apk-tools/libfetch/fetch.c

620 lines
14 KiB
C
Raw Normal View History

/* $NetBSD: fetch.c,v 1.19 2009/08/11 20:48:06 joerg Exp $ */
/*-
* Copyright (c) 1998-2004 Dag-Erling Co<EFBFBD>dan Sm<EFBFBD>rgrav
* Copyright (c) 2008 Joerg Sonnenberger <joerg@NetBSD.org>
* 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: fetch.c,v 1.41 2007/12/19 00:26:36 des Exp $
*/
#include <ctype.h>
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "fetch.h"
#include "common.h"
auth_t fetchAuthMethod;
int fetchLastErrCode;
char fetchLastErrString[MAXERRSTRING];
int fetchTimeout;
volatile 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.
*/
fetchIO *
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.
*/
fetchIO *
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.
*/
fetchIO *
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.
*/
int
fetchList(struct url_list *ue, struct url *URL, const char *pattern,
const char *flags)
{
if (strcasecmp(URL->scheme, SCHEME_FILE) == 0)
return (fetchListFile(ue, URL, pattern, flags));
else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0)
return (fetchListFTP(ue, URL, pattern, flags));
else if (strcasecmp(URL->scheme, SCHEME_HTTP) == 0)
return (fetchListHTTP(ue, URL, pattern, flags));
else if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0)
return (fetchListHTTP(ue, URL, pattern, flags));
url_seterr(URL_BAD_SCHEME);
return -1;
}
/*
* Attempt to parse the given URL; if successful, call fetchXGet().
*/
fetchIO *
fetchXGetURL(const char *URL, struct url_stat *us, const char *flags)
{
struct url *u;
fetchIO *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().
*/
fetchIO *
fetchGetURL(const char *URL, const char *flags)
{
return (fetchXGetURL(URL, NULL, flags));
}
/*
* Attempt to parse the given URL; if successful, call fetchPut().
*/
fetchIO *
fetchPutURL(const char *URL, const char *flags)
{
struct url *u;
fetchIO *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().
*/
int
fetchListURL(struct url_list *ue, const char *URL, const char *pattern,
const char *flags)
{
struct url *u;
int rv;
if ((u = fetchParseURL(URL)) == NULL)
return -1;
rv = fetchList(ue, u, pattern, flags);
fetchFreeURL(u);
return rv;
}
/*
* 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);
}
int
fetch_urlpath_safe(char x)
{
if ((x >= '0' && x <= '9') || (x >= 'A' && x <= 'Z') ||
(x >= 'a' && x <= 'z'))
return 1;
switch (x) {
case '$':
case '-':
case '_':
case '.':
case '+':
case '!':
case '*':
case '\'':
case '(':
case ')':
case ',':
/* The following are allowed in segment and path components: */
case '?':
case ':':
case '@':
case '&':
case '=':
case '/':
case ';':
/* If something is already quoted... */
case '%':
return 1;
default:
return 0;
}
}
/*
* Copy an existing URL.
*/
struct url *
fetchCopyURL(const struct url *src)
{
struct url *dst;
char *doc;
/* allocate struct url */
if ((dst = malloc(sizeof(*dst))) == NULL) {
fetch_syserr();
return (NULL);
}
if ((doc = strdup(src->doc)) == NULL) {
fetch_syserr();
free(dst);
return (NULL);
}
*dst = *src;
dst->doc = doc;
return dst;
}
/*
* 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)
{
const char *p, *q;
struct url *u;
size_t i, count;
int pre_quoted;
/* allocate struct url */
if ((u = calloc(1, sizeof(*u))) == NULL) {
fetch_syserr();
return (NULL);
}
if (*URL == '/') {
pre_quoted = 0;
strcpy(u->scheme, SCHEME_FILE);
p = URL;
goto quote_doc;
}
if (strncmp(URL, "file:", 5) == 0) {
pre_quoted = 1;
strcpy(u->scheme, SCHEME_FILE);
URL += 5;
if (URL[0] != '/' || URL[1] != '/' || URL[2] != '/') {
url_seterr(URL_MALFORMED);
goto ouch;
}
p = URL + 2;
goto quote_doc;
}
if (strncmp(URL, "http:", 5) == 0 ||
strncmp(URL, "https:", 6) == 0) {
pre_quoted = 1;
if (URL[4] == ':') {
strcpy(u->scheme, SCHEME_HTTP);
URL += 5;
} else {
strcpy(u->scheme, SCHEME_HTTPS);
URL += 6;
}
if (URL[0] != '/' || URL[1] != '/') {
url_seterr(URL_MALFORMED);
goto ouch;
}
URL += 2;
p = URL;
goto find_user;
}
if (strncmp(URL, "ftp:", 4) == 0) {
pre_quoted = 1;
strcpy(u->scheme, SCHEME_FTP);
URL += 4;
if (URL[0] != '/' || URL[1] != '/') {
url_seterr(URL_MALFORMED);
goto ouch;
}
URL += 2;
p = URL;
goto find_user;
}
url_seterr(URL_BAD_SCHEME);
goto ouch;
find_user:
p = strpbrk(URL, "/@");
if (p != NULL && *p == '@') {
/* username */
for (q = URL, i = 0; (*q != ':') && (*q != '@'); q++) {
if (i < URL_USERLEN)
u->user[i++] = *q;
}
/* password */
if (*q == ':') {
for (q++, i = 0; (*q != '@'); q++)
if (i < URL_PWDLEN)
u->pwd[i++] = *q;
}
p++;
} else {
p = URL;
}
/* hostname */
if (*p == '[' && (q = strchr(p + 1, ']')) != NULL &&
(*++q == '\0' || *q == '/' || *q == ':')) {
if ((i = q - p - 2) > URL_HOSTLEN)
i = URL_HOSTLEN;
strncpy(u->host, ++p, i);
p = q;
} else {
for (i = 0; *p && (*p != '/') && (*p != ':'); p++)
if (i < URL_HOSTLEN)
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;
}
/* document */
if (!*p)
p = "/";
quote_doc:
count = 1;
for (i = 0; p[i] != '\0'; ++i) {
if ((!pre_quoted && p[i] == '%') ||
!fetch_urlpath_safe(p[i]))
count += 3;
else
++count;
}
if ((u->doc = malloc(count)) == NULL) {
fetch_syserr();
goto ouch;
}
for (i = 0; *p != '\0'; ++p) {
if ((!pre_quoted && *p == '%') ||
!fetch_urlpath_safe(*p)) {
u->doc[i++] = '%';
if ((unsigned char)*p < 160)
u->doc[i++] = '0' + ((unsigned char)*p) / 16;
else
u->doc[i++] = 'a' - 10 + ((unsigned char)*p) / 16;
if ((unsigned char)*p % 16 < 10)
u->doc[i++] = '0' + ((unsigned char)*p) % 16;
else
u->doc[i++] = 'a' - 10 + ((unsigned char)*p) % 16;
} else
u->doc[i++] = *p;
}
u->doc[i] = '\0';
return (u);
ouch:
free(u);
return (NULL);
}
/*
* Free a URL
*/
void
fetchFreeURL(struct url *u)
{
free(u->doc);
free(u);
}
static char
xdigit2digit(char digit)
{
digit = tolower((unsigned char)digit);
if (digit >= 'a' && digit <= 'f')
digit = digit - 'a' + 10;
else
digit = digit - '0';
return digit;
}
/*
* Unquote whole URL.
* Skips optional parts like query or fragment identifier.
*/
char *
fetchUnquotePath(struct url *url)
{
char *unquoted;
const char *iter;
size_t i;
if ((unquoted = malloc(strlen(url->doc) + 1)) == NULL)
return NULL;
for (i = 0, iter = url->doc; *iter != '\0'; ++iter) {
if (*iter == '#' || *iter == '?')
break;
if (iter[0] != '%' ||
!isxdigit((unsigned char)iter[1]) ||
!isxdigit((unsigned char)iter[2])) {
unquoted[i++] = *iter;
continue;
}
unquoted[i++] = xdigit2digit(iter[1]) * 16 +
xdigit2digit(iter[2]);
iter += 2;
}
unquoted[i] = '\0';
return unquoted;
}
/*
* Extract the file name component of a URL.
*/
char *
fetchUnquoteFilename(struct url *url)
{
char *unquoted, *filename;
const char *last_slash;
if ((unquoted = fetchUnquotePath(url)) == NULL)
return NULL;
if ((last_slash = strrchr(unquoted, '/')) == NULL)
return unquoted;
filename = strdup(last_slash + 1);
free(unquoted);
return filename;
}
char *
fetchStringifyURL(const struct url *url)
{
size_t total;
char *doc;
/* scheme :// user : pwd @ host :port doc */
total = strlen(url->scheme) + 3 + strlen(url->user) + 1 +
strlen(url->pwd) + 1 + strlen(url->host) + 6 + strlen(url->doc) + 1;
if ((doc = malloc(total)) == NULL)
return NULL;
if (url->port != 0)
snprintf(doc, total, "%s%s%s%s%s%s%s:%d%s",
url->scheme,
url->scheme[0] != '\0' ? "://" : "",
url->user,
url->pwd[0] != '\0' ? ":" : "",
url->pwd,
url->user[0] != '\0' || url->pwd[0] != '\0' ? "@" : "",
url->host,
(int)url->port,
url->doc);
else {
snprintf(doc, total, "%s%s%s%s%s%s%s%s",
url->scheme,
url->scheme[0] != '\0' ? "://" : "",
url->user,
url->pwd[0] != '\0' ? ":" : "",
url->pwd,
url->user[0] != '\0' || url->pwd[0] != '\0' ? "@" : "",
url->host,
url->doc);
}
return doc;
}