2006-01-26 04:38:36 +08:00
|
|
|
#include "git-compat-util.h"
|
2005-07-05 02:57:58 +08:00
|
|
|
#include "cache.h"
|
2017-06-15 02:07:36 +08:00
|
|
|
#include "config.h"
|
2005-07-06 06:44:09 +08:00
|
|
|
#include "pkt-line.h"
|
2005-07-08 15:02:52 +08:00
|
|
|
#include "quote.h"
|
2005-10-16 15:25:26 +08:00
|
|
|
#include "refs.h"
|
2007-03-13 07:00:19 +08:00
|
|
|
#include "run-command.h"
|
2007-05-12 23:45:59 +08:00
|
|
|
#include "remote.h"
|
2013-07-09 04:56:53 +08:00
|
|
|
#include "connect.h"
|
2010-05-23 17:19:44 +08:00
|
|
|
#include "url.h"
|
2013-09-18 10:10:31 +08:00
|
|
|
#include "string-list.h"
|
2013-12-05 21:02:29 +08:00
|
|
|
#include "sha1-array.h"
|
transport: add a protocol-whitelist environment variable
If we are cloning an untrusted remote repository into a
sandbox, we may also want to fetch remote submodules in
order to get the complete view as intended by the other
side. However, that opens us up to attacks where a malicious
user gets us to clone something they would not otherwise
have access to (this is not necessarily a problem by itself,
but we may then act on the cloned contents in a way that
exposes them to the attacker).
Ideally such a setup would sandbox git entirely away from
high-value items, but this is not always practical or easy
to set up (e.g., OS network controls may block multiple
protocols, and we would want to enable some but not others).
We can help this case by providing a way to restrict
particular protocols. We use a whitelist in the environment.
This is more annoying to set up than a blacklist, but
defaults to safety if the set of protocols git supports
grows). If no whitelist is specified, we continue to default
to allowing all protocols (this is an "unsafe" default, but
since the minority of users will want this sandboxing
effect, it is the only sensible one).
A note on the tests: ideally these would all be in a single
test file, but the git-daemon and httpd test infrastructure
is an all-or-nothing proposition rather than a test-by-test
prerequisite. By putting them all together, we would be
unable to test the file-local code on machines without
apache.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-17 01:12:52 +08:00
|
|
|
#include "transport.h"
|
2017-09-27 07:56:19 +08:00
|
|
|
#include "strbuf.h"
|
2017-10-17 01:55:27 +08:00
|
|
|
#include "protocol.h"
|
2005-07-05 02:57:58 +08:00
|
|
|
|
2006-08-16 01:23:48 +08:00
|
|
|
static char *server_capabilities;
|
2013-09-18 07:29:28 +08:00
|
|
|
static const char *parse_feature_value(const char *, const char *, int *);
|
2005-10-28 10:48:54 +08:00
|
|
|
|
2014-08-30 17:46:54 +08:00
|
|
|
static int check_ref(const char *name, unsigned int flags)
|
Improve git-peek-remote
This makes git-peek-remote able to basically do everything that
git-ls-remote does (but obviously just for the native protocol, so no
http[s]: or rsync: support).
The default behaviour is the same, but you can now give a mixture of
"--refs", "--tags" and "--heads" flags, where "--refs" forces
git-peek-remote to only show real refs (ie none of the fakey tag lookups,
but also not the special pseudo-refs like HEAD and MERGE_HEAD).
The "--tags" and "--heads" flags respectively limit the output to just
regular tags and heads, of course.
You can still also ask to limit them by name too.
You can combine the flags, so
git peek-remote --refs --tags .
will show all local _true_ tags, without the generated tag lookups
(compare the output without the "--refs" flag).
And "--tags --heads" will show both tags and heads, but will avoid (for
example) any special refs outside of the standard locations.
I'm also planning on adding a "--ignore-local" flag that allows us to ask
it to ignore any refs that we already have in the local tree, but that's
an independent thing.
All this is obviously gearing up to making "git fetch" cheaper.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-05 03:29:10 +08:00
|
|
|
{
|
|
|
|
if (!flags)
|
|
|
|
return 1;
|
|
|
|
|
2014-08-30 17:46:54 +08:00
|
|
|
if (!skip_prefix(name, "refs/", &name))
|
Improve git-peek-remote
This makes git-peek-remote able to basically do everything that
git-ls-remote does (but obviously just for the native protocol, so no
http[s]: or rsync: support).
The default behaviour is the same, but you can now give a mixture of
"--refs", "--tags" and "--heads" flags, where "--refs" forces
git-peek-remote to only show real refs (ie none of the fakey tag lookups,
but also not the special pseudo-refs like HEAD and MERGE_HEAD).
The "--tags" and "--heads" flags respectively limit the output to just
regular tags and heads, of course.
You can still also ask to limit them by name too.
You can combine the flags, so
git peek-remote --refs --tags .
will show all local _true_ tags, without the generated tag lookups
(compare the output without the "--refs" flag).
And "--tags --heads" will show both tags and heads, but will avoid (for
example) any special refs outside of the standard locations.
I'm also planning on adding a "--ignore-local" flag that allows us to ask
it to ignore any refs that we already have in the local tree, but that's
an independent thing.
All this is obviously gearing up to making "git fetch" cheaper.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-05 03:29:10 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* REF_NORMAL means that we don't want the magic fake tag refs */
|
2011-09-16 05:10:25 +08:00
|
|
|
if ((flags & REF_NORMAL) && check_refname_format(name, 0))
|
Improve git-peek-remote
This makes git-peek-remote able to basically do everything that
git-ls-remote does (but obviously just for the native protocol, so no
http[s]: or rsync: support).
The default behaviour is the same, but you can now give a mixture of
"--refs", "--tags" and "--heads" flags, where "--refs" forces
git-peek-remote to only show real refs (ie none of the fakey tag lookups,
but also not the special pseudo-refs like HEAD and MERGE_HEAD).
The "--tags" and "--heads" flags respectively limit the output to just
regular tags and heads, of course.
You can still also ask to limit them by name too.
You can combine the flags, so
git peek-remote --refs --tags .
will show all local _true_ tags, without the generated tag lookups
(compare the output without the "--refs" flag).
And "--tags --heads" will show both tags and heads, but will avoid (for
example) any special refs outside of the standard locations.
I'm also planning on adding a "--ignore-local" flag that allows us to ask
it to ignore any refs that we already have in the local tree, but that's
an independent thing.
All this is obviously gearing up to making "git fetch" cheaper.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-05 03:29:10 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* REF_HEADS means that we want regular branch heads */
|
2014-08-30 17:46:54 +08:00
|
|
|
if ((flags & REF_HEADS) && starts_with(name, "heads/"))
|
Improve git-peek-remote
This makes git-peek-remote able to basically do everything that
git-ls-remote does (but obviously just for the native protocol, so no
http[s]: or rsync: support).
The default behaviour is the same, but you can now give a mixture of
"--refs", "--tags" and "--heads" flags, where "--refs" forces
git-peek-remote to only show real refs (ie none of the fakey tag lookups,
but also not the special pseudo-refs like HEAD and MERGE_HEAD).
The "--tags" and "--heads" flags respectively limit the output to just
regular tags and heads, of course.
You can still also ask to limit them by name too.
You can combine the flags, so
git peek-remote --refs --tags .
will show all local _true_ tags, without the generated tag lookups
(compare the output without the "--refs" flag).
And "--tags --heads" will show both tags and heads, but will avoid (for
example) any special refs outside of the standard locations.
I'm also planning on adding a "--ignore-local" flag that allows us to ask
it to ignore any refs that we already have in the local tree, but that's
an independent thing.
All this is obviously gearing up to making "git fetch" cheaper.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-05 03:29:10 +08:00
|
|
|
return 1;
|
|
|
|
|
|
|
|
/* REF_TAGS means that we want tags */
|
2014-08-30 17:46:54 +08:00
|
|
|
if ((flags & REF_TAGS) && starts_with(name, "tags/"))
|
Improve git-peek-remote
This makes git-peek-remote able to basically do everything that
git-ls-remote does (but obviously just for the native protocol, so no
http[s]: or rsync: support).
The default behaviour is the same, but you can now give a mixture of
"--refs", "--tags" and "--heads" flags, where "--refs" forces
git-peek-remote to only show real refs (ie none of the fakey tag lookups,
but also not the special pseudo-refs like HEAD and MERGE_HEAD).
The "--tags" and "--heads" flags respectively limit the output to just
regular tags and heads, of course.
You can still also ask to limit them by name too.
You can combine the flags, so
git peek-remote --refs --tags .
will show all local _true_ tags, without the generated tag lookups
(compare the output without the "--refs" flag).
And "--tags --heads" will show both tags and heads, but will avoid (for
example) any special refs outside of the standard locations.
I'm also planning on adding a "--ignore-local" flag that allows us to ask
it to ignore any refs that we already have in the local tree, but that's
an independent thing.
All this is obviously gearing up to making "git fetch" cheaper.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-07-05 03:29:10 +08:00
|
|
|
return 1;
|
|
|
|
|
|
|
|
/* All type bits clear means that we are ok with anything */
|
|
|
|
return !(flags & ~REF_NORMAL);
|
|
|
|
}
|
|
|
|
|
2007-10-30 09:05:40 +08:00
|
|
|
int check_ref_type(const struct ref *ref, int flags)
|
|
|
|
{
|
2014-08-30 17:46:54 +08:00
|
|
|
return check_ref(ref->name, flags);
|
2007-10-30 09:05:40 +08:00
|
|
|
}
|
|
|
|
|
2016-09-10 01:36:29 +08:00
|
|
|
static void die_initial_contact(int unexpected)
|
2012-06-20 02:24:50 +08:00
|
|
|
{
|
2016-09-10 01:36:29 +08:00
|
|
|
if (unexpected)
|
2016-09-19 21:08:17 +08:00
|
|
|
die(_("The remote end hung up upon initial contact"));
|
2012-06-20 02:24:50 +08:00
|
|
|
else
|
2016-09-19 21:08:17 +08:00
|
|
|
die(_("Could not read from remote repository.\n\n"
|
|
|
|
"Please make sure you have the correct access rights\n"
|
|
|
|
"and the repository exists."));
|
2012-06-20 02:24:50 +08:00
|
|
|
}
|
|
|
|
|
2013-09-18 10:10:31 +08:00
|
|
|
static void parse_one_symref_info(struct string_list *symref, const char *val, int len)
|
|
|
|
{
|
|
|
|
char *sym, *target;
|
|
|
|
struct string_list_item *item;
|
|
|
|
|
|
|
|
if (!len)
|
|
|
|
return; /* just "symref" */
|
|
|
|
/* e.g. "symref=HEAD:refs/heads/master" */
|
2014-07-19 23:35:34 +08:00
|
|
|
sym = xmemdupz(val, len);
|
2013-09-18 10:10:31 +08:00
|
|
|
target = strchr(sym, ':');
|
|
|
|
if (!target)
|
|
|
|
/* just "symref=something" */
|
|
|
|
goto reject;
|
|
|
|
*(target++) = '\0';
|
|
|
|
if (check_refname_format(sym, REFNAME_ALLOW_ONELEVEL) ||
|
|
|
|
check_refname_format(target, REFNAME_ALLOW_ONELEVEL))
|
|
|
|
/* "symref=bogus:pair */
|
|
|
|
goto reject;
|
connect.c: fix leak in parse_one_symref_info()
If we successfully parse a symref value like
"HEAD:refs/heads/master", we add the result to a string
list. But because the string list is marked
STRING_LIST_INIT_DUP, the string list code will make a copy
of the string and add the copy.
This patch fixes it by adding the entry with
string_list_append_nodup(), which lets the string list take
ownership of our newly allocated string. There are two
alternatives that seem like they would work, but aren't the
right solution.
The first is to initialize the list with the "NODUP"
initializer. That would avoid the copy, but then the string
list would not realize that it owns the strings. When we
eventually call string_list_clear(), it would not free the
strings, causing a leak.
The second option would be to use the normal
string_list_append(), but free the local copy in our
function. We can't do this because the local copy actually
contains _two_ strings; the symref name and its target. We
point to the target pointer via the "util" field, and its
memory must last as long as the string list does.
You may also wonder whether it's safe to ever free the local
copy, since the target points into it. The answer is yes,
because we duplicate it in annotaate_refs_with_symref_info
before clearing the string list.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-05-26 03:33:05 +08:00
|
|
|
item = string_list_append_nodup(symref, sym);
|
2013-09-18 10:10:31 +08:00
|
|
|
item->util = target;
|
|
|
|
return;
|
|
|
|
reject:
|
|
|
|
free(sym);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void annotate_refs_with_symref_info(struct ref *ref)
|
|
|
|
{
|
|
|
|
struct string_list symref = STRING_LIST_INIT_DUP;
|
|
|
|
const char *feature_list = server_capabilities;
|
|
|
|
|
|
|
|
while (feature_list) {
|
|
|
|
int len;
|
|
|
|
const char *val;
|
|
|
|
|
|
|
|
val = parse_feature_value(feature_list, "symref", &len);
|
|
|
|
if (!val)
|
|
|
|
break;
|
|
|
|
parse_one_symref_info(&symref, val, len);
|
|
|
|
feature_list = val + 1;
|
|
|
|
}
|
2014-11-25 16:02:35 +08:00
|
|
|
string_list_sort(&symref);
|
2013-09-18 10:10:31 +08:00
|
|
|
|
|
|
|
for (; ref; ref = ref->next) {
|
|
|
|
struct string_list_item *item;
|
|
|
|
item = string_list_lookup(&symref, ref->name);
|
|
|
|
if (!item)
|
|
|
|
continue;
|
|
|
|
ref->symref = xstrdup((char *)item->util);
|
|
|
|
}
|
|
|
|
string_list_clear(&symref, 0);
|
|
|
|
}
|
|
|
|
|
2017-09-27 07:56:19 +08:00
|
|
|
/*
|
|
|
|
* Read one line of a server's ref advertisement into packet_buffer.
|
|
|
|
*/
|
|
|
|
static int read_remote_ref(int in, char **src_buf, size_t *src_len,
|
|
|
|
int *responded)
|
|
|
|
{
|
|
|
|
int len = packet_read(in, src_buf, src_len,
|
|
|
|
packet_buffer, sizeof(packet_buffer),
|
|
|
|
PACKET_READ_GENTLE_ON_EOF |
|
|
|
|
PACKET_READ_CHOMP_NEWLINE);
|
|
|
|
const char *arg;
|
|
|
|
if (len < 0)
|
|
|
|
die_initial_contact(*responded);
|
|
|
|
if (len > 4 && skip_prefix(packet_buffer, "ERR ", &arg))
|
|
|
|
die("remote error: %s", arg);
|
|
|
|
|
|
|
|
*responded = 1;
|
|
|
|
|
|
|
|
return len;
|
|
|
|
}
|
|
|
|
|
2017-10-17 01:55:27 +08:00
|
|
|
#define EXPECTING_PROTOCOL_VERSION 0
|
|
|
|
#define EXPECTING_FIRST_REF 1
|
|
|
|
#define EXPECTING_REF 2
|
|
|
|
#define EXPECTING_SHALLOW 3
|
|
|
|
|
|
|
|
/* Returns 1 if packet_buffer is a protocol version pkt-line, 0 otherwise. */
|
|
|
|
static int process_protocol_version(void)
|
|
|
|
{
|
|
|
|
switch (determine_protocol_version_client(packet_buffer)) {
|
|
|
|
case protocol_v1:
|
|
|
|
return 1;
|
|
|
|
case protocol_v0:
|
|
|
|
return 0;
|
|
|
|
default:
|
|
|
|
die("server is speaking an unknown protocol");
|
|
|
|
}
|
|
|
|
}
|
2017-09-27 07:56:19 +08:00
|
|
|
|
|
|
|
static void process_capabilities(int *len)
|
|
|
|
{
|
|
|
|
int nul_location = strlen(packet_buffer);
|
|
|
|
if (nul_location == *len)
|
|
|
|
return;
|
|
|
|
server_capabilities = xstrdup(packet_buffer + nul_location + 1);
|
|
|
|
*len = nul_location;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int process_dummy_ref(void)
|
|
|
|
{
|
|
|
|
struct object_id oid;
|
|
|
|
const char *name;
|
|
|
|
|
|
|
|
if (parse_oid_hex(packet_buffer, &oid, &name))
|
|
|
|
return 0;
|
|
|
|
if (*name != ' ')
|
|
|
|
return 0;
|
|
|
|
name++;
|
|
|
|
|
|
|
|
return !oidcmp(&null_oid, &oid) && !strcmp(name, "capabilities^{}");
|
|
|
|
}
|
|
|
|
|
|
|
|
static void check_no_capabilities(int len)
|
|
|
|
{
|
|
|
|
if (strlen(packet_buffer) != len)
|
|
|
|
warning("Ignoring capabilities after first line '%s'",
|
|
|
|
packet_buffer + strlen(packet_buffer));
|
|
|
|
}
|
|
|
|
|
|
|
|
static int process_ref(int len, struct ref ***list, unsigned int flags,
|
|
|
|
struct oid_array *extra_have)
|
|
|
|
{
|
|
|
|
struct object_id old_oid;
|
|
|
|
const char *name;
|
|
|
|
|
|
|
|
if (parse_oid_hex(packet_buffer, &old_oid, &name))
|
|
|
|
return 0;
|
|
|
|
if (*name != ' ')
|
|
|
|
return 0;
|
|
|
|
name++;
|
|
|
|
|
|
|
|
if (extra_have && !strcmp(name, ".have")) {
|
|
|
|
oid_array_append(extra_have, &old_oid);
|
|
|
|
} else if (!strcmp(name, "capabilities^{}")) {
|
|
|
|
die("protocol error: unexpected capabilities^{}");
|
|
|
|
} else if (check_ref(name, flags)) {
|
|
|
|
struct ref *ref = alloc_ref(name);
|
|
|
|
oidcpy(&ref->old_oid, &old_oid);
|
|
|
|
**list = ref;
|
|
|
|
*list = &ref->next;
|
|
|
|
}
|
|
|
|
check_no_capabilities(len);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int process_shallow(int len, struct oid_array *shallow_points)
|
|
|
|
{
|
|
|
|
const char *arg;
|
|
|
|
struct object_id old_oid;
|
|
|
|
|
|
|
|
if (!skip_prefix(packet_buffer, "shallow ", &arg))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (get_oid_hex(arg, &old_oid))
|
|
|
|
die("protocol error: expected shallow sha-1, got '%s'", arg);
|
|
|
|
if (!shallow_points)
|
|
|
|
die("repository on the other end cannot be shallow");
|
|
|
|
oid_array_append(shallow_points, &old_oid);
|
|
|
|
check_no_capabilities(len);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2005-07-17 04:55:50 +08:00
|
|
|
/*
|
|
|
|
* Read all the refs from the other end
|
|
|
|
*/
|
2013-02-21 04:06:45 +08:00
|
|
|
struct ref **get_remote_heads(int in, char *src_buf, size_t src_len,
|
|
|
|
struct ref **list, unsigned int flags,
|
2017-03-31 09:40:00 +08:00
|
|
|
struct oid_array *extra_have,
|
|
|
|
struct oid_array *shallow_points)
|
2005-07-17 04:55:50 +08:00
|
|
|
{
|
2013-09-18 10:10:31 +08:00
|
|
|
struct ref **orig_list = list;
|
2016-09-10 01:36:29 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* A hang-up after seeing some response from the other end
|
|
|
|
* means that it is unexpected, as we know the other end is
|
|
|
|
* willing to talk to us. A hang-up before seeing any
|
|
|
|
* response does not necessarily mean an ACL problem, though.
|
|
|
|
*/
|
2017-09-27 07:56:19 +08:00
|
|
|
int responded = 0;
|
|
|
|
int len;
|
2017-10-17 01:55:27 +08:00
|
|
|
int state = EXPECTING_PROTOCOL_VERSION;
|
2012-06-20 02:24:50 +08:00
|
|
|
|
2005-07-17 04:55:50 +08:00
|
|
|
*list = NULL;
|
2005-10-14 09:57:40 +08:00
|
|
|
|
2017-09-27 07:56:19 +08:00
|
|
|
while ((len = read_remote_ref(in, &src_buf, &src_len, &responded))) {
|
|
|
|
switch (state) {
|
2017-10-17 01:55:27 +08:00
|
|
|
case EXPECTING_PROTOCOL_VERSION:
|
|
|
|
if (process_protocol_version()) {
|
|
|
|
state = EXPECTING_FIRST_REF;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state = EXPECTING_FIRST_REF;
|
|
|
|
/* fallthrough */
|
2017-09-27 07:56:19 +08:00
|
|
|
case EXPECTING_FIRST_REF:
|
|
|
|
process_capabilities(&len);
|
|
|
|
if (process_dummy_ref()) {
|
|
|
|
state = EXPECTING_SHALLOW;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
state = EXPECTING_REF;
|
|
|
|
/* fallthrough */
|
|
|
|
case EXPECTING_REF:
|
|
|
|
if (process_ref(len, &list, flags, extra_have))
|
|
|
|
break;
|
|
|
|
state = EXPECTING_SHALLOW;
|
|
|
|
/* fallthrough */
|
|
|
|
case EXPECTING_SHALLOW:
|
|
|
|
if (process_shallow(len, shallow_points))
|
|
|
|
break;
|
|
|
|
die("protocol error: unexpected '%s'", packet_buffer);
|
|
|
|
default:
|
|
|
|
die("unexpected state %d", state);
|
2005-10-28 10:48:54 +08:00
|
|
|
}
|
2005-07-17 04:55:50 +08:00
|
|
|
}
|
2013-09-18 10:10:31 +08:00
|
|
|
|
|
|
|
annotate_refs_with_symref_info(*orig_list);
|
|
|
|
|
2005-07-17 04:55:50 +08:00
|
|
|
return list;
|
|
|
|
}
|
|
|
|
|
2013-09-18 07:29:28 +08:00
|
|
|
static const char *parse_feature_value(const char *feature_list, const char *feature, int *lenp)
|
2012-01-09 05:06:19 +08:00
|
|
|
{
|
|
|
|
int len;
|
|
|
|
|
|
|
|
if (!feature_list)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
len = strlen(feature);
|
|
|
|
while (*feature_list) {
|
|
|
|
const char *found = strstr(feature_list, feature);
|
|
|
|
if (!found)
|
|
|
|
return NULL;
|
parse_feature_request: make it easier to see feature values
We already take care to parse key/value capabilities like
"foo=bar", but the code does not provide a good way of
actually finding out what is on the right-hand side of the
"=".
A server using "parse_feature_request" could accomplish this
with some extra parsing. You must skip past the "key"
portion manually, check for "=" versus NUL or space, and
then find the length by searching for the next space (or
NUL). But clients can't even do that, since the
"server_supports" interface does not even return the
pointer.
Instead, let's have our parser share more information by
providing a pointer to the value and its length. The
"parse_feature_value" function returns a pointer to the
feature's value portion, along with the length of the value.
If the feature is missing, NULL is returned. If it does not
have an "=", then a zero-length value is returned.
Similarly, "server_feature_value" behaves in the same way,
but always checks the static server_feature_list variable.
We can then implement "server_supports" in terms of
"server_feature_value". We cannot implement the original
"parse_feature_request" in terms of our new function,
because it returned a pointer to the beginning of the
feature. However, no callers actually cared about the value
of the returned pointer, so we can simplify it to a boolean
just as we do for "server_supports".
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-08-14 09:59:27 +08:00
|
|
|
if (feature_list == found || isspace(found[-1])) {
|
|
|
|
const char *value = found + len;
|
|
|
|
/* feature with no value (e.g., "thin-pack") */
|
|
|
|
if (!*value || isspace(*value)) {
|
|
|
|
if (lenp)
|
|
|
|
*lenp = 0;
|
|
|
|
return value;
|
|
|
|
}
|
|
|
|
/* feature with a value (e.g., "agent=git/1.2.3") */
|
|
|
|
else if (*value == '=') {
|
|
|
|
value++;
|
|
|
|
if (lenp)
|
|
|
|
*lenp = strcspn(value, " \t\n");
|
|
|
|
return value;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* otherwise we matched a substring of another feature;
|
|
|
|
* keep looking
|
|
|
|
*/
|
|
|
|
}
|
2012-01-09 05:06:19 +08:00
|
|
|
feature_list = found + 1;
|
|
|
|
}
|
|
|
|
return NULL;
|
2005-10-28 10:48:54 +08:00
|
|
|
}
|
|
|
|
|
parse_feature_request: make it easier to see feature values
We already take care to parse key/value capabilities like
"foo=bar", but the code does not provide a good way of
actually finding out what is on the right-hand side of the
"=".
A server using "parse_feature_request" could accomplish this
with some extra parsing. You must skip past the "key"
portion manually, check for "=" versus NUL or space, and
then find the length by searching for the next space (or
NUL). But clients can't even do that, since the
"server_supports" interface does not even return the
pointer.
Instead, let's have our parser share more information by
providing a pointer to the value and its length. The
"parse_feature_value" function returns a pointer to the
feature's value portion, along with the length of the value.
If the feature is missing, NULL is returned. If it does not
have an "=", then a zero-length value is returned.
Similarly, "server_feature_value" behaves in the same way,
but always checks the static server_feature_list variable.
We can then implement "server_supports" in terms of
"server_feature_value". We cannot implement the original
"parse_feature_request" in terms of our new function,
because it returned a pointer to the beginning of the
feature. However, no callers actually cared about the value
of the returned pointer, so we can simplify it to a boolean
just as we do for "server_supports".
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-08-14 09:59:27 +08:00
|
|
|
int parse_feature_request(const char *feature_list, const char *feature)
|
|
|
|
{
|
|
|
|
return !!parse_feature_value(feature_list, feature, NULL);
|
|
|
|
}
|
|
|
|
|
|
|
|
const char *server_feature_value(const char *feature, int *len)
|
|
|
|
{
|
|
|
|
return parse_feature_value(server_capabilities, feature, len);
|
|
|
|
}
|
|
|
|
|
|
|
|
int server_supports(const char *feature)
|
|
|
|
{
|
|
|
|
return !!server_feature_value(feature, NULL);
|
|
|
|
}
|
|
|
|
|
2005-07-14 09:46:20 +08:00
|
|
|
enum protocol {
|
|
|
|
PROTO_LOCAL = 1,
|
2013-11-29 03:50:03 +08:00
|
|
|
PROTO_FILE,
|
2005-07-14 09:46:20 +08:00
|
|
|
PROTO_SSH,
|
2010-05-14 17:31:35 +08:00
|
|
|
PROTO_GIT
|
2005-07-14 09:46:20 +08:00
|
|
|
};
|
|
|
|
|
2013-11-29 03:50:03 +08:00
|
|
|
int url_is_local_not_ssh(const char *url)
|
|
|
|
{
|
|
|
|
const char *colon = strchr(url, ':');
|
|
|
|
const char *slash = strchr(url, '/');
|
|
|
|
return !colon || (slash && slash < colon) ||
|
|
|
|
has_dos_drive_prefix(url);
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:49:17 +08:00
|
|
|
static const char *prot_name(enum protocol protocol)
|
|
|
|
{
|
|
|
|
switch (protocol) {
|
|
|
|
case PROTO_LOCAL:
|
2013-11-29 03:50:03 +08:00
|
|
|
case PROTO_FILE:
|
2013-11-29 03:49:17 +08:00
|
|
|
return "file";
|
|
|
|
case PROTO_SSH:
|
|
|
|
return "ssh";
|
|
|
|
case PROTO_GIT:
|
|
|
|
return "git";
|
|
|
|
default:
|
2015-09-24 20:44:49 +08:00
|
|
|
return "unknown protocol";
|
2013-11-29 03:49:17 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-07-14 09:46:20 +08:00
|
|
|
static enum protocol get_protocol(const char *name)
|
|
|
|
{
|
|
|
|
if (!strcmp(name, "ssh"))
|
|
|
|
return PROTO_SSH;
|
|
|
|
if (!strcmp(name, "git"))
|
|
|
|
return PROTO_GIT;
|
2016-02-15 22:29:06 +08:00
|
|
|
if (!strcmp(name, "git+ssh")) /* deprecated - do not use */
|
2005-10-15 08:14:56 +08:00
|
|
|
return PROTO_SSH;
|
2016-02-15 22:29:06 +08:00
|
|
|
if (!strcmp(name, "ssh+git")) /* deprecated - do not use */
|
2005-10-15 08:14:56 +08:00
|
|
|
return PROTO_SSH;
|
2007-08-02 01:03:37 +08:00
|
|
|
if (!strcmp(name, "file"))
|
2013-11-29 03:50:03 +08:00
|
|
|
return PROTO_FILE;
|
2005-07-14 09:46:20 +08:00
|
|
|
die("I don't handle protocol '%s'", name);
|
|
|
|
}
|
|
|
|
|
2015-02-21 23:52:48 +08:00
|
|
|
static char *host_end(char **hoststart, int removebrackets)
|
|
|
|
{
|
|
|
|
char *host = *hoststart;
|
|
|
|
char *end;
|
|
|
|
char *start = strstr(host, "@[");
|
|
|
|
if (start)
|
|
|
|
start++; /* Jump over '@' */
|
|
|
|
else
|
|
|
|
start = host;
|
|
|
|
if (start[0] == '[') {
|
|
|
|
end = strchr(start + 1, ']');
|
|
|
|
if (end) {
|
|
|
|
if (removebrackets) {
|
|
|
|
*end = 0;
|
|
|
|
memmove(start, start + 1, end - start);
|
|
|
|
end++;
|
|
|
|
}
|
|
|
|
} else
|
|
|
|
end = host;
|
|
|
|
} else
|
|
|
|
end = host;
|
|
|
|
return end;
|
|
|
|
}
|
|
|
|
|
2005-07-21 21:10:36 +08:00
|
|
|
#define STR_(s) # s
|
|
|
|
#define STR(s) STR_(s)
|
2005-07-14 09:46:20 +08:00
|
|
|
|
2010-02-18 04:56:02 +08:00
|
|
|
static void get_host_and_port(char **host, const char **port)
|
|
|
|
{
|
|
|
|
char *colon, *end;
|
2015-02-21 23:52:48 +08:00
|
|
|
end = host_end(host, 1);
|
2010-02-18 04:56:02 +08:00
|
|
|
colon = strchr(end, ':');
|
|
|
|
if (colon) {
|
2015-02-21 23:52:48 +08:00
|
|
|
long portnr = strtol(colon + 1, &end, 10);
|
|
|
|
if (end != colon + 1 && *end == '\0' && 0 <= portnr && portnr < 65536) {
|
|
|
|
*colon = 0;
|
|
|
|
*port = colon + 1;
|
2015-04-08 04:03:25 +08:00
|
|
|
} else if (!colon[1]) {
|
|
|
|
*colon = 0;
|
2015-02-21 23:52:48 +08:00
|
|
|
}
|
2010-02-18 04:56:02 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-12-06 12:39:36 +08:00
|
|
|
static void enable_keepalive(int sockfd)
|
|
|
|
{
|
|
|
|
int ka = 1;
|
|
|
|
|
|
|
|
if (setsockopt(sockfd, SOL_SOCKET, SO_KEEPALIVE, &ka, sizeof(ka)) < 0)
|
|
|
|
fprintf(stderr, "unable to set SO_KEEPALIVE on socket: %s\n",
|
|
|
|
strerror(errno));
|
|
|
|
}
|
|
|
|
|
2005-09-29 07:52:21 +08:00
|
|
|
#ifndef NO_IPV6
|
2005-09-29 07:37:58 +08:00
|
|
|
|
2007-05-24 05:34:27 +08:00
|
|
|
static const char *ai_name(const struct addrinfo *ai)
|
|
|
|
{
|
2009-04-24 20:16:41 +08:00
|
|
|
static char addr[NI_MAXHOST];
|
|
|
|
if (getnameinfo(ai->ai_addr, ai->ai_addrlen, addr, sizeof(addr), NULL, 0,
|
|
|
|
NI_NUMERICHOST) != 0)
|
2015-09-25 05:06:08 +08:00
|
|
|
xsnprintf(addr, sizeof(addr), "(unknown)");
|
2009-04-24 20:16:41 +08:00
|
|
|
|
2007-05-24 05:34:27 +08:00
|
|
|
return addr;
|
|
|
|
}
|
|
|
|
|
2006-06-07 11:58:41 +08:00
|
|
|
/*
|
|
|
|
* Returns a connected socket() fd, or else die()s.
|
|
|
|
*/
|
2007-05-17 01:09:41 +08:00
|
|
|
static int git_tcp_connect_sock(char *host, int flags)
|
2005-07-14 09:46:20 +08:00
|
|
|
{
|
2011-07-13 00:28:34 +08:00
|
|
|
struct strbuf error_message = STRBUF_INIT;
|
|
|
|
int sockfd = -1;
|
2006-06-28 17:04:39 +08:00
|
|
|
const char *port = STR(DEFAULT_GIT_PORT);
|
2005-07-21 21:10:36 +08:00
|
|
|
struct addrinfo hints, *ai0, *ai;
|
|
|
|
int gai;
|
2007-05-24 05:34:27 +08:00
|
|
|
int cnt = 0;
|
2005-07-21 21:10:36 +08:00
|
|
|
|
2010-02-18 04:56:02 +08:00
|
|
|
get_host_and_port(&host, &port);
|
|
|
|
if (!*port)
|
|
|
|
port = "<none>";
|
2005-07-21 21:10:36 +08:00
|
|
|
|
|
|
|
memset(&hints, 0, sizeof(hints));
|
2016-02-03 12:09:14 +08:00
|
|
|
if (flags & CONNECT_IPV4)
|
|
|
|
hints.ai_family = AF_INET;
|
|
|
|
else if (flags & CONNECT_IPV6)
|
|
|
|
hints.ai_family = AF_INET6;
|
2005-07-21 21:10:36 +08:00
|
|
|
hints.ai_socktype = SOCK_STREAM;
|
|
|
|
hints.ai_protocol = IPPROTO_TCP;
|
|
|
|
|
2007-05-17 01:09:41 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "Looking up %s ... ", host);
|
|
|
|
|
2005-07-21 21:10:36 +08:00
|
|
|
gai = getaddrinfo(host, port, &hints, &ai);
|
|
|
|
if (gai)
|
Fix "getaddrinfo()" buglet
At least in Linux glibc, "getaddrinfo()" has a very irritating feature (or
bug, who knows..).
Namely if you pass it in an empty string for the service name, it will
happily and quietly consider it identical to a NULL port pointer, and
return port number zero and no errors. Which obviously will not work.
Maybe that's what it's really expected to do, although the man-page for
getaddrinfo() certainly implies that it's a bug.
So when somebody passes me a "please pull" request pointing to something
like the following
git://git.kernel.org:/pub/scm/linux/kernel/git/mchehab/v4l-dvb.git
(note the extraneous colon at the end of the host name), git would happily
try to connect to port 0, which would generally just cause the remote to
not even answer, and the "connect()" will take a long time to time out.
So to work around the glibc feature/bug, just notice this empty port case
automatically. Also, add the port information to the error information
when it fails to look up (maybe it's the host-name that fails, maybe it's
the port-name - we should print out both).
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-03-28 00:50:20 +08:00
|
|
|
die("Unable to look up %s (port %s) (%s)", host, port, gai_strerror(gai));
|
2005-07-21 21:10:36 +08:00
|
|
|
|
2007-05-17 01:09:41 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "done.\nConnecting to %s (port %s) ... ", host, port);
|
|
|
|
|
2011-08-01 19:16:09 +08:00
|
|
|
for (ai0 = ai; ai; ai = ai->ai_next, cnt++) {
|
2006-06-07 11:58:41 +08:00
|
|
|
sockfd = socket(ai->ai_family,
|
|
|
|
ai->ai_socktype, ai->ai_protocol);
|
2011-07-13 00:28:34 +08:00
|
|
|
if ((sockfd < 0) ||
|
|
|
|
(connect(sockfd, ai->ai_addr, ai->ai_addrlen) < 0)) {
|
|
|
|
strbuf_addf(&error_message, "%s[%d: %s]: errno=%s\n",
|
|
|
|
host, cnt, ai_name(ai), strerror(errno));
|
|
|
|
if (0 <= sockfd)
|
|
|
|
close(sockfd);
|
2005-07-21 21:10:36 +08:00
|
|
|
sockfd = -1;
|
|
|
|
continue;
|
2005-07-14 09:46:20 +08:00
|
|
|
}
|
2007-05-24 05:34:27 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "%s ", ai_name(ai));
|
2005-07-21 21:10:36 +08:00
|
|
|
break;
|
2005-07-14 09:46:20 +08:00
|
|
|
}
|
|
|
|
|
2005-07-21 21:10:36 +08:00
|
|
|
freeaddrinfo(ai0);
|
2005-07-14 09:46:20 +08:00
|
|
|
|
|
|
|
if (sockfd < 0)
|
2011-07-13 00:28:34 +08:00
|
|
|
die("unable to connect to %s:\n%s", host, error_message.buf);
|
2005-07-21 21:10:36 +08:00
|
|
|
|
2011-12-06 12:39:36 +08:00
|
|
|
enable_keepalive(sockfd);
|
|
|
|
|
2007-05-17 01:09:41 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "done.\n");
|
|
|
|
|
2011-07-13 00:28:34 +08:00
|
|
|
strbuf_release(&error_message);
|
|
|
|
|
2006-06-07 11:58:41 +08:00
|
|
|
return sockfd;
|
2005-07-14 09:46:20 +08:00
|
|
|
}
|
|
|
|
|
2005-09-29 07:52:21 +08:00
|
|
|
#else /* NO_IPV6 */
|
2005-09-29 07:37:58 +08:00
|
|
|
|
2006-06-07 11:58:41 +08:00
|
|
|
/*
|
|
|
|
* Returns a connected socket() fd, or else die()s.
|
|
|
|
*/
|
2007-05-17 01:09:41 +08:00
|
|
|
static int git_tcp_connect_sock(char *host, int flags)
|
2005-09-29 07:37:58 +08:00
|
|
|
{
|
2011-08-01 19:16:10 +08:00
|
|
|
struct strbuf error_message = STRBUF_INIT;
|
|
|
|
int sockfd = -1;
|
2010-02-18 04:56:02 +08:00
|
|
|
const char *port = STR(DEFAULT_GIT_PORT);
|
|
|
|
char *ep;
|
2005-09-29 07:37:58 +08:00
|
|
|
struct hostent *he;
|
|
|
|
struct sockaddr_in sa;
|
|
|
|
char **ap;
|
|
|
|
unsigned int nport;
|
2007-05-24 05:34:27 +08:00
|
|
|
int cnt;
|
2005-09-29 07:37:58 +08:00
|
|
|
|
2010-02-18 04:56:02 +08:00
|
|
|
get_host_and_port(&host, &port);
|
2005-09-29 07:37:58 +08:00
|
|
|
|
2007-05-17 01:09:41 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "Looking up %s ... ", host);
|
|
|
|
|
2005-09-29 07:37:58 +08:00
|
|
|
he = gethostbyname(host);
|
|
|
|
if (!he)
|
|
|
|
die("Unable to look up %s (%s)", host, hstrerror(h_errno));
|
|
|
|
nport = strtoul(port, &ep, 10);
|
|
|
|
if ( ep == port || *ep ) {
|
|
|
|
/* Not numeric */
|
|
|
|
struct servent *se = getservbyname(port,"tcp");
|
|
|
|
if ( !se )
|
2009-01-05 02:38:41 +08:00
|
|
|
die("Unknown port %s", port);
|
2005-09-29 07:37:58 +08:00
|
|
|
nport = se->s_port;
|
|
|
|
}
|
|
|
|
|
2007-05-17 01:09:41 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "done.\nConnecting to %s (port %s) ... ", host, port);
|
|
|
|
|
2007-05-24 05:34:27 +08:00
|
|
|
for (cnt = 0, ap = he->h_addr_list; *ap; ap++, cnt++) {
|
2005-09-29 07:37:58 +08:00
|
|
|
memset(&sa, 0, sizeof sa);
|
|
|
|
sa.sin_family = he->h_addrtype;
|
2005-09-29 08:26:44 +08:00
|
|
|
sa.sin_port = htons(nport);
|
2005-11-22 21:54:23 +08:00
|
|
|
memcpy(&sa.sin_addr, *ap, he->h_length);
|
2005-09-29 07:37:58 +08:00
|
|
|
|
2011-08-01 19:16:10 +08:00
|
|
|
sockfd = socket(he->h_addrtype, SOCK_STREAM, 0);
|
|
|
|
if ((sockfd < 0) ||
|
|
|
|
connect(sockfd, (struct sockaddr *)&sa, sizeof sa) < 0) {
|
|
|
|
strbuf_addf(&error_message, "%s[%d: %s]: errno=%s\n",
|
2007-05-24 05:34:27 +08:00
|
|
|
host,
|
|
|
|
cnt,
|
|
|
|
inet_ntoa(*(struct in_addr *)&sa.sin_addr),
|
2011-08-01 19:16:10 +08:00
|
|
|
strerror(errno));
|
|
|
|
if (0 <= sockfd)
|
|
|
|
close(sockfd);
|
2005-09-29 07:37:58 +08:00
|
|
|
sockfd = -1;
|
|
|
|
continue;
|
|
|
|
}
|
2007-05-24 05:34:27 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "%s ",
|
|
|
|
inet_ntoa(*(struct in_addr *)&sa.sin_addr));
|
2005-09-29 07:37:58 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (sockfd < 0)
|
2011-08-01 19:16:10 +08:00
|
|
|
die("unable to connect to %s:\n%s", host, error_message.buf);
|
2005-09-29 07:37:58 +08:00
|
|
|
|
2011-12-06 12:39:36 +08:00
|
|
|
enable_keepalive(sockfd);
|
|
|
|
|
2007-05-17 01:09:41 +08:00
|
|
|
if (flags & CONNECT_VERBOSE)
|
|
|
|
fprintf(stderr, "done.\n");
|
|
|
|
|
2006-06-07 11:58:41 +08:00
|
|
|
return sockfd;
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* NO_IPV6 */
|
|
|
|
|
|
|
|
|
2017-11-21 05:22:54 +08:00
|
|
|
/*
|
|
|
|
* Dummy child_process returned by git_connect() if the transport protocol
|
|
|
|
* does not need fork(2).
|
|
|
|
*/
|
|
|
|
static struct child_process no_fork = CHILD_PROCESS_INIT;
|
|
|
|
|
|
|
|
int git_connection_is_socket(struct child_process *conn)
|
|
|
|
{
|
|
|
|
return conn == &no_fork;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct child_process *git_tcp_connect(int fd[2], char *host, int flags)
|
2006-06-07 11:58:41 +08:00
|
|
|
{
|
2007-05-17 01:09:41 +08:00
|
|
|
int sockfd = git_tcp_connect_sock(host, flags);
|
2006-06-07 11:58:41 +08:00
|
|
|
|
2005-09-29 07:37:58 +08:00
|
|
|
fd[0] = sockfd;
|
2007-01-22 09:10:51 +08:00
|
|
|
fd[1] = dup(sockfd);
|
2017-11-21 05:22:54 +08:00
|
|
|
|
|
|
|
return &no_fork;
|
2005-09-29 07:37:58 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2006-08-16 01:23:48 +08:00
|
|
|
static char *git_proxy_command;
|
2005-11-04 22:57:16 +08:00
|
|
|
|
2008-05-15 01:46:53 +08:00
|
|
|
static int git_proxy_command_options(const char *var, const char *value,
|
|
|
|
void *cb)
|
2005-11-04 22:57:16 +08:00
|
|
|
{
|
2005-11-19 19:48:56 +08:00
|
|
|
if (!strcmp(var, "core.gitproxy")) {
|
2005-11-22 11:18:23 +08:00
|
|
|
const char *for_pos;
|
|
|
|
int matchlen = -1;
|
|
|
|
int hostlen;
|
2009-03-11 10:38:12 +08:00
|
|
|
const char *rhost_name = cb;
|
|
|
|
int rhost_len = strlen(rhost_name);
|
2005-11-22 11:18:23 +08:00
|
|
|
|
2005-11-19 19:48:56 +08:00
|
|
|
if (git_proxy_command)
|
2005-11-04 22:57:16 +08:00
|
|
|
return 0;
|
2008-02-12 02:52:15 +08:00
|
|
|
if (!value)
|
|
|
|
return config_error_nonbool(var);
|
2005-11-19 19:48:56 +08:00
|
|
|
/* [core]
|
|
|
|
* ;# matches www.kernel.org as well
|
|
|
|
* gitproxy = netcatter-1 for kernel.org
|
|
|
|
* gitproxy = netcatter-2 for sample.xz
|
|
|
|
* gitproxy = netcatter-default
|
|
|
|
*/
|
2005-11-22 11:18:23 +08:00
|
|
|
for_pos = strstr(value, " for ");
|
2005-11-19 19:48:56 +08:00
|
|
|
if (!for_pos)
|
|
|
|
/* matches everybody */
|
|
|
|
matchlen = strlen(value);
|
|
|
|
else {
|
|
|
|
hostlen = strlen(for_pos + 5);
|
|
|
|
if (rhost_len < hostlen)
|
|
|
|
matchlen = -1;
|
|
|
|
else if (!strncmp(for_pos + 5,
|
|
|
|
rhost_name + rhost_len - hostlen,
|
|
|
|
hostlen) &&
|
|
|
|
((rhost_len == hostlen) ||
|
|
|
|
rhost_name[rhost_len - hostlen -1] == '.'))
|
|
|
|
matchlen = for_pos - value;
|
|
|
|
else
|
|
|
|
matchlen = -1;
|
|
|
|
}
|
|
|
|
if (0 <= matchlen) {
|
|
|
|
/* core.gitproxy = none for kernel.org */
|
2007-06-07 15:04:01 +08:00
|
|
|
if (matchlen == 4 &&
|
2005-11-19 19:48:56 +08:00
|
|
|
!memcmp(value, "none", 4))
|
|
|
|
matchlen = 0;
|
2007-09-16 06:32:36 +08:00
|
|
|
git_proxy_command = xmemdupz(value, matchlen);
|
2005-11-04 22:57:16 +08:00
|
|
|
}
|
2005-11-19 19:48:56 +08:00
|
|
|
return 0;
|
2005-11-04 22:57:16 +08:00
|
|
|
}
|
|
|
|
|
2008-05-15 01:46:53 +08:00
|
|
|
return git_default_config(var, value, cb);
|
2005-11-04 22:57:16 +08:00
|
|
|
}
|
|
|
|
|
2005-11-19 19:48:56 +08:00
|
|
|
static int git_use_proxy(const char *host)
|
2005-11-04 22:57:16 +08:00
|
|
|
{
|
|
|
|
git_proxy_command = getenv("GIT_PROXY_COMMAND");
|
2009-03-11 10:38:12 +08:00
|
|
|
git_config(git_proxy_command_options, (void*)host);
|
2005-11-19 19:48:56 +08:00
|
|
|
return (git_proxy_command && *git_proxy_command);
|
2005-11-04 22:57:16 +08:00
|
|
|
}
|
|
|
|
|
connect: treat generic proxy processes like ssh processes
The git_connect function returns two ends of a pipe for
talking with a remote, plus a struct child_process
representing the other end of the pipe. If we have a direct
socket connection, then this points to a special "no_fork"
child process.
The code path for doing git-over-pipes or git-over-ssh sets
up this child process to point to the child git command or
the ssh process. When we call finish_connect eventually, we
check wait() on the command and report its return value.
The code path for git://, on the other hand, always sets it
to no_fork. In the case of a direct TCP connection, this
makes sense; we have no child process. But in the case of a
proxy command (configured by core.gitproxy), we do have a
child process, but we throw away its pid, and therefore
ignore its return code.
Instead, let's keep that information in the proxy case, and
respect its return code, which can help catch some errors
(though depending on your proxy command, it will be errors
reported by the proxy command itself, and not propagated
from git commands. Still, it is probably better to propagate
such errors than to ignore them).
It also means that the child_process field can reliably be
used to determine whether the returned descriptors are
actually a full-duplex socket, which means we should be
using shutdown() instead of a simple close.
Signed-off-by: Jeff King <peff@peff.net>
Helped-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-05-16 14:46:07 +08:00
|
|
|
static struct child_process *git_proxy_connect(int fd[2], char *host)
|
2005-11-04 22:57:16 +08:00
|
|
|
{
|
2006-06-28 17:04:39 +08:00
|
|
|
const char *port = STR(DEFAULT_GIT_PORT);
|
connect: treat generic proxy processes like ssh processes
The git_connect function returns two ends of a pipe for
talking with a remote, plus a struct child_process
representing the other end of the pipe. If we have a direct
socket connection, then this points to a special "no_fork"
child process.
The code path for doing git-over-pipes or git-over-ssh sets
up this child process to point to the child git command or
the ssh process. When we call finish_connect eventually, we
check wait() on the command and report its return value.
The code path for git://, on the other hand, always sets it
to no_fork. In the case of a direct TCP connection, this
makes sense; we have no child process. But in the case of a
proxy command (configured by core.gitproxy), we do have a
child process, but we throw away its pid, and therefore
ignore its return code.
Instead, let's keep that information in the proxy case, and
respect its return code, which can help catch some errors
(though depending on your proxy command, it will be errors
reported by the proxy command itself, and not propagated
from git commands. Still, it is probably better to propagate
such errors than to ignore them).
It also means that the child_process field can reliably be
used to determine whether the returned descriptors are
actually a full-duplex socket, which means we should be
using shutdown() instead of a simple close.
Signed-off-by: Jeff King <peff@peff.net>
Helped-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-05-16 14:46:07 +08:00
|
|
|
struct child_process *proxy;
|
2005-11-04 22:57:16 +08:00
|
|
|
|
2010-02-18 04:56:02 +08:00
|
|
|
get_host_and_port(&host, &port);
|
2005-11-04 22:57:16 +08:00
|
|
|
|
2017-07-29 03:26:50 +08:00
|
|
|
if (looks_like_command_line_option(host))
|
|
|
|
die("strange hostname '%s' blocked", host);
|
|
|
|
if (looks_like_command_line_option(port))
|
|
|
|
die("strange port '%s' blocked", port);
|
|
|
|
|
2014-08-20 03:10:48 +08:00
|
|
|
proxy = xmalloc(sizeof(*proxy));
|
|
|
|
child_process_init(proxy);
|
git_connect: use argv_array
This avoids magic numbers when we allocate fixed-size argv
arrays, and makes it more obvious that we are not
overflowing.
It is also the first step to fixing a memory leak. When
git_connect returns a child_process struct, the argv array
in the struct is dynamically allocated, but the individual
strings are not (they are either owned elsewhere, or are
freed). Later, in finish_connect, we free the array but
leave the strings alone.
This works for the child_process created by git_connect, but
if we use transport_take_over, we may also end up with a
child_process created by transport-helper's get_helper.
In that case, the strings are freshly allocated, and we
would want to free them. However, we have no idea in
finish_connect which type we have.
By consistently using run-command's internal argv-array, we
do not have to worry about this issue at all; finish_command
takes care of it for us, and we can drop our manual free
entirely.
Note that this actually makes the get_helper leak slightly
worse; now we are leaking both the strings and the array.
But when we adjust it in a future patch, that leak will go
away entirely.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-05-15 16:34:09 +08:00
|
|
|
argv_array_push(&proxy->args, git_proxy_command);
|
|
|
|
argv_array_push(&proxy->args, host);
|
|
|
|
argv_array_push(&proxy->args, port);
|
connect: treat generic proxy processes like ssh processes
The git_connect function returns two ends of a pipe for
talking with a remote, plus a struct child_process
representing the other end of the pipe. If we have a direct
socket connection, then this points to a special "no_fork"
child process.
The code path for doing git-over-pipes or git-over-ssh sets
up this child process to point to the child git command or
the ssh process. When we call finish_connect eventually, we
check wait() on the command and report its return value.
The code path for git://, on the other hand, always sets it
to no_fork. In the case of a direct TCP connection, this
makes sense; we have no child process. But in the case of a
proxy command (configured by core.gitproxy), we do have a
child process, but we throw away its pid, and therefore
ignore its return code.
Instead, let's keep that information in the proxy case, and
respect its return code, which can help catch some errors
(though depending on your proxy command, it will be errors
reported by the proxy command itself, and not propagated
from git commands. Still, it is probably better to propagate
such errors than to ignore them).
It also means that the child_process field can reliably be
used to determine whether the returned descriptors are
actually a full-duplex socket, which means we should be
using shutdown() instead of a simple close.
Signed-off-by: Jeff King <peff@peff.net>
Helped-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-05-16 14:46:07 +08:00
|
|
|
proxy->in = -1;
|
|
|
|
proxy->out = -1;
|
|
|
|
if (start_command(proxy))
|
git_connect: use argv_array
This avoids magic numbers when we allocate fixed-size argv
arrays, and makes it more obvious that we are not
overflowing.
It is also the first step to fixing a memory leak. When
git_connect returns a child_process struct, the argv array
in the struct is dynamically allocated, but the individual
strings are not (they are either owned elsewhere, or are
freed). Later, in finish_connect, we free the array but
leave the strings alone.
This works for the child_process created by git_connect, but
if we use transport_take_over, we may also end up with a
child_process created by transport-helper's get_helper.
In that case, the strings are freshly allocated, and we
would want to free them. However, we have no idea in
finish_connect which type we have.
By consistently using run-command's internal argv-array, we
do not have to worry about this issue at all; finish_command
takes care of it for us, and we can drop our manual free
entirely.
Note that this actually makes the get_helper leak slightly
worse; now we are leaking both the strings and the array.
But when we adjust it in a future patch, that leak will go
away entirely.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-05-15 16:34:09 +08:00
|
|
|
die("cannot start proxy %s", git_proxy_command);
|
connect: treat generic proxy processes like ssh processes
The git_connect function returns two ends of a pipe for
talking with a remote, plus a struct child_process
representing the other end of the pipe. If we have a direct
socket connection, then this points to a special "no_fork"
child process.
The code path for doing git-over-pipes or git-over-ssh sets
up this child process to point to the child git command or
the ssh process. When we call finish_connect eventually, we
check wait() on the command and report its return value.
The code path for git://, on the other hand, always sets it
to no_fork. In the case of a direct TCP connection, this
makes sense; we have no child process. But in the case of a
proxy command (configured by core.gitproxy), we do have a
child process, but we throw away its pid, and therefore
ignore its return code.
Instead, let's keep that information in the proxy case, and
respect its return code, which can help catch some errors
(though depending on your proxy command, it will be errors
reported by the proxy command itself, and not propagated
from git commands. Still, it is probably better to propagate
such errors than to ignore them).
It also means that the child_process field can reliably be
used to determine whether the returned descriptors are
actually a full-duplex socket, which means we should be
using shutdown() instead of a simple close.
Signed-off-by: Jeff King <peff@peff.net>
Helped-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-05-16 14:46:07 +08:00
|
|
|
fd[0] = proxy->out; /* read from proxy stdout */
|
|
|
|
fd[1] = proxy->in; /* write to proxy stdin */
|
|
|
|
return proxy;
|
2005-11-04 22:57:16 +08:00
|
|
|
}
|
|
|
|
|
2015-02-21 23:52:48 +08:00
|
|
|
static char *get_port(char *host)
|
2007-09-01 17:36:31 +08:00
|
|
|
{
|
|
|
|
char *end;
|
2015-02-21 23:52:48 +08:00
|
|
|
char *p = strchr(host, ':');
|
|
|
|
|
2007-09-01 17:36:31 +08:00
|
|
|
if (p) {
|
2008-12-21 09:12:11 +08:00
|
|
|
long port = strtol(p + 1, &end, 10);
|
|
|
|
if (end != p + 1 && *end == '\0' && 0 <= port && port < 65536) {
|
2015-02-21 23:52:48 +08:00
|
|
|
*p = '\0';
|
|
|
|
return p+1;
|
2007-09-01 17:36:31 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-07-05 02:57:58 +08:00
|
|
|
/*
|
2013-11-29 03:49:01 +08:00
|
|
|
* Extract protocol and relevant parts from the specified connection URL.
|
|
|
|
* The caller must free() the returned strings.
|
2005-07-05 02:57:58 +08:00
|
|
|
*/
|
2013-11-29 03:49:01 +08:00
|
|
|
static enum protocol parse_connect_url(const char *url_orig, char **ret_host,
|
2013-11-29 03:49:54 +08:00
|
|
|
char **ret_path)
|
2005-07-05 02:57:58 +08:00
|
|
|
{
|
2010-05-23 17:19:44 +08:00
|
|
|
char *url;
|
2009-03-13 20:51:33 +08:00
|
|
|
char *host, *path;
|
2005-12-21 18:23:42 +08:00
|
|
|
char *end;
|
2013-11-29 03:50:03 +08:00
|
|
|
int separator = '/';
|
2005-11-18 03:37:14 +08:00
|
|
|
enum protocol protocol = PROTO_LOCAL;
|
2006-06-20 09:25:21 +08:00
|
|
|
|
2010-05-23 17:19:44 +08:00
|
|
|
if (is_url(url_orig))
|
|
|
|
url = url_decode(url_orig);
|
|
|
|
else
|
|
|
|
url = xstrdup(url_orig);
|
|
|
|
|
2005-11-18 03:37:14 +08:00
|
|
|
host = strstr(url, "://");
|
2009-09-01 13:35:10 +08:00
|
|
|
if (host) {
|
2005-11-18 03:37:14 +08:00
|
|
|
*host = '\0';
|
|
|
|
protocol = get_protocol(url);
|
|
|
|
host += 3;
|
2005-12-21 18:23:42 +08:00
|
|
|
} else {
|
2005-07-05 02:57:58 +08:00
|
|
|
host = url;
|
2013-11-29 03:50:03 +08:00
|
|
|
if (!url_is_local_not_ssh(url)) {
|
|
|
|
protocol = PROTO_SSH;
|
|
|
|
separator = ':';
|
|
|
|
}
|
2005-12-21 18:23:42 +08:00
|
|
|
}
|
|
|
|
|
2010-01-27 02:24:42 +08:00
|
|
|
/*
|
2013-11-29 03:49:54 +08:00
|
|
|
* Don't do destructive transforms as protocol code does
|
|
|
|
* '[]' unwrapping in get_host_and_port()
|
2010-01-27 02:24:42 +08:00
|
|
|
*/
|
2015-02-21 23:52:48 +08:00
|
|
|
end = host_end(&host, 0);
|
2005-12-21 18:23:42 +08:00
|
|
|
|
2013-11-29 03:50:03 +08:00
|
|
|
if (protocol == PROTO_LOCAL)
|
2007-08-02 01:03:37 +08:00
|
|
|
path = end;
|
2013-11-29 03:50:03 +08:00
|
|
|
else if (protocol == PROTO_FILE && has_dos_drive_prefix(end))
|
|
|
|
path = end; /* "file://$(pwd)" may be "file://C:/projects/repo" */
|
|
|
|
else
|
|
|
|
path = strchr(end, separator);
|
2005-07-14 09:46:20 +08:00
|
|
|
|
2005-11-18 03:37:14 +08:00
|
|
|
if (!path || !*path)
|
|
|
|
die("No path specified. See 'man git-pull' for valid url syntax");
|
|
|
|
|
|
|
|
/*
|
|
|
|
* null-terminate hostname and point path to ~ for URL's like this:
|
|
|
|
* ssh://host.xz/~user/repo
|
|
|
|
*/
|
2013-11-29 03:50:03 +08:00
|
|
|
|
|
|
|
end = path; /* Need to \0 terminate host here */
|
|
|
|
if (separator == ':')
|
|
|
|
path++; /* path starts after ':' */
|
|
|
|
if (protocol == PROTO_GIT || protocol == PROTO_SSH) {
|
2005-11-18 03:37:14 +08:00
|
|
|
if (path[1] == '~')
|
|
|
|
path++;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:50:03 +08:00
|
|
|
path = xstrdup(path);
|
|
|
|
*end = '\0';
|
|
|
|
|
2013-11-29 03:49:01 +08:00
|
|
|
*ret_host = xstrdup(host);
|
2013-11-29 03:50:03 +08:00
|
|
|
*ret_path = path;
|
2013-11-29 03:49:01 +08:00
|
|
|
free(url);
|
|
|
|
return protocol;
|
|
|
|
}
|
|
|
|
|
2016-06-26 19:16:35 +08:00
|
|
|
static const char *get_ssh_command(void)
|
|
|
|
{
|
|
|
|
const char *ssh;
|
|
|
|
|
|
|
|
if ((ssh = getenv("GIT_SSH_COMMAND")))
|
|
|
|
return ssh;
|
|
|
|
|
|
|
|
if (!git_config_get_string_const("core.sshcommand", &ssh))
|
|
|
|
return ssh;
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
enum ssh_variant {
|
|
|
|
VARIANT_SIMPLE,
|
|
|
|
VARIANT_SSH,
|
|
|
|
VARIANT_PLINK,
|
|
|
|
VARIANT_PUTTY,
|
|
|
|
VARIANT_TORTOISEPLINK,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int override_ssh_variant(enum ssh_variant *ssh_variant)
|
2017-02-01 20:01:10 +08:00
|
|
|
{
|
2017-10-17 01:55:31 +08:00
|
|
|
const char *variant = getenv("GIT_SSH_VARIANT");
|
2017-02-10 01:20:25 +08:00
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
if (!variant && git_config_get_string_const("ssh.variant", &variant))
|
2017-02-10 01:20:25 +08:00
|
|
|
return 0;
|
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
if (!strcmp(variant, "plink"))
|
|
|
|
*ssh_variant = VARIANT_PLINK;
|
|
|
|
else if (!strcmp(variant, "putty"))
|
|
|
|
*ssh_variant = VARIANT_PUTTY;
|
|
|
|
else if (!strcmp(variant, "tortoiseplink"))
|
|
|
|
*ssh_variant = VARIANT_TORTOISEPLINK;
|
|
|
|
else if (!strcmp(variant, "simple"))
|
|
|
|
*ssh_variant = VARIANT_SIMPLE;
|
|
|
|
else
|
|
|
|
*ssh_variant = VARIANT_SSH;
|
|
|
|
|
2017-02-10 01:20:25 +08:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
static enum ssh_variant determine_ssh_variant(const char *ssh_command,
|
|
|
|
int is_cmdline)
|
2017-02-10 01:20:25 +08:00
|
|
|
{
|
2017-10-17 01:55:31 +08:00
|
|
|
enum ssh_variant ssh_variant = VARIANT_SIMPLE;
|
2017-02-10 01:20:25 +08:00
|
|
|
const char *variant;
|
2017-02-01 20:01:10 +08:00
|
|
|
char *p = NULL;
|
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
if (override_ssh_variant(&ssh_variant))
|
|
|
|
return ssh_variant;
|
2017-02-10 01:20:25 +08:00
|
|
|
|
|
|
|
if (!is_cmdline) {
|
2017-02-01 20:01:10 +08:00
|
|
|
p = xstrdup(ssh_command);
|
|
|
|
variant = basename(p);
|
|
|
|
} else {
|
|
|
|
const char **ssh_argv;
|
|
|
|
|
|
|
|
p = xstrdup(ssh_command);
|
2017-04-11 08:30:23 +08:00
|
|
|
if (split_cmdline(p, &ssh_argv) > 0) {
|
2017-02-01 20:01:10 +08:00
|
|
|
variant = basename((char *)ssh_argv[0]);
|
|
|
|
/*
|
|
|
|
* At this point, variant points into the buffer
|
|
|
|
* referenced by p, hence we do not need ssh_argv
|
|
|
|
* any longer.
|
|
|
|
*/
|
|
|
|
free(ssh_argv);
|
2017-04-21 04:21:58 +08:00
|
|
|
} else {
|
|
|
|
free(p);
|
2017-10-17 01:55:31 +08:00
|
|
|
return ssh_variant;
|
2017-04-21 04:21:58 +08:00
|
|
|
}
|
2017-02-01 20:01:10 +08:00
|
|
|
}
|
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
if (!strcasecmp(variant, "ssh") ||
|
|
|
|
!strcasecmp(variant, "ssh.exe"))
|
|
|
|
ssh_variant = VARIANT_SSH;
|
|
|
|
else if (!strcasecmp(variant, "plink") ||
|
|
|
|
!strcasecmp(variant, "plink.exe"))
|
|
|
|
ssh_variant = VARIANT_PLINK;
|
2017-02-01 20:01:10 +08:00
|
|
|
else if (!strcasecmp(variant, "tortoiseplink") ||
|
2017-10-17 01:55:31 +08:00
|
|
|
!strcasecmp(variant, "tortoiseplink.exe"))
|
|
|
|
ssh_variant = VARIANT_TORTOISEPLINK;
|
|
|
|
|
2017-02-01 20:01:10 +08:00
|
|
|
free(p);
|
2017-10-17 01:55:31 +08:00
|
|
|
return ssh_variant;
|
2017-02-01 20:01:10 +08:00
|
|
|
}
|
|
|
|
|
2017-11-21 05:23:27 +08:00
|
|
|
/*
|
|
|
|
* Open a connection using Git's native protocol.
|
|
|
|
*
|
|
|
|
* The caller is responsible for freeing hostandport, but this function may
|
|
|
|
* modify it (for example, to truncate it to remove the port part).
|
|
|
|
*/
|
|
|
|
static struct child_process *git_connect_git(int fd[2], char *hostandport,
|
|
|
|
const char *path, const char *prog,
|
|
|
|
int flags)
|
|
|
|
{
|
|
|
|
struct child_process *conn;
|
|
|
|
struct strbuf request = STRBUF_INIT;
|
|
|
|
/*
|
|
|
|
* Set up virtual host information based on where we will
|
|
|
|
* connect, unless the user has overridden us in
|
|
|
|
* the environment.
|
|
|
|
*/
|
|
|
|
char *target_host = getenv("GIT_OVERRIDE_VIRTUAL_HOST");
|
|
|
|
if (target_host)
|
|
|
|
target_host = xstrdup(target_host);
|
|
|
|
else
|
|
|
|
target_host = xstrdup(hostandport);
|
|
|
|
|
|
|
|
transport_check_allowed("git");
|
|
|
|
|
|
|
|
/* These underlying connection commands die() if they
|
|
|
|
* cannot connect.
|
|
|
|
*/
|
|
|
|
if (git_use_proxy(hostandport))
|
|
|
|
conn = git_proxy_connect(fd, hostandport);
|
|
|
|
else
|
|
|
|
conn = git_tcp_connect(fd, hostandport, flags);
|
|
|
|
/*
|
|
|
|
* Separate original protocol components prog and path
|
|
|
|
* from extended host header with a NUL byte.
|
|
|
|
*
|
|
|
|
* Note: Do not add any other headers here! Doing so
|
|
|
|
* will cause older git-daemon servers to crash.
|
|
|
|
*/
|
|
|
|
strbuf_addf(&request,
|
|
|
|
"%s %s%chost=%s%c",
|
|
|
|
prog, path, 0,
|
|
|
|
target_host, 0);
|
|
|
|
|
|
|
|
/* If using a new version put that stuff here after a second null byte */
|
|
|
|
if (get_protocol_version_config() > 0) {
|
|
|
|
strbuf_addch(&request, '\0');
|
|
|
|
strbuf_addf(&request, "version=%d%c",
|
|
|
|
get_protocol_version_config(), '\0');
|
|
|
|
}
|
|
|
|
|
|
|
|
packet_write(fd[1], request.buf, request.len);
|
|
|
|
|
|
|
|
free(target_host);
|
|
|
|
strbuf_release(&request);
|
|
|
|
return conn;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:49:01 +08:00
|
|
|
/*
|
2017-11-21 05:22:54 +08:00
|
|
|
* This returns the dummy child_process `no_fork` if the transport protocol
|
|
|
|
* does not need fork(2), or a struct child_process object if it does. Once
|
|
|
|
* done, finish the connection with finish_connect() with the value returned
|
|
|
|
* from this function (it is safe to call finish_connect() with NULL to
|
|
|
|
* support the former case).
|
2013-11-29 03:49:01 +08:00
|
|
|
*
|
|
|
|
* If it returns, the connect is successful; it just dies on errors (this
|
|
|
|
* will hopefully be changed in a libification effort, to return NULL when
|
|
|
|
* the connection failed).
|
|
|
|
*/
|
|
|
|
struct child_process *git_connect(int fd[2], const char *url,
|
|
|
|
const char *prog, int flags)
|
|
|
|
{
|
2013-11-29 03:50:15 +08:00
|
|
|
char *hostandport, *path;
|
2017-11-21 05:22:54 +08:00
|
|
|
struct child_process *conn;
|
2013-11-29 03:49:01 +08:00
|
|
|
enum protocol protocol;
|
|
|
|
|
|
|
|
/* Without this we cannot rely on waitpid() to tell
|
|
|
|
* what happened to our children.
|
2007-09-01 17:36:31 +08:00
|
|
|
*/
|
2013-11-29 03:49:01 +08:00
|
|
|
signal(SIGCHLD, SIG_DFL);
|
2007-09-01 17:36:31 +08:00
|
|
|
|
2013-11-29 03:50:15 +08:00
|
|
|
protocol = parse_connect_url(url, &hostandport, &path);
|
2015-02-21 23:52:55 +08:00
|
|
|
if ((flags & CONNECT_DIAG_URL) && (protocol != PROTO_SSH)) {
|
2013-11-29 03:49:17 +08:00
|
|
|
printf("Diag: url=%s\n", url ? url : "NULL");
|
|
|
|
printf("Diag: protocol=%s\n", prot_name(protocol));
|
2013-11-29 03:50:15 +08:00
|
|
|
printf("Diag: hostandport=%s\n", hostandport ? hostandport : "NULL");
|
2013-11-29 03:49:17 +08:00
|
|
|
printf("Diag: path=%s\n", path ? path : "NULL");
|
2013-11-29 03:50:15 +08:00
|
|
|
conn = NULL;
|
|
|
|
} else if (protocol == PROTO_GIT) {
|
2017-11-21 05:23:27 +08:00
|
|
|
conn = git_connect_git(fd, hostandport, path, prog, flags);
|
2013-11-29 03:50:15 +08:00
|
|
|
} else {
|
2017-08-31 01:49:39 +08:00
|
|
|
struct strbuf cmd = STRBUF_INIT;
|
2017-10-17 01:55:28 +08:00
|
|
|
const char *const *var;
|
2017-08-31 01:49:39 +08:00
|
|
|
|
2014-08-20 03:10:48 +08:00
|
|
|
conn = xmalloc(sizeof(*conn));
|
|
|
|
child_process_init(conn);
|
2013-11-29 03:50:15 +08:00
|
|
|
|
2017-07-29 03:28:55 +08:00
|
|
|
if (looks_like_command_line_option(path))
|
|
|
|
die("strange pathname '%s' blocked", path);
|
|
|
|
|
2013-11-29 03:50:15 +08:00
|
|
|
strbuf_addstr(&cmd, prog);
|
|
|
|
strbuf_addch(&cmd, ' ');
|
|
|
|
sq_quote_buf(&cmd, path);
|
|
|
|
|
2015-09-05 06:40:08 +08:00
|
|
|
/* remove repo-local variables from the environment */
|
2017-10-17 01:55:28 +08:00
|
|
|
for (var = local_repo_env; *var; var++)
|
|
|
|
argv_array_push(&conn->env_array, *var);
|
|
|
|
|
2015-09-08 16:33:14 +08:00
|
|
|
conn->use_shell = 1;
|
2013-11-29 03:50:15 +08:00
|
|
|
conn->in = conn->out = -1;
|
|
|
|
if (protocol == PROTO_SSH) {
|
2014-11-10 06:42:32 +08:00
|
|
|
const char *ssh;
|
2017-10-17 01:55:31 +08:00
|
|
|
enum ssh_variant variant;
|
2013-11-29 03:50:15 +08:00
|
|
|
char *ssh_host = hostandport;
|
|
|
|
const char *port = NULL;
|
transport: add a protocol-whitelist environment variable
If we are cloning an untrusted remote repository into a
sandbox, we may also want to fetch remote submodules in
order to get the complete view as intended by the other
side. However, that opens us up to attacks where a malicious
user gets us to clone something they would not otherwise
have access to (this is not necessarily a problem by itself,
but we may then act on the cloned contents in a way that
exposes them to the attacker).
Ideally such a setup would sandbox git entirely away from
high-value items, but this is not always practical or easy
to set up (e.g., OS network controls may block multiple
protocols, and we would want to enable some but not others).
We can help this case by providing a way to restrict
particular protocols. We use a whitelist in the environment.
This is more annoying to set up than a blacklist, but
defaults to safety if the set of protocols git supports
grows). If no whitelist is specified, we continue to default
to allowing all protocols (this is an "unsafe" default, but
since the minority of users will want this sandboxing
effect, it is the only sensible one).
A note on the tests: ideally these would all be in a single
test file, but the git-daemon and httpd test infrastructure
is an all-or-nothing proposition rather than a test-by-test
prerequisite. By putting them all together, we would be
unable to test the file-local code on machines without
apache.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-17 01:12:52 +08:00
|
|
|
transport_check_allowed("ssh");
|
2013-11-29 03:50:15 +08:00
|
|
|
get_host_and_port(&ssh_host, &port);
|
|
|
|
|
2015-02-21 23:52:48 +08:00
|
|
|
if (!port)
|
|
|
|
port = get_port(ssh_host);
|
2015-03-06 04:45:44 +08:00
|
|
|
|
2015-02-21 23:52:55 +08:00
|
|
|
if (flags & CONNECT_DIAG_URL) {
|
|
|
|
printf("Diag: url=%s\n", url ? url : "NULL");
|
|
|
|
printf("Diag: protocol=%s\n", prot_name(protocol));
|
|
|
|
printf("Diag: userandhost=%s\n", ssh_host ? ssh_host : "NULL");
|
|
|
|
printf("Diag: port=%s\n", port ? port : "NONE");
|
|
|
|
printf("Diag: path=%s\n", path ? path : "NULL");
|
2013-11-29 03:50:15 +08:00
|
|
|
|
2015-02-21 23:52:55 +08:00
|
|
|
free(hostandport);
|
|
|
|
free(path);
|
2015-03-10 00:58:22 +08:00
|
|
|
free(conn);
|
2017-08-31 01:49:39 +08:00
|
|
|
strbuf_release(&cmd);
|
2015-02-21 23:52:55 +08:00
|
|
|
return NULL;
|
2015-04-27 04:30:10 +08:00
|
|
|
}
|
|
|
|
|
2017-07-29 03:25:45 +08:00
|
|
|
if (looks_like_command_line_option(ssh_host))
|
connect: reject ssh hostname that begins with a dash
When commands like "git fetch" talk with ssh://$rest_of_URL/, the
code splits $rest_of_URL into components like host, port, etc., and
then spawns the underlying "ssh" program by formulating argv[] array
that has:
- the path to ssh command taken from GIT_SSH_COMMAND, etc.
- dashed options like '-batch' (for Tortoise), '-p <port>' as
needed.
- ssh_host, which is supposed to be the hostname parsed out of
$rest_of_URL.
- then the command to be run on the other side, e.g. git
upload-pack.
If the ssh_host ends up getting '-<anything>', the argv[] that is
used to spawn the command becomes something like:
{ "ssh", "-p", "22", "-<anything>", "command", "to", "run", NULL }
which obviously is bogus, but depending on the actual value of
"<anything>", will make "ssh" parse and use it as an option.
Prevent this by forbidding ssh_host that begins with a "-".
Noticed-by: Joern Schneeweisz of Recurity Labs
Reported-by: Brian at GitLab
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-07-27 01:24:20 +08:00
|
|
|
die("strange hostname '%s' blocked", ssh_host);
|
|
|
|
|
2016-06-26 19:16:35 +08:00
|
|
|
ssh = get_ssh_command();
|
2017-10-17 01:55:31 +08:00
|
|
|
if (ssh) {
|
|
|
|
variant = determine_ssh_variant(ssh, 1);
|
|
|
|
} else {
|
2015-09-08 16:33:14 +08:00
|
|
|
/*
|
|
|
|
* GIT_SSH is the no-shell version of
|
|
|
|
* GIT_SSH_COMMAND (and must remain so for
|
|
|
|
* historical compatibility).
|
|
|
|
*/
|
|
|
|
conn->use_shell = 0;
|
|
|
|
|
2015-04-27 04:30:10 +08:00
|
|
|
ssh = getenv("GIT_SSH");
|
|
|
|
if (!ssh)
|
|
|
|
ssh = "ssh";
|
2017-10-17 01:55:31 +08:00
|
|
|
variant = determine_ssh_variant(ssh, 0);
|
2015-04-27 04:30:10 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
argv_array_push(&conn->args, ssh);
|
2017-10-17 01:55:28 +08:00
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
if (variant == VARIANT_SSH &&
|
|
|
|
get_protocol_version_config() > 0) {
|
2017-10-17 01:55:28 +08:00
|
|
|
argv_array_push(&conn->args, "-o");
|
|
|
|
argv_array_push(&conn->args, "SendEnv=" GIT_PROTOCOL_ENVIRONMENT);
|
|
|
|
argv_array_pushf(&conn->env_array, GIT_PROTOCOL_ENVIRONMENT "=version=%d",
|
|
|
|
get_protocol_version_config());
|
|
|
|
}
|
|
|
|
|
2017-10-17 01:55:31 +08:00
|
|
|
if (variant != VARIANT_SIMPLE) {
|
|
|
|
if (flags & CONNECT_IPV4)
|
|
|
|
argv_array_push(&conn->args, "-4");
|
|
|
|
else if (flags & CONNECT_IPV6)
|
|
|
|
argv_array_push(&conn->args, "-6");
|
|
|
|
}
|
|
|
|
|
|
|
|
if (variant == VARIANT_TORTOISEPLINK)
|
2015-04-27 04:30:10 +08:00
|
|
|
argv_array_push(&conn->args, "-batch");
|
2017-10-17 01:55:31 +08:00
|
|
|
|
|
|
|
if (port && variant != VARIANT_SIMPLE) {
|
|
|
|
if (variant == VARIANT_SSH)
|
|
|
|
argv_array_push(&conn->args, "-p");
|
|
|
|
else
|
|
|
|
argv_array_push(&conn->args, "-P");
|
|
|
|
|
2015-04-27 04:30:10 +08:00
|
|
|
argv_array_push(&conn->args, port);
|
2013-11-29 03:50:15 +08:00
|
|
|
}
|
2017-10-17 01:55:31 +08:00
|
|
|
|
2015-04-27 04:30:10 +08:00
|
|
|
argv_array_push(&conn->args, ssh_host);
|
2014-03-13 19:45:31 +08:00
|
|
|
} else {
|
transport: add a protocol-whitelist environment variable
If we are cloning an untrusted remote repository into a
sandbox, we may also want to fetch remote submodules in
order to get the complete view as intended by the other
side. However, that opens us up to attacks where a malicious
user gets us to clone something they would not otherwise
have access to (this is not necessarily a problem by itself,
but we may then act on the cloned contents in a way that
exposes them to the attacker).
Ideally such a setup would sandbox git entirely away from
high-value items, but this is not always practical or easy
to set up (e.g., OS network controls may block multiple
protocols, and we would want to enable some but not others).
We can help this case by providing a way to restrict
particular protocols. We use a whitelist in the environment.
This is more annoying to set up than a blacklist, but
defaults to safety if the set of protocols git supports
grows). If no whitelist is specified, we continue to default
to allowing all protocols (this is an "unsafe" default, but
since the minority of users will want this sandboxing
effect, it is the only sensible one).
A note on the tests: ideally these would all be in a single
test file, but the git-daemon and httpd test infrastructure
is an all-or-nothing proposition rather than a test-by-test
prerequisite. By putting them all together, we would be
unable to test the file-local code on machines without
apache.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-09-17 01:12:52 +08:00
|
|
|
transport_check_allowed("file");
|
2017-10-17 01:55:28 +08:00
|
|
|
if (get_protocol_version_config() > 0) {
|
|
|
|
argv_array_pushf(&conn->env_array, GIT_PROTOCOL_ENVIRONMENT "=version=%d",
|
|
|
|
get_protocol_version_config());
|
|
|
|
}
|
2005-08-03 23:15:42 +08:00
|
|
|
}
|
git_connect: use argv_array
This avoids magic numbers when we allocate fixed-size argv
arrays, and makes it more obvious that we are not
overflowing.
It is also the first step to fixing a memory leak. When
git_connect returns a child_process struct, the argv array
in the struct is dynamically allocated, but the individual
strings are not (they are either owned elsewhere, or are
freed). Later, in finish_connect, we free the array but
leave the strings alone.
This works for the child_process created by git_connect, but
if we use transport_take_over, we may also end up with a
child_process created by transport-helper's get_helper.
In that case, the strings are freshly allocated, and we
would want to free them. However, we have no idea in
finish_connect which type we have.
By consistently using run-command's internal argv-array, we
do not have to worry about this issue at all; finish_command
takes care of it for us, and we can drop our manual free
entirely.
Note that this actually makes the get_helper leak slightly
worse; now we are leaking both the strings and the array.
But when we adjust it in a future patch, that leak will go
away entirely.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-05-15 16:34:09 +08:00
|
|
|
argv_array_push(&conn->args, cmd.buf);
|
2007-10-20 03:47:54 +08:00
|
|
|
|
2013-11-29 03:50:15 +08:00
|
|
|
if (start_command(conn))
|
|
|
|
die("unable to fork");
|
2007-10-20 03:47:54 +08:00
|
|
|
|
2013-11-29 03:50:15 +08:00
|
|
|
fd[0] = conn->out; /* read from child's stdout */
|
|
|
|
fd[1] = conn->in; /* write to child's stdin */
|
|
|
|
strbuf_release(&cmd);
|
|
|
|
}
|
|
|
|
free(hostandport);
|
2013-11-29 03:49:01 +08:00
|
|
|
free(path);
|
2007-10-20 03:47:53 +08:00
|
|
|
return conn;
|
2005-07-05 02:57:58 +08:00
|
|
|
}
|
|
|
|
|
2007-10-20 03:47:53 +08:00
|
|
|
int finish_connect(struct child_process *conn)
|
2005-07-05 02:57:58 +08:00
|
|
|
{
|
2007-10-20 03:47:54 +08:00
|
|
|
int code;
|
2011-05-16 14:52:11 +08:00
|
|
|
if (!conn || git_connection_is_socket(conn))
|
2006-09-12 17:00:13 +08:00
|
|
|
return 0;
|
|
|
|
|
2007-10-20 03:47:54 +08:00
|
|
|
code = finish_command(conn);
|
2007-10-20 03:47:53 +08:00
|
|
|
free(conn);
|
2007-10-20 03:47:54 +08:00
|
|
|
return code;
|
2005-07-05 02:57:58 +08:00
|
|
|
}
|