, but do not support the + * MODE command at all. + * + * If "MODE S" fails, it is unlikely that we + * previously succeeded in setting a different + * mode. Therefore, we simply hope that the + * server is already in the correct mode, and + * silently ignore the failure. + */ + } else { + return (e); + } + } + + switch (type) { + case 0: + case 'i': + type = 'I'; + case 'I': + break; + case 'a': + type = 'A'; + case 'A': + break; + case 'd': + type = 'D'; + case 'D': + /* can't handle yet */ + default: + return (FTP_PROTOCOL_ERROR); + } + if ((e = ftp_cmd(conn, "TYPE %c", type)) != FTP_OK) + return (e); + + return (FTP_OK); +} + +/* + * Request and parse file stats + */ +static int +ftp_stat(conn_t *conn, const char *file, struct url_stat *us) +{ + char *ln; + const char *filename; + int filenamelen, type; + struct tm tm; + time_t t; + int e; + + us->size = -1; + us->atime = us->mtime = 0; + + filename = ftp_filename(file, &filenamelen, &type); + + if ((e = ftp_mode_type(conn, 0, type)) != FTP_OK) { + ftp_seterr(e); + return (-1); + } + + e = ftp_cmd(conn, "SIZE %.*s", filenamelen, filename); + if (e != FTP_FILE_STATUS) { + ftp_seterr(e); + return (-1); + } + for (ln = conn->buf + 4; *ln && isspace((unsigned char)*ln); ln++) + /* nothing */ ; + for (us->size = 0; *ln && isdigit((unsigned char)*ln); ln++) + us->size = us->size * 10 + *ln - '0'; + if (*ln && !isspace((unsigned char)*ln)) { + ftp_seterr(FTP_PROTOCOL_ERROR); + us->size = -1; + return (-1); + } + if (us->size == 0) + us->size = -1; + DEBUG(fprintf(stderr, "size: [%lld]\n", (long long)us->size)); + + e = ftp_cmd(conn, "MDTM %.*s", filenamelen, filename); + if (e != FTP_FILE_STATUS) { + ftp_seterr(e); + return (-1); + } + for (ln = conn->buf + 4; *ln && isspace((unsigned char)*ln); ln++) + /* nothing */ ; + switch (strspn(ln, "0123456789")) { + case 14: + break; + case 15: + ln++; + ln[0] = '2'; + ln[1] = '0'; + break; + default: + ftp_seterr(FTP_PROTOCOL_ERROR); + return (-1); + } + if (sscanf(ln, "%04d%02d%02d%02d%02d%02d", + &tm.tm_year, &tm.tm_mon, &tm.tm_mday, + &tm.tm_hour, &tm.tm_min, &tm.tm_sec) != 6) { + ftp_seterr(FTP_PROTOCOL_ERROR); + return (-1); + } + tm.tm_mon--; + tm.tm_year -= 1900; + tm.tm_isdst = -1; + t = timegm(&tm); + if (t == (time_t)-1) + t = time(NULL); + us->mtime = t; + us->atime = t; + DEBUG(fprintf(stderr, + "last modified: [%04d-%02d-%02d %02d:%02d:%02d]\n", + tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday, + tm.tm_hour, tm.tm_min, tm.tm_sec)); + return (0); +} + +/* + * I/O functions for FTP + */ +struct ftpio { + conn_t *cconn; /* Control connection */ + conn_t *dconn; /* Data connection */ + int dir; /* Direction */ + int eof; /* EOF reached */ + int err; /* Error code */ +}; + +static int ftp_readfn(void *, char *, int); +static int ftp_writefn(void *, const char *, int); +static fpos_t ftp_seekfn(void *, fpos_t, int); +static int ftp_closefn(void *); + +static int +ftp_readfn(void *v, char *buf, int len) +{ + struct ftpio *io; + int r; + + io = (struct ftpio *)v; + if (io == NULL) { + errno = EBADF; + return (-1); + } + if (io->cconn == NULL || io->dconn == NULL || io->dir == O_WRONLY) { + errno = EBADF; + return (-1); + } + if (io->err) { + errno = io->err; + return (-1); + } + if (io->eof) + return (0); + r = fetch_read(io->dconn, buf, len); + if (r > 0) + return (r); + if (r == 0) { + io->eof = 1; + return (0); + } + if (errno != EINTR) + io->err = errno; + return (-1); +} + +static int +ftp_writefn(void *v, const char *buf, int len) +{ + struct ftpio *io; + int w; + + io = (struct ftpio *)v; + if (io == NULL) { + errno = EBADF; + return (-1); + } + if (io->cconn == NULL || io->dconn == NULL || io->dir == O_RDONLY) { + errno = EBADF; + return (-1); + } + if (io->err) { + errno = io->err; + return (-1); + } + w = fetch_write(io->dconn, buf, len); + if (w >= 0) + return (w); + if (errno != EINTR) + io->err = errno; + return (-1); +} + +static fpos_t +ftp_seekfn(void *v, fpos_t pos __unused, int whence __unused) +{ + struct ftpio *io; + + io = (struct ftpio *)v; + if (io == NULL) { + errno = EBADF; + return (-1); + } + errno = ESPIPE; + return (-1); +} + +static int +ftp_closefn(void *v) +{ + struct ftpio *io; + int r; + + io = (struct ftpio *)v; + if (io == NULL) { + errno = EBADF; + return (-1); + } + if (io->dir == -1) + return (0); + if (io->cconn == NULL || io->dconn == NULL) { + errno = EBADF; + return (-1); + } + fetch_close(io->dconn); + io->dir = -1; + io->dconn = NULL; + DEBUG(fprintf(stderr, "Waiting for final status\n")); + r = ftp_chkerr(io->cconn); + if (io->cconn == cached_connection && io->cconn->ref == 1) + cached_connection = NULL; + fetch_close(io->cconn); + free(io); + return (r == FTP_TRANSFER_COMPLETE) ? 0 : -1; +} + +static FILE * +ftp_setup(conn_t *cconn, conn_t *dconn, int mode) +{ + struct ftpio *io; + FILE *f; + + if (cconn == NULL || dconn == NULL) + return (NULL); + if ((io = malloc(sizeof(*io))) == NULL) + return (NULL); + io->cconn = cconn; + io->dconn = dconn; + io->dir = mode; + io->eof = io->err = 0; + f = funopen(io, ftp_readfn, ftp_writefn, ftp_seekfn, ftp_closefn); + if (f == NULL) + free(io); + return (f); +} + +/* + * Transfer file + */ +static FILE * +ftp_transfer(conn_t *conn, const char *oper, const char *file, + int mode, off_t offset, const char *flags) +{ + struct sockaddr_storage sa; + struct sockaddr_in6 *sin6; + struct sockaddr_in *sin4; + const char *bindaddr; + const char *filename; + int filenamelen, type; + int low, pasv, verbose; + int e, sd = -1; + socklen_t l; + char *s; + FILE *df; + + /* check flags */ + low = CHECK_FLAG('l'); + pasv = CHECK_FLAG('p') || !CHECK_FLAG('P'); + verbose = CHECK_FLAG('v'); + + /* passive mode */ + if ((s = getenv("FTP_PASSIVE_MODE")) != NULL) + pasv = (strncasecmp(s, "no", 2) != 0); + + /* isolate filename */ + filename = ftp_filename(file, &filenamelen, &type); + + /* set transfer mode and data type */ + if ((e = ftp_mode_type(conn, 0, type)) != FTP_OK) + goto ouch; + + /* find our own address, bind, and listen */ + l = sizeof(sa); + if (getsockname(conn->sd, (struct sockaddr *)&sa, &l) == -1) + goto sysouch; + if (sa.ss_family == AF_INET6) + unmappedaddr((struct sockaddr_in6 *)&sa); + + /* open data socket */ + if ((sd = socket(sa.ss_family, SOCK_STREAM, IPPROTO_TCP)) == -1) { + fetch_syserr(); + return (NULL); + } + + if (pasv) { + u_char addr[64]; + char *ln, *p; + unsigned int i; + int port; + + /* send PASV command */ + if (verbose) + fetch_info("setting passive mode"); + switch (sa.ss_family) { + case AF_INET: + if ((e = ftp_cmd(conn, "PASV")) != FTP_PASSIVE_MODE) + goto ouch; + break; + case AF_INET6: + if ((e = ftp_cmd(conn, "EPSV")) != FTP_EPASSIVE_MODE) { + if (e == -1) + goto ouch; + if ((e = ftp_cmd(conn, "LPSV")) != + FTP_LPASSIVE_MODE) + goto ouch; + } + break; + default: + e = FTP_PROTOCOL_ERROR; /* XXX: error code should be prepared */ + goto ouch; + } + + /* + * Find address and port number. The reply to the PASV command + * is IMHO the one and only weak point in the FTP protocol. + */ + ln = conn->buf; + switch (e) { + case FTP_PASSIVE_MODE: + case FTP_LPASSIVE_MODE: + for (p = ln + 3; *p && !isdigit((unsigned char)*p); p++) + /* nothing */ ; + if (!*p) { + e = FTP_PROTOCOL_ERROR; + goto ouch; + } + l = (e == FTP_PASSIVE_MODE ? 6 : 21); + for (i = 0; *p && i < l; i++, p++) + addr[i] = strtol(p, &p, 10); + if (i < l) { + e = FTP_PROTOCOL_ERROR; + goto ouch; + } + break; + case FTP_EPASSIVE_MODE: + for (p = ln + 3; *p && *p != '('; p++) + /* nothing */ ; + if (!*p) { + e = FTP_PROTOCOL_ERROR; + goto ouch; + } + ++p; + if (sscanf(p, "%c%c%c%d%c", &addr[0], &addr[1], &addr[2], + &port, &addr[3]) != 5 || + addr[0] != addr[1] || + addr[0] != addr[2] || addr[0] != addr[3]) { + e = FTP_PROTOCOL_ERROR; + goto ouch; + } + break; + } + + /* seek to required offset */ + if (offset) + if (ftp_cmd(conn, "REST %lu", (u_long)offset) != FTP_FILE_OK) + goto sysouch; + + /* construct sockaddr for data socket */ + l = sizeof(sa); + if (getpeername(conn->sd, (struct sockaddr *)&sa, &l) == -1) + goto sysouch; + if (sa.ss_family == AF_INET6) + unmappedaddr((struct sockaddr_in6 *)&sa); + switch (sa.ss_family) { + case AF_INET6: + sin6 = (struct sockaddr_in6 *)&sa; + if (e == FTP_EPASSIVE_MODE) + sin6->sin6_port = htons(port); + else { + memcpy(&sin6->sin6_addr, addr + 2, 16); + memcpy(&sin6->sin6_port, addr + 19, 2); + } + break; + case AF_INET: + sin4 = (struct sockaddr_in *)&sa; + if (e == FTP_EPASSIVE_MODE) + sin4->sin_port = htons(port); + else { + memcpy(&sin4->sin_addr, addr, 4); + memcpy(&sin4->sin_port, addr + 4, 2); + } + break; + default: + e = FTP_PROTOCOL_ERROR; /* XXX: error code should be prepared */ + break; + } + + /* connect to data port */ + if (verbose) + fetch_info("opening data connection"); + bindaddr = getenv("FETCH_BIND_ADDRESS"); + if (bindaddr != NULL && *bindaddr != '\0' && + fetch_bind(sd, sa.ss_family, bindaddr) != 0) + goto sysouch; + if (connect(sd, (struct sockaddr *)&sa, sa.ss_len) == -1) + goto sysouch; + + /* make the server initiate the transfer */ + if (verbose) + fetch_info("initiating transfer"); + e = ftp_cmd(conn, "%s %.*s", oper, filenamelen, filename); + if (e != FTP_CONNECTION_ALREADY_OPEN && e != FTP_OPEN_DATA_CONNECTION) + goto ouch; + + } else { + u_int32_t a; + u_short p; + int arg, d; + char *ap; + char hname[INET6_ADDRSTRLEN]; + + switch (sa.ss_family) { + case AF_INET6: + ((struct sockaddr_in6 *)&sa)->sin6_port = 0; +#ifdef IPV6_PORTRANGE + arg = low ? IPV6_PORTRANGE_DEFAULT : IPV6_PORTRANGE_HIGH; + if (setsockopt(sd, IPPROTO_IPV6, IPV6_PORTRANGE, + (char *)&arg, sizeof(arg)) == -1) + goto sysouch; +#endif + break; + case AF_INET: + ((struct sockaddr_in *)&sa)->sin_port = 0; + arg = low ? IP_PORTRANGE_DEFAULT : IP_PORTRANGE_HIGH; + if (setsockopt(sd, IPPROTO_IP, IP_PORTRANGE, + (char *)&arg, sizeof(arg)) == -1) + goto sysouch; + break; + } + if (verbose) + fetch_info("binding data socket"); + if (bind(sd, (struct sockaddr *)&sa, sa.ss_len) == -1) + goto sysouch; + if (listen(sd, 1) == -1) + goto sysouch; + + /* find what port we're on and tell the server */ + if (getsockname(sd, (struct sockaddr *)&sa, &l) == -1) + goto sysouch; + switch (sa.ss_family) { + case AF_INET: + sin4 = (struct sockaddr_in *)&sa; + a = ntohl(sin4->sin_addr.s_addr); + p = ntohs(sin4->sin_port); + e = ftp_cmd(conn, "PORT %d,%d,%d,%d,%d,%d", + (a >> 24) & 0xff, (a >> 16) & 0xff, + (a >> 8) & 0xff, a & 0xff, + (p >> 8) & 0xff, p & 0xff); + break; + case AF_INET6: +#define UC(b) (((int)b)&0xff) + e = -1; + sin6 = (struct sockaddr_in6 *)&sa; + sin6->sin6_scope_id = 0; + if (getnameinfo((struct sockaddr *)&sa, sa.ss_len, + hname, sizeof(hname), + NULL, 0, NI_NUMERICHOST) == 0) { + e = ftp_cmd(conn, "EPRT |%d|%s|%d|", 2, hname, + htons(sin6->sin6_port)); + if (e == -1) + goto ouch; + } + if (e != FTP_OK) { + ap = (char *)&sin6->sin6_addr; + e = ftp_cmd(conn, + "LPRT %d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d", + 6, 16, + UC(ap[0]), UC(ap[1]), UC(ap[2]), UC(ap[3]), + UC(ap[4]), UC(ap[5]), UC(ap[6]), UC(ap[7]), + UC(ap[8]), UC(ap[9]), UC(ap[10]), UC(ap[11]), + UC(ap[12]), UC(ap[13]), UC(ap[14]), UC(ap[15]), + 2, + (ntohs(sin6->sin6_port) >> 8) & 0xff, + ntohs(sin6->sin6_port) & 0xff); + } + break; + default: + e = FTP_PROTOCOL_ERROR; /* XXX: error code should be prepared */ + goto ouch; + } + if (e != FTP_OK) + goto ouch; + + /* seek to required offset */ + if (offset) + if (ftp_cmd(conn, "REST %ju", (uintmax_t)offset) != FTP_FILE_OK) + goto sysouch; + + /* make the server initiate the transfer */ + if (verbose) + fetch_info("initiating transfer"); + e = ftp_cmd(conn, "%s %.*s", oper, filenamelen, filename); + if (e != FTP_CONNECTION_ALREADY_OPEN && e != FTP_OPEN_DATA_CONNECTION) + goto ouch; + + /* accept the incoming connection and go to town */ + if ((d = accept(sd, NULL, NULL)) == -1) + goto sysouch; + close(sd); + sd = d; + } + + if ((df = ftp_setup(conn, fetch_reopen(sd), mode)) == NULL) + goto sysouch; + return (df); + +sysouch: + fetch_syserr(); + if (sd >= 0) + close(sd); + return (NULL); + +ouch: + if (e != -1) + ftp_seterr(e); + if (sd >= 0) + close(sd); + return (NULL); +} + +/* + * Authenticate + */ +static int +ftp_authenticate(conn_t *conn, struct url *url, struct url *purl) +{ + const char *user, *pwd, *logname; + char pbuf[MAXHOSTNAMELEN + MAXLOGNAME + 1]; + int e, len; + + /* XXX FTP_AUTH, and maybe .netrc */ + + /* send user name and password */ + if (url->user[0] == '\0') + fetch_netrc_auth(url); + user = url->user; + if (*user == '\0') + user = getenv("FTP_LOGIN"); + if (user == NULL || *user == '\0') + user = FTP_ANONYMOUS_USER; + if (purl && url->port == fetch_default_port(url->scheme)) + e = ftp_cmd(conn, "USER %s@%s", user, url->host); + else if (purl) + e = ftp_cmd(conn, "USER %s@%s@%d", user, url->host, url->port); + else + e = ftp_cmd(conn, "USER %s", user); + + /* did the server request a password? */ + if (e == FTP_NEED_PASSWORD) { + pwd = url->pwd; + if (*pwd == '\0') + pwd = getenv("FTP_PASSWORD"); + if (pwd == NULL || *pwd == '\0') { + if ((logname = getlogin()) == 0) + logname = FTP_ANONYMOUS_USER; + if ((len = snprintf(pbuf, MAXLOGNAME + 1, "%s@", logname)) < 0) + len = 0; + else if (len > MAXLOGNAME) + len = MAXLOGNAME; + gethostname(pbuf + len, sizeof(pbuf) - len); + pwd = pbuf; + } + e = ftp_cmd(conn, "PASS %s", pwd); + } + + return (e); +} + +/* + * Log on to FTP server + */ +static conn_t * +ftp_connect(struct url *url, struct url *purl, const char *flags) +{ + conn_t *conn; + int e, direct, verbose; +#ifdef INET6 + int af = AF_UNSPEC; +#else + int af = AF_INET; +#endif + + direct = CHECK_FLAG('d'); + verbose = CHECK_FLAG('v'); + if (CHECK_FLAG('4')) + af = AF_INET; + else if (CHECK_FLAG('6')) + af = AF_INET6; + + if (direct) + purl = NULL; + + /* check for proxy */ + if (purl) { + /* XXX proxy authentication! */ + conn = fetch_connect(purl->host, purl->port, af, verbose); + } else { + /* no proxy, go straight to target */ + conn = fetch_connect(url->host, url->port, af, verbose); + purl = NULL; + } + + /* check connection */ + if (conn == NULL) + /* fetch_connect() has already set an error code */ + return (NULL); + + /* expect welcome message */ + if ((e = ftp_chkerr(conn)) != FTP_SERVICE_READY) + goto fouch; + + /* authenticate */ + if ((e = ftp_authenticate(conn, url, purl)) != FTP_LOGGED_IN) + goto fouch; + + /* TODO: Request extended features supported, if any (RFC 3659). */ + + /* done */ + return (conn); + +fouch: + if (e != -1) + ftp_seterr(e); + fetch_close(conn); + return (NULL); +} + +/* + * Disconnect from server + */ +static void +ftp_disconnect(conn_t *conn) +{ + (void)ftp_cmd(conn, "QUIT"); + if (conn == cached_connection && conn->ref == 1) + cached_connection = NULL; + fetch_close(conn); +} + +/* + * Check if we're already connected + */ +static int +ftp_isconnected(struct url *url) +{ + return (cached_connection + && (strcmp(url->host, cached_host.host) == 0) + && (strcmp(url->user, cached_host.user) == 0) + && (strcmp(url->pwd, cached_host.pwd) == 0) + && (url->port == cached_host.port)); +} + +/* + * Check the cache, reconnect if no luck + */ +static conn_t * +ftp_cached_connect(struct url *url, struct url *purl, const char *flags) +{ + conn_t *conn; + int e; + + /* set default port */ + if (!url->port) + url->port = fetch_default_port(url->scheme); + + /* try to use previously cached connection */ + if (ftp_isconnected(url)) { + e = ftp_cmd(cached_connection, "NOOP"); + if (e == FTP_OK || e == FTP_SYNTAX_ERROR) + return (fetch_ref(cached_connection)); + } + + /* connect to server */ + if ((conn = ftp_connect(url, purl, flags)) == NULL) + return (NULL); + if (cached_connection) + ftp_disconnect(cached_connection); + cached_connection = fetch_ref(conn); + memcpy(&cached_host, url, sizeof(*url)); + return (conn); +} + +/* + * Check the proxy settings + */ +static struct url * +ftp_get_proxy(struct url * url, const char *flags) +{ + struct url *purl; + char *p; + + if (flags != NULL && strchr(flags, 'd') != NULL) + return (NULL); + if (fetch_no_proxy_match(url->host)) + return (NULL); + if (((p = getenv("FTP_PROXY")) || (p = getenv("ftp_proxy")) || + (p = getenv("HTTP_PROXY")) || (p = getenv("http_proxy"))) && + *p && (purl = fetchParseURL(p)) != NULL) { + if (!*purl->scheme) { + if (getenv("FTP_PROXY") || getenv("ftp_proxy")) + strcpy(purl->scheme, SCHEME_FTP); + else + strcpy(purl->scheme, SCHEME_HTTP); + } + if (!purl->port) + purl->port = fetch_default_proxy_port(purl->scheme); + if (strcasecmp(purl->scheme, SCHEME_FTP) == 0 || + strcasecmp(purl->scheme, SCHEME_HTTP) == 0) + return (purl); + fetchFreeURL(purl); + } + return (NULL); +} + +/* + * Process an FTP request + */ +FILE * +ftp_request(struct url *url, const char *op, struct url_stat *us, + struct url *purl, const char *flags) +{ + conn_t *conn; + int oflag; + + /* check if we should use HTTP instead */ + if (purl && strcasecmp(purl->scheme, SCHEME_HTTP) == 0) { + if (strcmp(op, "STAT") == 0) + return (http_request(url, "HEAD", us, purl, flags)); + else if (strcmp(op, "RETR") == 0) + return (http_request(url, "GET", us, purl, flags)); + /* + * Our HTTP code doesn't support PUT requests yet, so try + * a direct connection. + */ + } + + /* connect to server */ + conn = ftp_cached_connect(url, purl, flags); + if (purl) + fetchFreeURL(purl); + if (conn == NULL) + return (NULL); + + /* change directory */ + if (ftp_cwd(conn, url->doc) == -1) + goto errsock; + + /* stat file */ + if (us && ftp_stat(conn, url->doc, us) == -1 + && fetchLastErrCode != FETCH_PROTO + && fetchLastErrCode != FETCH_UNAVAIL) + goto errsock; + + /* just a stat */ + if (strcmp(op, "STAT") == 0) { + --conn->ref; + ftp_disconnect(conn); + return (FILE *)1; /* bogus return value */ + } + if (strcmp(op, "STOR") == 0 || strcmp(op, "APPE") == 0) + oflag = O_WRONLY; + else + oflag = O_RDONLY; + + /* initiate the transfer */ + return (ftp_transfer(conn, op, url->doc, oflag, url->offset, flags)); + +errsock: + ftp_disconnect(conn); + return (NULL); +} + +/* + * Get and stat file + */ +FILE * +fetchXGetFTP(struct url *url, struct url_stat *us, const char *flags) +{ + return (ftp_request(url, "RETR", us, ftp_get_proxy(url, flags), flags)); +} + +/* + * Get file + */ +FILE * +fetchGetFTP(struct url *url, const char *flags) +{ + return (fetchXGetFTP(url, NULL, flags)); +} + +/* + * Put file + */ +FILE * +fetchPutFTP(struct url *url, const char *flags) +{ + return (ftp_request(url, CHECK_FLAG('a') ? "APPE" : "STOR", NULL, + ftp_get_proxy(url, flags), flags)); +} + +/* + * Get file stats + */ +int +fetchStatFTP(struct url *url, struct url_stat *us, const char *flags) +{ + FILE *f; + + f = ftp_request(url, "STAT", us, ftp_get_proxy(url, flags), flags); + if (f == NULL) + return (-1); + /* + * When op is "STAT", ftp_request() will return either NULL or + * (FILE *)1, never a valid FILE *, so we mustn't fclose(f) before + * returning, as it would cause a segfault. + */ + return (0); +} + +/* + * List a directory + */ +struct url_ent * +fetchListFTP(struct url *url __unused, const char *flags __unused) +{ + warnx("fetchListFTP(): not implemented"); + return (NULL); +} diff --git a/libfetch/ftp.errors b/libfetch/ftp.errors new file mode 100644 index 0000000..9a4a11e --- /dev/null +++ b/libfetch/ftp.errors @@ -0,0 +1,47 @@ +# $FreeBSD$ +# +# This list is taken from RFC 959. +# It probably needs a going over. +# +110 OK Restart marker reply +120 TEMP Service ready in a few minutes +125 OK Data connection already open; transfer starting +150 OK File status okay; about to open data connection +200 OK Command okay +202 PROTO Command not implemented, superfluous at this site +211 INFO System status, or system help reply +212 INFO Directory status +213 INFO File status +214 INFO Help message +215 INFO Set system type +220 OK Service ready for new user +221 OK Service closing control connection +225 OK Data connection open; no transfer in progress +226 OK Requested file action successful +227 OK Entering Passive Mode +229 OK Entering Extended Passive Mode +230 OK User logged in, proceed +250 OK Requested file action okay, completed +257 OK File/directory created +331 AUTH User name okay, need password +332 AUTH Need account for login +350 OK Requested file action pending further information +421 DOWN Service not available, closing control connection +425 NETWORK Can't open data connection +426 ABORT Connection closed; transfer aborted +450 UNAVAIL File unavailable (e.g., file busy) +451 SERVER Requested action aborted: local error in processing +452 FULL Insufficient storage space in system +500 PROTO Syntax error, command unrecognized +501 PROTO Syntax error in parameters or arguments +502 PROTO Command not implemented +503 PROTO Bad sequence of commands +504 PROTO Command not implemented for that parameter +530 AUTH Not logged in +532 AUTH Need account for storing files +535 PROTO Bug in MediaHawk Video Kernel FTP server +550 UNAVAIL File unavailable (e.g., file not found, no access) +551 PROTO Requested action aborted. Page type unknown +552 FULL Exceeded storage allocation +553 EXISTS File name not allowed +999 PROTO Protocol error diff --git a/libfetch/http.c b/libfetch/http.c new file mode 100644 index 0000000..f6e063a --- /dev/null +++ b/libfetch/http.c @@ -0,0 +1,2005 @@ +/*- + * Copyright (c) 2000-2011 Dag-Erling Smørgrav + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer + * in this position and unchanged. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. The name of the author may not be used to endorse or promote products + * derived from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +/* + * The following copyright applies to the base64 code: + * + *- + * Copyright 1997 Massachusetts Institute of Technology + * + * Permission to use, copy, modify, and distribute this software and + * its documentation for any purpose and without fee is hereby + * granted, provided that both the above copyright notice and this + * permission notice appear in all copies, that both the above + * copyright notice and this permission notice appear in all + * supporting documentation, and that the name of M.I.T. not be used + * in advertising or publicity pertaining to distribution of the + * software without specific, written prior permission. M.I.T. makes + * no representations about the suitability of this software for any + * purpose. It is provided "as is" without express or implied + * warranty. + * + * THIS SOFTWARE IS PROVIDED BY M.I.T. ``AS IS''. M.I.T. DISCLAIMS + * ALL EXPRESS OR IMPLIED WARRANTIES WITH REGARD TO THIS SOFTWARE, + * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT + * SHALL M.I.T. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF + * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND + * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, + * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT + * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#include "fetch.h" +#include "common.h" +#include "httperr.h" + +/* Maximum number of redirects to follow */ +#define MAX_REDIRECT 5 + +/* Symbolic names for reply codes we care about */ +#define HTTP_OK 200 +#define HTTP_PARTIAL 206 +#define HTTP_MOVED_PERM 301 +#define HTTP_MOVED_TEMP 302 +#define HTTP_SEE_OTHER 303 +#define HTTP_NOT_MODIFIED 304 +#define HTTP_TEMP_REDIRECT 307 +#define HTTP_NEED_AUTH 401 +#define HTTP_NEED_PROXY_AUTH 407 +#define HTTP_BAD_RANGE 416 +#define HTTP_PROTOCOL_ERROR 999 + +#define HTTP_REDIRECT(xyz) ((xyz) == HTTP_MOVED_PERM \ + || (xyz) == HTTP_MOVED_TEMP \ + || (xyz) == HTTP_TEMP_REDIRECT \ + || (xyz) == HTTP_SEE_OTHER) + +#define HTTP_ERROR(xyz) ((xyz) > 400 && (xyz) < 599) + + +/***************************************************************************** + * I/O functions for decoding chunked streams + */ + +struct httpio +{ + conn_t *conn; /* connection */ + int chunked; /* chunked mode */ + char *buf; /* chunk buffer */ + size_t bufsize; /* size of chunk buffer */ + ssize_t buflen; /* amount of data currently in buffer */ + int bufpos; /* current read offset in buffer */ + int eof; /* end-of-file flag */ + int error; /* error flag */ + size_t chunksize; /* remaining size of current chunk */ +#ifndef NDEBUG + size_t total; +#endif +}; + +/* + * Get next chunk header + */ +static int +http_new_chunk(struct httpio *io) +{ + char *p; + + if (fetch_getln(io->conn) == -1) + return (-1); + + if (io->conn->buflen < 2 || !isxdigit((unsigned char)*io->conn->buf)) + return (-1); + + for (p = io->conn->buf; *p && !isspace((unsigned char)*p); ++p) { + if (*p == ';') + break; + if (!isxdigit((unsigned char)*p)) + return (-1); + if (isdigit((unsigned char)*p)) { + io->chunksize = io->chunksize * 16 + + *p - '0'; + } else { + io->chunksize = io->chunksize * 16 + + 10 + tolower((unsigned char)*p) - 'a'; + } + } + +#ifndef NDEBUG + if (fetchDebug) { + io->total += io->chunksize; + if (io->chunksize == 0) + fprintf(stderr, "%s(): end of last chunk\n", __func__); + else + fprintf(stderr, "%s(): new chunk: %lu (%lu)\n", + __func__, (unsigned long)io->chunksize, + (unsigned long)io->total); + } +#endif + + return (io->chunksize); +} + +/* + * Grow the input buffer to at least len bytes + */ +static inline int +http_growbuf(struct httpio *io, size_t len) +{ + char *tmp; + + if (io->bufsize >= len) + return (0); + + if ((tmp = realloc(io->buf, len)) == NULL) + return (-1); + io->buf = tmp; + io->bufsize = len; + return (0); +} + +/* + * Fill the input buffer, do chunk decoding on the fly + */ +static int +http_fillbuf(struct httpio *io, size_t len) +{ + ssize_t nbytes; + + if (io->error) + return (-1); + if (io->eof) + return (0); + + if (io->chunked == 0) { + if (http_growbuf(io, len) == -1) + return (-1); + if ((nbytes = fetch_read(io->conn, io->buf, len)) == -1) { + io->error = errno; + return (-1); + } + io->buflen = nbytes; + io->bufpos = 0; + return (io->buflen); + } + + if (io->chunksize == 0) { + switch (http_new_chunk(io)) { + case -1: + io->error = 1; + return (-1); + case 0: + io->eof = 1; + return (0); + } + } + + if (len > io->chunksize) + len = io->chunksize; + if (http_growbuf(io, len) == -1) + return (-1); + if ((nbytes = fetch_read(io->conn, io->buf, len)) == -1) { + io->error = errno; + return (-1); + } + io->buflen = nbytes; + io->chunksize -= io->buflen; + + if (io->chunksize == 0) { + char endl[2]; + + if (fetch_read(io->conn, endl, 2) != 2 || + endl[0] != '\r' || endl[1] != '\n') + return (-1); + } + + io->bufpos = 0; + + return (io->buflen); +} + +/* + * Read function + */ +static int +http_readfn(void *v, char *buf, int len) +{ + struct httpio *io = (struct httpio *)v; + int l, pos; + + if (io->error) + return (-1); + if (io->eof) + return (0); + + for (pos = 0; len > 0; pos += l, len -= l) { + /* empty buffer */ + if (!io->buf || io->bufpos == io->buflen) + if (http_fillbuf(io, len) < 1) + break; + l = io->buflen - io->bufpos; + if (len < l) + l = len; + memcpy(buf + pos, io->buf + io->bufpos, l); + io->bufpos += l; + } + + if (!pos && io->error) { + if (io->error == EINTR) + io->error = 0; + return (-1); + } + return (pos); +} + +/* + * Write function + */ +static int +http_writefn(void *v, const char *buf, int len) +{ + struct httpio *io = (struct httpio *)v; + + return (fetch_write(io->conn, buf, len)); +} + +/* + * Close function + */ +static int +http_closefn(void *v) +{ + struct httpio *io = (struct httpio *)v; + int r; + + r = fetch_close(io->conn); + if (io->buf) + free(io->buf); + free(io); + return (r); +} + +/* + * Wrap a file descriptor up + */ +static FILE * +http_funopen(conn_t *conn, int chunked) +{ + struct httpio *io; + FILE *f; + + if ((io = calloc(1, sizeof(*io))) == NULL) { + fetch_syserr(); + return (NULL); + } + io->conn = conn; + io->chunked = chunked; + f = funopen(io, http_readfn, http_writefn, NULL, http_closefn); + if (f == NULL) { + fetch_syserr(); + free(io); + return (NULL); + } + return (f); +} + + +/***************************************************************************** + * Helper functions for talking to the server and parsing its replies + */ + +/* Header types */ +typedef enum { + hdr_syserror = -2, + hdr_error = -1, + hdr_end = 0, + hdr_unknown = 1, + hdr_content_length, + hdr_content_range, + hdr_last_modified, + hdr_location, + hdr_transfer_encoding, + hdr_www_authenticate, + hdr_proxy_authenticate, +} hdr_t; + +/* Names of interesting headers */ +static struct { + hdr_t num; + const char *name; +} hdr_names[] = { + { hdr_content_length, "Content-Length" }, + { hdr_content_range, "Content-Range" }, + { hdr_last_modified, "Last-Modified" }, + { hdr_location, "Location" }, + { hdr_transfer_encoding, "Transfer-Encoding" }, + { hdr_www_authenticate, "WWW-Authenticate" }, + { hdr_proxy_authenticate, "Proxy-Authenticate" }, + { hdr_unknown, NULL }, +}; + +/* + * Send a formatted line; optionally echo to terminal + */ +static int +http_cmd(conn_t *conn, const char *fmt, ...) +{ + va_list ap; + size_t len; + char *msg; + int r; + + va_start(ap, fmt); + len = vasprintf(&msg, fmt, ap); + va_end(ap); + + if (msg == NULL) { + errno = ENOMEM; + fetch_syserr(); + return (-1); + } + + r = fetch_putln(conn, msg, len); + free(msg); + + if (r == -1) { + fetch_syserr(); + return (-1); + } + + return (0); +} + +/* + * Get and parse status line + */ +static int +http_get_reply(conn_t *conn) +{ + char *p; + + if (fetch_getln(conn) == -1) + return (-1); + /* + * A valid status line looks like "HTTP/m.n xyz reason" where m + * and n are the major and minor protocol version numbers and xyz + * is the reply code. + * Unfortunately, there are servers out there (NCSA 1.5.1, to name + * just one) that do not send a version number, so we can't rely + * on finding one, but if we do, insist on it being 1.0 or 1.1. + * We don't care about the reason phrase. + */ + if (strncmp(conn->buf, "HTTP", 4) != 0) + return (HTTP_PROTOCOL_ERROR); + p = conn->buf + 4; + if (*p == '/') { + if (p[1] != '1' || p[2] != '.' || (p[3] != '0' && p[3] != '1')) + return (HTTP_PROTOCOL_ERROR); + p += 4; + } + if (*p != ' ' || + !isdigit((unsigned char)p[1]) || + !isdigit((unsigned char)p[2]) || + !isdigit((unsigned char)p[3])) + return (HTTP_PROTOCOL_ERROR); + + conn->err = (p[1] - '0') * 100 + (p[2] - '0') * 10 + (p[3] - '0'); + return (conn->err); +} + +/* + * Check a header; if the type matches the given string, return a pointer + * to the beginning of the value. + */ +static const char * +http_match(const char *str, const char *hdr) +{ + while (*str && *hdr && + tolower((unsigned char)*str++) == tolower((unsigned char)*hdr++)) + /* nothing */; + if (*str || *hdr != ':') + return (NULL); + while (*hdr && isspace((unsigned char)*++hdr)) + /* nothing */; + return (hdr); +} + + +/* + * Get the next header and return the appropriate symbolic code. We + * need to read one line ahead for checking for a continuation line + * belonging to the current header (continuation lines start with + * white space). + * + * We get called with a fresh line already in the conn buffer, either + * from the previous http_next_header() invocation, or, the first + * time, from a fetch_getln() performed by our caller. + * + * This stops when we encounter an empty line (we dont read beyond the header + * area). + * + * Note that the "headerbuf" is just a place to return the result. Its + * contents are not used for the next call. This means that no cleanup + * is needed when ie doing another connection, just call the cleanup when + * fully done to deallocate memory. + */ + +/* Limit the max number of continuation lines to some reasonable value */ +#define HTTP_MAX_CONT_LINES 10 + +/* Place into which to build a header from one or several lines */ +typedef struct { + char *buf; /* buffer */ + size_t bufsize; /* buffer size */ + size_t buflen; /* length of buffer contents */ +} http_headerbuf_t; + +static void +init_http_headerbuf(http_headerbuf_t *buf) +{ + buf->buf = NULL; + buf->bufsize = 0; + buf->buflen = 0; +} + +static void +clean_http_headerbuf(http_headerbuf_t *buf) +{ + if (buf->buf) + free(buf->buf); + init_http_headerbuf(buf); +} + +/* Remove whitespace at the end of the buffer */ +static void +http_conn_trimright(conn_t *conn) +{ + while (conn->buflen && + isspace((unsigned char)conn->buf[conn->buflen - 1])) + conn->buflen--; + conn->buf[conn->buflen] = '\0'; +} + +static hdr_t +http_next_header(conn_t *conn, http_headerbuf_t *hbuf, const char **p) +{ + unsigned int i, len; + + /* + * Have to do the stripping here because of the first line. So + * it's done twice for the subsequent lines. No big deal + */ + http_conn_trimright(conn); + if (conn->buflen == 0) + return (hdr_end); + + /* Copy the line to the headerbuf */ + if (hbuf->bufsize < conn->buflen + 1) { + if ((hbuf->buf = realloc(hbuf->buf, conn->buflen + 1)) == NULL) + return (hdr_syserror); + hbuf->bufsize = conn->buflen + 1; + } + strcpy(hbuf->buf, conn->buf); + hbuf->buflen = conn->buflen; + + /* + * Fetch possible continuation lines. Stop at 1st non-continuation + * and leave it in the conn buffer + */ + for (i = 0; i < HTTP_MAX_CONT_LINES; i++) { + if (fetch_getln(conn) == -1) + return (hdr_syserror); + + /* + * Note: we carry on the idea from the previous version + * that a pure whitespace line is equivalent to an empty + * one (so it's not continuation and will be handled when + * we are called next) + */ + http_conn_trimright(conn); + if (conn->buf[0] != ' ' && conn->buf[0] != "\t"[0]) + break; + + /* Got a continuation line. Concatenate to previous */ + len = hbuf->buflen + conn->buflen; + if (hbuf->bufsize < len + 1) { + len *= 2; + if ((hbuf->buf = realloc(hbuf->buf, len + 1)) == NULL) + return (hdr_syserror); + hbuf->bufsize = len + 1; + } + strcpy(hbuf->buf + hbuf->buflen, conn->buf); + hbuf->buflen += conn->buflen; + } + + /* + * We could check for malformed headers but we don't really care. + * A valid header starts with a token immediately followed by a + * colon; a token is any sequence of non-control, non-whitespace + * characters except "()<>@,;:\\\"{}". + */ + for (i = 0; hdr_names[i].num != hdr_unknown; i++) + if ((*p = http_match(hdr_names[i].name, hbuf->buf)) != NULL) + return (hdr_names[i].num); + + return (hdr_unknown); +} + +/************************** + * [Proxy-]Authenticate header parsing + */ + +/* + * Read doublequote-delimited string into output buffer obuf (allocated + * by caller, whose responsibility it is to ensure that it's big enough) + * cp points to the first char after the initial '"' + * Handles \ quoting + * Returns pointer to the first char after the terminating double quote, or + * NULL for error. + */ +static const char * +http_parse_headerstring(const char *cp, char *obuf) +{ + for (;;) { + switch (*cp) { + case 0: /* Unterminated string */ + *obuf = 0; + return (NULL); + case '"': /* Ending quote */ + *obuf = 0; + return (++cp); + case '\\': + if (*++cp == 0) { + *obuf = 0; + return (NULL); + } + /* FALLTHROUGH */ + default: + *obuf++ = *cp++; + } + } +} + +/* Http auth challenge schemes */ +typedef enum {HTTPAS_UNKNOWN, HTTPAS_BASIC,HTTPAS_DIGEST} http_auth_schemes_t; + +/* Data holder for a Basic or Digest challenge. */ +typedef struct { + http_auth_schemes_t scheme; + char *realm; + char *qop; + char *nonce; + char *opaque; + char *algo; + int stale; + int nc; /* Nonce count */ +} http_auth_challenge_t; + +static void +init_http_auth_challenge(http_auth_challenge_t *b) +{ + b->scheme = HTTPAS_UNKNOWN; + b->realm = b->qop = b->nonce = b->opaque = b->algo = NULL; + b->stale = b->nc = 0; +} + +static void +clean_http_auth_challenge(http_auth_challenge_t *b) +{ + if (b->realm) + free(b->realm); + if (b->qop) + free(b->qop); + if (b->nonce) + free(b->nonce); + if (b->opaque) + free(b->opaque); + if (b->algo) + free(b->algo); + init_http_auth_challenge(b); +} + +/* Data holder for an array of challenges offered in an http response. */ +#define MAX_CHALLENGES 10 +typedef struct { + http_auth_challenge_t *challenges[MAX_CHALLENGES]; + int count; /* Number of parsed challenges in the array */ + int valid; /* We did parse an authenticate header */ +} http_auth_challenges_t; + +static void +init_http_auth_challenges(http_auth_challenges_t *cs) +{ + int i; + for (i = 0; i < MAX_CHALLENGES; i++) + cs->challenges[i] = NULL; + cs->count = cs->valid = 0; +} + +static void +clean_http_auth_challenges(http_auth_challenges_t *cs) +{ + int i; + /* We rely on non-zero pointers being allocated, not on the count */ + for (i = 0; i < MAX_CHALLENGES; i++) { + if (cs->challenges[i] != NULL) { + clean_http_auth_challenge(cs->challenges[i]); + free(cs->challenges[i]); + } + } + init_http_auth_challenges(cs); +} + +/* + * Enumeration for lexical elements. Separators will be returned as their own + * ascii value + */ +typedef enum {HTTPHL_WORD=256, HTTPHL_STRING=257, HTTPHL_END=258, + HTTPHL_ERROR = 259} http_header_lex_t; + +/* + * Determine what kind of token comes next and return possible value + * in buf, which is supposed to have been allocated big enough by + * caller. Advance input pointer and return element type. + */ +static int +http_header_lex(const char **cpp, char *buf) +{ + size_t l; + /* Eat initial whitespace */ + *cpp += strspn(*cpp, " \t"); + if (**cpp == 0) + return (HTTPHL_END); + + /* Separator ? */ + if (**cpp == ',' || **cpp == '=') + return (*((*cpp)++)); + + /* String ? */ + if (**cpp == '"') { + *cpp = http_parse_headerstring(++*cpp, buf); + if (*cpp == NULL) + return (HTTPHL_ERROR); + return (HTTPHL_STRING); + } + + /* Read other token, until separator or whitespace */ + l = strcspn(*cpp, " \t,="); + memcpy(buf, *cpp, l); + buf[l] = 0; + *cpp += l; + return (HTTPHL_WORD); +} + +/* + * Read challenges from http xxx-authenticate header and accumulate them + * in the challenges list structure. + * + * Headers with multiple challenges are specified by rfc2617, but + * servers (ie: squid) often send them in separate headers instead, + * which in turn is forbidden by the http spec (multiple headers with + * the same name are only allowed for pure comma-separated lists, see + * rfc2616 sec 4.2). + * + * We support both approaches anyway + */ +static int +http_parse_authenticate(const char *cp, http_auth_challenges_t *cs) +{ + int ret = -1; + http_header_lex_t lex; + char *key = malloc(strlen(cp) + 1); + char *value = malloc(strlen(cp) + 1); + char *buf = malloc(strlen(cp) + 1); + + if (key == NULL || value == NULL || buf == NULL) { + fetch_syserr(); + goto out; + } + + /* In any case we've seen the header and we set the valid bit */ + cs->valid = 1; + + /* Need word first */ + lex = http_header_lex(&cp, key); + if (lex != HTTPHL_WORD) + goto out; + + /* Loop on challenges */ + for (; cs->count < MAX_CHALLENGES; cs->count++) { + cs->challenges[cs->count] = + malloc(sizeof(http_auth_challenge_t)); + if (cs->challenges[cs->count] == NULL) { + fetch_syserr(); + goto out; + } + init_http_auth_challenge(cs->challenges[cs->count]); + if (!strcasecmp(key, "basic")) { + cs->challenges[cs->count]->scheme = HTTPAS_BASIC; + } else if (!strcasecmp(key, "digest")) { + cs->challenges[cs->count]->scheme = HTTPAS_DIGEST; + } else { + cs->challenges[cs->count]->scheme = HTTPAS_UNKNOWN; + /* + * Continue parsing as basic or digest may + * follow, and the syntax is the same for + * all. We'll just ignore this one when + * looking at the list + */ + } + + /* Loop on attributes */ + for (;;) { + /* Key */ + lex = http_header_lex(&cp, key); + if (lex != HTTPHL_WORD) + goto out; + + /* Equal sign */ + lex = http_header_lex(&cp, buf); + if (lex != '=') + goto out; + + /* Value */ + lex = http_header_lex(&cp, value); + if (lex != HTTPHL_WORD && lex != HTTPHL_STRING) + goto out; + + if (!strcasecmp(key, "realm")) + cs->challenges[cs->count]->realm = + strdup(value); + else if (!strcasecmp(key, "qop")) + cs->challenges[cs->count]->qop = + strdup(value); + else if (!strcasecmp(key, "nonce")) + cs->challenges[cs->count]->nonce = + strdup(value); + else if (!strcasecmp(key, "opaque")) + cs->challenges[cs->count]->opaque = + strdup(value); + else if (!strcasecmp(key, "algorithm")) + cs->challenges[cs->count]->algo = + strdup(value); + else if (!strcasecmp(key, "stale")) + cs->challenges[cs->count]->stale = + strcasecmp(value, "no"); + /* Else ignore unknown attributes */ + + /* Comma or Next challenge or End */ + lex = http_header_lex(&cp, key); + /* + * If we get a word here, this is the beginning of the + * next challenge. Break the attributes loop + */ + if (lex == HTTPHL_WORD) + break; + + if (lex == HTTPHL_END) { + /* End while looking for ',' is normal exit */ + cs->count++; + ret = 0; + goto out; + } + /* Anything else is an error */ + if (lex != ',') + goto out; + + } /* End attributes loop */ + } /* End challenge loop */ + + /* + * Challenges max count exceeded. This really can't happen + * with normal data, something's fishy -> error + */ + +out: + if (key) + free(key); + if (value) + free(value); + if (buf) + free(buf); + return (ret); +} + + +/* + * Parse a last-modified header + */ +static int +http_parse_mtime(const char *p, time_t *mtime) +{ + char locale[64], *r; + struct tm tm; + + strncpy(locale, setlocale(LC_TIME, NULL), sizeof(locale)); + setlocale(LC_TIME, "C"); + r = strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm); + /* XXX should add support for date-2 and date-3 */ + setlocale(LC_TIME, locale); + if (r == NULL) + return (-1); + DEBUG(fprintf(stderr, "last modified: [%04d-%02d-%02d " + "%02d:%02d:%02d]\n", + tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday, + tm.tm_hour, tm.tm_min, tm.tm_sec)); + *mtime = timegm(&tm); + return (0); +} + +/* + * Parse a content-length header + */ +static int +http_parse_length(const char *p, off_t *length) +{ + off_t len; + + for (len = 0; *p && isdigit((unsigned char)*p); ++p) + len = len * 10 + (*p - '0'); + if (*p) + return (-1); + DEBUG(fprintf(stderr, "content length: [%lld]\n", + (long long)len)); + *length = len; + return (0); +} + +/* + * Parse a content-range header + */ +static int +http_parse_range(const char *p, off_t *offset, off_t *length, off_t *size) +{ + off_t first, last, len; + + if (strncasecmp(p, "bytes ", 6) != 0) + return (-1); + p += 6; + if (*p == '*') { + first = last = -1; + ++p; + } else { + for (first = 0; *p && isdigit((unsigned char)*p); ++p) + first = first * 10 + *p - '0'; + if (*p != '-') + return (-1); + for (last = 0, ++p; *p && isdigit((unsigned char)*p); ++p) + last = last * 10 + *p - '0'; + } + if (first > last || *p != '/') + return (-1); + for (len = 0, ++p; *p && isdigit((unsigned char)*p); ++p) + len = len * 10 + *p - '0'; + if (*p || len < last - first + 1) + return (-1); + if (first == -1) { + DEBUG(fprintf(stderr, "content range: [*/%lld]\n", + (long long)len)); + *length = 0; + } else { + DEBUG(fprintf(stderr, "content range: [%lld-%lld/%lld]\n", + (long long)first, (long long)last, (long long)len)); + *length = last - first + 1; + } + *offset = first; + *size = len; + return (0); +} + + +/***************************************************************************** + * Helper functions for authorization + */ + +/* + * Base64 encoding + */ +static char * +http_base64(const char *src) +{ + static const char base64[] = + "ABCDEFGHIJKLMNOPQRSTUVWXYZ" + "abcdefghijklmnopqrstuvwxyz" + "0123456789+/"; + char *str, *dst; + size_t l; + int t, r; + + l = strlen(src); + if ((str = malloc(((l + 2) / 3) * 4 + 1)) == NULL) + return (NULL); + dst = str; + r = 0; + + while (l >= 3) { + t = (src[0] << 16) | (src[1] << 8) | src[2]; + dst[0] = base64[(t >> 18) & 0x3f]; + dst[1] = base64[(t >> 12) & 0x3f]; + dst[2] = base64[(t >> 6) & 0x3f]; + dst[3] = base64[(t >> 0) & 0x3f]; + src += 3; l -= 3; + dst += 4; r += 4; + } + + switch (l) { + case 2: + t = (src[0] << 16) | (src[1] << 8); + dst[0] = base64[(t >> 18) & 0x3f]; + dst[1] = base64[(t >> 12) & 0x3f]; + dst[2] = base64[(t >> 6) & 0x3f]; + dst[3] = '='; + dst += 4; + r += 4; + break; + case 1: + t = src[0] << 16; + dst[0] = base64[(t >> 18) & 0x3f]; + dst[1] = base64[(t >> 12) & 0x3f]; + dst[2] = dst[3] = '='; + dst += 4; + r += 4; + break; + case 0: + break; + } + + *dst = 0; + return (str); +} + + +/* + * Extract authorization parameters from environment value. + * The value is like scheme:realm:user:pass + */ +typedef struct { + char *scheme; + char *realm; + char *user; + char *password; +} http_auth_params_t; + +static void +init_http_auth_params(http_auth_params_t *s) +{ + s->scheme = s->realm = s->user = s->password = 0; +} + +static void +clean_http_auth_params(http_auth_params_t *s) +{ + if (s->scheme) + free(s->scheme); + if (s->realm) + free(s->realm); + if (s->user) + free(s->user); + if (s->password) + free(s->password); + init_http_auth_params(s); +} + +static int +http_authfromenv(const char *p, http_auth_params_t *parms) +{ + int ret = -1; + char *v, *ve; + char *str = strdup(p); + + if (str == NULL) { + fetch_syserr(); + return (-1); + } + v = str; + + if ((ve = strchr(v, ':')) == NULL) + goto out; + + *ve = 0; + if ((parms->scheme = strdup(v)) == NULL) { + fetch_syserr(); + goto out; + } + v = ve + 1; + + if ((ve = strchr(v, ':')) == NULL) + goto out; + + *ve = 0; + if ((parms->realm = strdup(v)) == NULL) { + fetch_syserr(); + goto out; + } + v = ve + 1; + + if ((ve = strchr(v, ':')) == NULL) + goto out; + + *ve = 0; + if ((parms->user = strdup(v)) == NULL) { + fetch_syserr(); + goto out; + } + v = ve + 1; + + + if ((parms->password = strdup(v)) == NULL) { + fetch_syserr(); + goto out; + } + ret = 0; +out: + if (ret == -1) + clean_http_auth_params(parms); + if (str) + free(str); + return (ret); +} + + +/* + * Digest response: the code to compute the digest is taken from the + * sample implementation in RFC2616 + */ +#define IN const +#define OUT + +#define HASHLEN 16 +typedef char HASH[HASHLEN]; +#define HASHHEXLEN 32 +typedef char HASHHEX[HASHHEXLEN+1]; + +static const char *hexchars = "0123456789abcdef"; +static void +CvtHex(IN HASH Bin, OUT HASHHEX Hex) +{ + unsigned short i; + unsigned char j; + + for (i = 0; i < HASHLEN; i++) { + j = (Bin[i] >> 4) & 0xf; + Hex[i*2] = hexchars[j]; + j = Bin[i] & 0xf; + Hex[i*2+1] = hexchars[j]; + }; + Hex[HASHHEXLEN] = '\0'; +}; + +/* calculate H(A1) as per spec */ +static void +DigestCalcHA1( + IN char * pszAlg, + IN char * pszUserName, + IN char * pszRealm, + IN char * pszPassword, + IN char * pszNonce, + IN char * pszCNonce, + OUT HASHHEX SessionKey + ) +{ + MD5_CTX Md5Ctx; + HASH HA1; + + MD5Init(&Md5Ctx); + MD5Update(&Md5Ctx, pszUserName, strlen(pszUserName)); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszRealm, strlen(pszRealm)); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszPassword, strlen(pszPassword)); + MD5Final(HA1, &Md5Ctx); + if (strcasecmp(pszAlg, "md5-sess") == 0) { + + MD5Init(&Md5Ctx); + MD5Update(&Md5Ctx, HA1, HASHLEN); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszNonce, strlen(pszNonce)); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszCNonce, strlen(pszCNonce)); + MD5Final(HA1, &Md5Ctx); + }; + CvtHex(HA1, SessionKey); +} + +/* calculate request-digest/response-digest as per HTTP Digest spec */ +static void +DigestCalcResponse( + IN HASHHEX HA1, /* H(A1) */ + IN char * pszNonce, /* nonce from server */ + IN char * pszNonceCount, /* 8 hex digits */ + IN char * pszCNonce, /* client nonce */ + IN char * pszQop, /* qop-value: "", "auth", "auth-int" */ + IN char * pszMethod, /* method from the request */ + IN char * pszDigestUri, /* requested URL */ + IN HASHHEX HEntity, /* H(entity body) if qop="auth-int" */ + OUT HASHHEX Response /* request-digest or response-digest */ + ) +{ +/* DEBUG(fprintf(stderr, + "Calc: HA1[%s] Nonce[%s] qop[%s] method[%s] URI[%s]\n", + HA1, pszNonce, pszQop, pszMethod, pszDigestUri));*/ + MD5_CTX Md5Ctx; + HASH HA2; + HASH RespHash; + HASHHEX HA2Hex; + + // calculate H(A2) + MD5Init(&Md5Ctx); + MD5Update(&Md5Ctx, pszMethod, strlen(pszMethod)); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszDigestUri, strlen(pszDigestUri)); + if (strcasecmp(pszQop, "auth-int") == 0) { + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, HEntity, HASHHEXLEN); + }; + MD5Final(HA2, &Md5Ctx); + CvtHex(HA2, HA2Hex); + + // calculate response + MD5Init(&Md5Ctx); + MD5Update(&Md5Ctx, HA1, HASHHEXLEN); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszNonce, strlen(pszNonce)); + MD5Update(&Md5Ctx, ":", 1); + if (*pszQop) { + MD5Update(&Md5Ctx, pszNonceCount, strlen(pszNonceCount)); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszCNonce, strlen(pszCNonce)); + MD5Update(&Md5Ctx, ":", 1); + MD5Update(&Md5Ctx, pszQop, strlen(pszQop)); + MD5Update(&Md5Ctx, ":", 1); + }; + MD5Update(&Md5Ctx, HA2Hex, HASHHEXLEN); + MD5Final(RespHash, &Md5Ctx); + CvtHex(RespHash, Response); +} + +/* + * Generate/Send a Digest authorization header + * This looks like: [Proxy-]Authorization: credentials + * + * credentials = "Digest" digest-response + * digest-response = 1#( username | realm | nonce | digest-uri + * | response | [ algorithm ] | [cnonce] | + * [opaque] | [message-qop] | + * [nonce-count] | [auth-param] ) + * username = "username" "=" username-value + * username-value = quoted-string + * digest-uri = "uri" "=" digest-uri-value + * digest-uri-value = request-uri ; As specified by HTTP/1.1 + * message-qop = "qop" "=" qop-value + * cnonce = "cnonce" "=" cnonce-value + * cnonce-value = nonce-value + * nonce-count = "nc" "=" nc-value + * nc-value = 8LHEX + * response = "response" "=" request-digest + * request-digest = <"> 32LHEX <"> + */ +static int +http_digest_auth(conn_t *conn, const char *hdr, http_auth_challenge_t *c, + http_auth_params_t *parms, struct url *url) +{ + int r; + char noncecount[10]; + char cnonce[40]; + char *options = 0; + + if (!c->realm || !c->nonce) { + DEBUG(fprintf(stderr, "realm/nonce not set in challenge\n")); + return(-1); + } + if (!c->algo) + c->algo = strdup(""); + + if (asprintf(&options, "%s%s%s%s", + *c->algo? ",algorithm=" : "", c->algo, + c->opaque? ",opaque=" : "", c->opaque?c->opaque:"")== -1) + return (-1); + + if (!c->qop) { + c->qop = strdup(""); + *noncecount = 0; + *cnonce = 0; + } else { + c->nc++; + sprintf(noncecount, "%08x", c->nc); + /* We don't try very hard with the cnonce ... */ + sprintf(cnonce, "%x%lx", getpid(), (unsigned long)time(0)); + } + + HASHHEX HA1; + DigestCalcHA1(c->algo, parms->user, c->realm, + parms->password, c->nonce, cnonce, HA1); + DEBUG(fprintf(stderr, "HA1: [%s]\n", HA1)); + HASHHEX digest; + DigestCalcResponse(HA1, c->nonce, noncecount, cnonce, c->qop, + "GET", url->doc, "", digest); + + if (c->qop[0]) { + r = http_cmd(conn, "%s: Digest username=\"%s\",realm=\"%s\"," + "nonce=\"%s\",uri=\"%s\",response=\"%s\"," + "qop=\"auth\", cnonce=\"%s\", nc=%s%s", + hdr, parms->user, c->realm, + c->nonce, url->doc, digest, + cnonce, noncecount, options); + } else { + r = http_cmd(conn, "%s: Digest username=\"%s\",realm=\"%s\"," + "nonce=\"%s\",uri=\"%s\",response=\"%s\"%s", + hdr, parms->user, c->realm, + c->nonce, url->doc, digest, options); + } + if (options) + free(options); + return (r); +} + +/* + * Encode username and password + */ +static int +http_basic_auth(conn_t *conn, const char *hdr, const char *usr, const char *pwd) +{ + char *upw, *auth; + int r; + + DEBUG(fprintf(stderr, "basic: usr: [%s]\n", usr)); + DEBUG(fprintf(stderr, "basic: pwd: [%s]\n", pwd)); + if (asprintf(&upw, "%s:%s", usr, pwd) == -1) + return (-1); + auth = http_base64(upw); + free(upw); + if (auth == NULL) + return (-1); + r = http_cmd(conn, "%s: Basic %s", hdr, auth); + free(auth); + return (r); +} + +/* + * Chose the challenge to answer and call the appropriate routine to + * produce the header. + */ +static int +http_authorize(conn_t *conn, const char *hdr, http_auth_challenges_t *cs, + http_auth_params_t *parms, struct url *url) +{ + http_auth_challenge_t *basic = NULL; + http_auth_challenge_t *digest = NULL; + int i; + + /* If user or pass are null we're not happy */ + if (!parms->user || !parms->password) { + DEBUG(fprintf(stderr, "NULL usr or pass\n")); + return (-1); + } + + /* Look for a Digest and a Basic challenge */ + for (i = 0; i < cs->count; i++) { + if (cs->challenges[i]->scheme == HTTPAS_BASIC) + basic = cs->challenges[i]; + if (cs->challenges[i]->scheme == HTTPAS_DIGEST) + digest = cs->challenges[i]; + } + + /* Error if "Digest" was specified and there is no Digest challenge */ + if (!digest && (parms->scheme && + !strcasecmp(parms->scheme, "digest"))) { + DEBUG(fprintf(stderr, + "Digest auth in env, not supported by peer\n")); + return (-1); + } + /* + * If "basic" was specified in the environment, or there is no Digest + * challenge, do the basic thing. Don't need a challenge for this, + * so no need to check basic!=NULL + */ + if (!digest || (parms->scheme && !strcasecmp(parms->scheme,"basic"))) + return (http_basic_auth(conn,hdr,parms->user,parms->password)); + + /* Else, prefer digest. We just checked that it's not NULL */ + return (http_digest_auth(conn, hdr, digest, parms, url)); +} + +/***************************************************************************** + * Helper functions for connecting to a server or proxy + */ + +/* + * Connect to the correct HTTP server or proxy. + */ +static conn_t * +http_connect(struct url *URL, struct url *purl, const char *flags) +{ + conn_t *conn; + int verbose; + int af, val; + +#ifdef INET6 + af = AF_UNSPEC; +#else + af = AF_INET; +#endif + + verbose = CHECK_FLAG('v'); + if (CHECK_FLAG('4')) + af = AF_INET; +#ifdef INET6 + else if (CHECK_FLAG('6')) + af = AF_INET6; +#endif + + if (purl && strcasecmp(URL->scheme, SCHEME_HTTPS) != 0) { + URL = purl; + } else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) { + /* can't talk http to an ftp server */ + /* XXX should set an error code */ + return (NULL); + } + + if ((conn = fetch_connect(URL->host, URL->port, af, verbose)) == NULL) + /* fetch_connect() has already set an error code */ + return (NULL); + if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0 && + fetch_ssl(conn, verbose) == -1) { + fetch_close(conn); + /* grrr */ + errno = EAUTH; + fetch_syserr(); + return (NULL); + } + + val = 1; + setsockopt(conn->sd, IPPROTO_TCP, TCP_NOPUSH, &val, sizeof(val)); + + return (conn); +} + +static struct url * +http_get_proxy(struct url * url, const char *flags) +{ + struct url *purl; + char *p; + + if (flags != NULL && strchr(flags, 'd') != NULL) + return (NULL); + if (fetch_no_proxy_match(url->host)) + return (NULL); + if (((p = getenv("HTTP_PROXY")) || (p = getenv("http_proxy"))) && + *p && (purl = fetchParseURL(p))) { + if (!*purl->scheme) + strcpy(purl->scheme, SCHEME_HTTP); + if (!purl->port) + purl->port = fetch_default_proxy_port(purl->scheme); + if (strcasecmp(purl->scheme, SCHEME_HTTP) == 0) + return (purl); + fetchFreeURL(purl); + } + return (NULL); +} + +static void +http_print_html(FILE *out, FILE *in) +{ + size_t len; + char *line, *p, *q; + int comment, tag; + + comment = tag = 0; + while ((line = fgetln(in, &len)) != NULL) { + while (len && isspace((unsigned char)line[len - 1])) + --len; + for (p = q = line; q < line + len; ++q) { + if (comment && *q == '-') { + if (q + 2 < line + len && + strcmp(q, "-->") == 0) { + tag = comment = 0; + q += 2; + } + } else if (tag && !comment && *q == '>') { + p = q + 1; + tag = 0; + } else if (!tag && *q == '<') { + if (q > p) + fwrite(p, q - p, 1, out); + tag = 1; + if (q + 3 < line + len && + strcmp(q, "