Tor Lillqvist [Thu, 24 Jan 2008 09:44:38 +0000 (09:44 +0000)]
Allow autogening even without AM_PATH_LIBGCRYPT available
2008-01-23 Tor Lillqvist <tml@novell.com>
* configure.in: Allow autogening even without AM_PATH_LIBGCRYPT
available
* libsoup/soup-date.c (soup_date_new_from_time_t): Correct use of
gmtime().
* libsoup/soup-headers.c (soup_headers_parse_status_line): Return
FALSE if neither HTTP nor ICY. Avoids crash in
tests/header-parsing.
* libsoup/soup-socket.c: On Windows SHUT_RDWR is called SD_BOTH.
(set_nonblocking): Fix typo.
* tests/continue-test.c: Seems to build fine without <pthread.h>,
so drop that.
svn path=/trunk/; revision=1051
Dan Winship [Fri, 18 Jan 2008 15:21:43 +0000 (15:21 +0000)]
test SOUP_AUTH_DOMAIN_REMOVE_PATH
* tests/server-auth-test.c: test SOUP_AUTH_DOMAIN_REMOVE_PATH
svn path=/trunk/; revision=1050
Dan Winship [Fri, 18 Jan 2008 14:55:31 +0000 (14:55 +0000)]
require glib 2.15.3, not 2.15.0, since AM_PATH_GLIB_2_0 didn't know about
* configure.in: require glib 2.15.3, not 2.15.0, since
AM_PATH_GLIB_2_0 didn't know about gio until post-2.15.2. Pointed
out by Matthew Barnes, #510216.
svn path=/trunk/; revision=1049
Dan Winship [Fri, 18 Jan 2008 14:53:35 +0000 (14:53 +0000)]
add a new generic auth callback that can be used with any subclass to do
* libsoup/soup-auth-domain.c
(soup_auth_domain_set_generic_auth_callback):
(soup_auth_domain_check_password): add a new generic auth callback
that can be used with any subclass to do cleartext password
checking against messages. Suggested by Mathias Hasselmann.
* libsoup/soup-auth-domain-basic.c: Implement generic auth
* libsoup/soup-auth-domain-digest.c: Implement generic auth.
(soup_auth_domain_digest_evil_check_password): Gone, use the
generic version now.
svn path=/trunk/; revision=1048
Dan Winship [Thu, 17 Jan 2008 17:06:47 +0000 (17:06 +0000)]
cast the second arg to g_checksum_update to (guchar *) to avoid warnings
* libsoup/soup-auth-digest.c (soup_auth_digest_compute_hex_urp)
(soup_auth_digest_compute_hex_a1)
(soup_auth_digest_compute_response): cast the second arg to
g_checksum_update to (guchar *) to avoid warnings
svn path=/trunk/; revision=1047
Dan Winship [Wed, 16 Jan 2008 22:00:32 +0000 (22:00 +0000)]
Deal with Shoutcast servers, which return "ICY 200 OK", but are otherwise
* libsoup/soup-headers.c (soup_headers_parse_status_line): Deal
with Shoutcast servers, which return "ICY 200 OK", but are
otherwise straight HTTP/1.0. #502325, Wouter Cloetens.
* tests/header-parsing.c (resptests): add a test for it
svn path=/trunk/; revision=1046
Dan Winship [Wed, 16 Jan 2008 21:49:54 +0000 (21:49 +0000)]
Allow the session authenticate signal to be handled asynchronously, by
* libsoup/soup-auth-manager.c (authorize_handler, etc): Allow the
session authenticate signal to be handled asynchronously, by
pausing the message and then authenticating the auth later.
(auth_type_compare_func): make this work. oops.
(extract_challenge): plug leak
* libsoup/soup-auth-manager-ntlm.c: Make this work async too.
* libsoup/soup-headers.c (soup_header_parse_list):
(soup_header_parse_param_list): plug leaks
* tests/auth-test.c (do_async_auth_test): test async auth
* docs/reference/client-howto.xml (Handling Authentication):
mention async auth
svn path=/trunk/; revision=1045
Dan Winship [Wed, 16 Jan 2008 19:16:51 +0000 (19:16 +0000)]
Bomb out if glib 2.15.0 isn't found. (AM_PATH_GLIB_2_0 doesn't do this
* configure.in: Bomb out if glib 2.15.0 isn't found.
(AM_PATH_GLIB_2_0 doesn't do this itself.)
svn path=/trunk/; revision=1044
Dan Winship [Tue, 15 Jan 2008 22:52:36 +0000 (22:52 +0000)]
Replaces SoupConnectionNTLM; now works as a SoupSession::request_started
* libsoup/soup-auth-manager-ntlm.c: Replaces SoupConnectionNTLM;
now works as a SoupSession::request_started watcher.
* libsoup/soup-connection.c: remove the no-longer-needed
"authenticate" signal
* libsoup/soup-session.c: Use a SoupAuthManagerNTLM if USE_NTLM is
set. Remove connection-authenticate-signal references.
svn path=/trunk/; revision=1043
Dan Winship [Tue, 15 Jan 2008 17:40:47 +0000 (17:40 +0000)]
Merge libsoup-2.4 branch to trunk
* Merge libsoup-2.4 branch to trunk
svn path=/trunk/; revision=1041
Benjamin Otte [Wed, 9 Jan 2008 21:28:48 +0000 (21:28 +0000)]
clarify docs for new-connetion signal.
* libsoup/soup-socket.c: (soup_socket_class_init): clarify docs for new-connetion signal.
svn path=/trunk/; revision=1024
Dan Winship [Mon, 26 Nov 2007 22:33:58 +0000 (22:33 +0000)]
wait for apache to exit, to avoid spurious test failures
svn path=/trunk/; revision=959
Dan Winship [Mon, 26 Nov 2007 22:19:02 +0000 (22:19 +0000)]
2.2.104
* configure.in: 2.2.104
* NEWS: update
svn path=/trunk/; revision=958
Dan Winship [Mon, 26 Nov 2007 16:59:21 +0000 (16:59 +0000)]
fix docs
svn path=/trunk/; revision=957
Dan Winship [Wed, 21 Nov 2007 16:23:26 +0000 (16:23 +0000)]
make this non-static.
* libsoup/soup-message-io.c (soup_message_io_cleanup): make this
non-static.
* libsoup/soup-message.c (finalize): Use soup_message_io_cleanup()
rather than soup_message_io_stop(), to avoid leaks when finalizing
an unfinished message. (Another part of #498509, Wouter Cloetens.)
svn path=/trunk/; revision=956
Dan Winship [Wed, 21 Nov 2007 04:17:54 +0000 (04:17 +0000)]
add bug reference to previous commit:
Fix up SOUP_SESSION_ASYNC_CONTEXT. #498509, Wouter Cloetens
svn path=/trunk/; revision=955
Dan Winship [Wed, 21 Nov 2007 04:15:53 +0000 (04:15 +0000)]
don't leak the async_context
* libsoup/soup-message-io.c (soup_message_io_unpause): don't leak
the async_context
* libsoup/soup-server.c (soup_server_quit): disconnect the
"new_connection" handler.
(soup_server_get_async_context): Convenience method to return the
server's async_context.
* libsoup/soup-server-message.c: don't circularly ref the server,
there's no need anyway.
* libsoup/soup-session.c (soup_session_get_async_context):
Convenience method to return the session's async_context.
* libsoup/soup-session-async.c (queue_message): call run_queue in
the session's async_context, not the main context.
(send_message): don't leak the async_context
* libsoup/soup-session-sync.c (queue_message_thread): don't leak
the async_context
* tests/context-test.c: test that SOUP_SESSION_ASYNC_CONTEXT works
and doesn't leak
svn path=/trunk/; revision=954
Dan Winship [Wed, 21 Nov 2007 03:51:24 +0000 (03:51 +0000)]
don't leak the SoupAddress.
* libsoup/soup-connection.c (soup_connection_connect_async): don't
leak the SoupAddress.
* libsoup/soup-dns.c (soup_dns_lookup_resolve_async): fix a leak
when re-looking up an address
* libsoup/soup-session.c (soup_session_abort): close all
connections in addition to cancelling messages (needed because
connections currently end up holding a ref on their session,
preventing them from being destroyed).
* tests/auth-test.c:
* tests/ntlm-test.c:
* tests/proxy-test.c:
* tests/pull-api.c:
* tests/ssl-test.c:
* tests/xmlrpc-test.c: clean up more memory on exit, to help find
leaks in the library
* tests/libsoup.supp: add a zillion new suppressions so we
can use --leak-resolution=med
svn path=/trunk/; revision=953
Dan Winship [Mon, 19 Nov 2007 17:51:42 +0000 (17:51 +0000)]
Fix the guards around the got_chunk emission so that it doesn't get messed
* libsoup/soup-message-io.c (read_body_chunk): Fix the guards
around the got_chunk emission so that it doesn't get messed up if
you pause the I/O from the got_chunk handler. (#452280, Marco
Barisione).
(soup_message_io_pause, soup_message_io_unpause): Update docs
again; these are now allowed with client-side I/O as well. Fix
unpause() to unpause asynchronously on async sockets.
* libsoup/soup-session-async.c (send_message): Iterate session's
async_context, not the default main context.
* tests/pull-api.c: Test/sample of creating a pull-style API using
SoupSessionAsync.
* tests/index.txt: new file to act as DirectoryIndex for the
tests. (In particular, pull-api wants this to be largeish.)
svn path=/trunk/; revision=952
Dan Winship [Sun, 28 Oct 2007 17:50:14 +0000 (17:50 +0000)]
2.2.103
* configure.in: 2.2.103
* NEWS: update
svn path=/trunk/; revision=949
Dan Winship [Sun, 28 Oct 2007 17:29:47 +0000 (17:29 +0000)]
remove the g_return_if_fail()s from the last patch so I don't end up breaking
anything that worked before...
svn path=/trunk/; revision=948
Dan Winship [Sun, 28 Oct 2007 17:23:38 +0000 (17:23 +0000)]
ref the socket around the processing of the message, since otherwise it
* libsoup/soup-server.c (start_request, request_finished): ref the
socket around the processing of the message, since otherwise it
might already be freed when request_finished runs. #459896.
* libsoup/soup-message-io.c (soup_message_io_pause)
(soup_message_io_unpause): Clarify the docs here; this is for
server-side use only. Inspired by #452280.
* docs/reference/server-howto.xml: You need to watch the
"finished" signal on the message if using soup_message_io_pause()
or chunked encoding, because the client might disconnect while
you're paused. Clarification inspired by #471385.
* tests/simple-proxy.c (client_msg_failed): Fix this to DTRT since
server-howto.xml points to it as an example of what to do.
svn path=/trunk/; revision=947
Dan Winship [Sun, 28 Oct 2007 16:11:02 +0000 (16:11 +0000)]
free proxy-related stuff
* libsoup/soup-session.c (finalize): free proxy-related stuff
* libsoup/soup-session-async.c (idle_run_queue): clean up the weak
pointer
* tests/*.c: fix leaks
svn path=/trunk/; revision=946
Dan Winship [Sun, 28 Oct 2007 15:57:55 +0000 (15:57 +0000)]
Make these less verbose by default (to make it easier to see what failed
* tests/auth-test.c:
* tests/date.c:
* tests/header-parsing.c:
* tests/ntlm-test.c:
* tests/proxy-test.c:
* tests/uri-parsing.c:
* tests/xmlrpc-test.c: Make these less verbose by default (to make
it easier to see what failed when a "make check" fails).
svn path=/trunk/; revision=945
Dan Winship [Mon, 15 Oct 2007 14:41:30 +0000 (14:41 +0000)]
2.2.102
* configure.in: 2.2.102
* NEWS: update
svn path=/trunk/; revision=942
Dan Winship [Mon, 15 Oct 2007 14:40:58 +0000 (14:40 +0000)]
update
svn path=/trunk/; revision=941
Dan Winship [Tue, 9 Oct 2007 20:55:22 +0000 (20:55 +0000)]
Update for current soup-ssl.h prototypes
* libsoup/soup-nossl.c: Update for current soup-ssl.h prototypes
* tests/proxy-test.c (run_test): wrap https tests in #if HAVE_SSL
svn path=/trunk/; revision=940
Dan Winship [Fri, 5 Oct 2007 16:34:32 +0000 (16:34 +0000)]
srcdir != builddir fixes
* tests/httpd.conf.in:
* tests/Makefile.am:
* tests/ssl-test.c: srcdir != builddir fixes
svn path=/trunk/; revision=937
Dan Winship [Fri, 5 Oct 2007 15:54:32 +0000 (15:54 +0000)]
2.2.101
* configure.in: 2.2.101
* NEWS: update
svn path=/trunk/; revision=936
Dan Winship [Fri, 5 Oct 2007 15:51:27 +0000 (15:51 +0000)]
Don't crash if the authenticate callback returns a username and no
* libsoup/soup-connection-ntlm.c (ntlm_authorize_pre): Don't crash
if the authenticate callback returns a username and no password
(even though it's not supposed to do that). #480987
svn path=/trunk/; revision=935
Dan Winship [Tue, 25 Sep 2007 03:11:18 +0000 (03:11 +0000)]
add "realm" field to the struct. (SoupAuthClass) remove "get_realm"
* libsoup/soup-auth.h (SoupAuth): add "realm" field to the struct.
(SoupAuthClass) remove "get_realm" virtual method.
* libsoup/soup-auth.c (soup_auth_new_from_header_list): Parse the
WWW-Authenticate/Proxy-Authenticate header here, set realm, and
pass the params hash to the construct method.
(soup_auth_get_info): Return an identifier for the auth:
"SCHEME:REALM"
* libsoup/soup-auth-basic.c:
* libsoup/soup-auth-digest.c: update
* libsoup/soup-session.c (invalidate_auth, update_auth_internal):
use soup_auth_get_info().
svn path=/trunk/; revision=934
Dan Winship [Tue, 25 Sep 2007 00:23:17 +0000 (00:23 +0000)]
minor rfc850-date parsing improvement suggested by RFC2616 19.3.
* libsoup/soup-date.c (soup_date_parse): minor rfc850-date parsing
improvement suggested by RFC2616 19.3.
* libsoup/soup-headers.c (soup_headers_parse_request): allow
erroneous trailing whitespace after HTTP version. #475169
* libsoup/soup-message-server-io.c (parse_request_headers): fix
the parsing of the Host header to assume it already includes the
port (which it should; the only reason this ever worked is because
SoupUri ignores the second port number when parse_request_headers
generates a URL like "http://localhost:9999:9999/").
* tests/header-parsing.c (reqtests): add a test for #475169
svn path=/trunk/; revision=933
Dan Winship [Mon, 24 Sep 2007 01:00:04 +0000 (01:00 +0000)]
remove a mysterious partial sentence in the ::wrote-chunk docstring.
* libsoup/soup-message.c (soup_message_class_init): remove a
mysterious partial sentence in the ::wrote-chunk docstring.
#458116
* docs/reference/libsoup-sections.txt: Remove documentation of MD5
methods, which are not public. #440092
svn path=/trunk/; revision=932
Dan Winship [Mon, 24 Sep 2007 00:53:31 +0000 (00:53 +0000)]
update
svn path=/trunk/; revision=931
Dan Winship [Mon, 24 Sep 2007 00:53:00 +0000 (00:53 +0000)]
Only remove the Authorization / Proxy-Authorization header from the
* libsoup/soup-message.c (soup_message_set_auth)
(soup_message_set_proxy_auth): Only remove the Authorization /
Proxy-Authorization header from the message if it was previously
set by soup_message_set_auth(). (Eg, not if it was added by
SoupConnectionNTLM.) #471389
* libsoup/soup-connection-ntlm.h: fix a search-and-replace-o
* tests/ntlm-test.c: Simple NTLM regression test; doesn't really
test the crypto/encoding bits, just that the right headers are
being sent at the right times.
svn path=/trunk/; revision=930
Dan Winship [Fri, 14 Sep 2007 19:50:23 +0000 (19:50 +0000)]
Make "make check" pass on Fedora 7:
* configure.in: update apache/php tests with additional filenames
* tests/httpd.conf.in: updates for configure.in changes and
slightly-more-recent apache
* tests/ssl-test.c (start_writing): fix uninitialized struct field
(main): start server after setting up client since otherwise
there's a race condition since soup_gnutls_init() isn't actually
thread-safe. (FIXME)
* tests/xmlrpc-server.php: rewrite to not use $HTTP_RAW_POST_DATA
(which only exists if register_globals is set)
svn path=/trunk/; revision=929
Dan Winship [Thu, 9 Aug 2007 14:45:43 +0000 (14:45 +0000)]
update for new maintainers format (and remove old email address)
svn path=/trunk/; revision=927
Dan Winship [Fri, 1 Jun 2007 20:46:15 +0000 (20:46 +0000)]
fix. noted by "cascardo" on libsoup-list.
* libsoup/soup-message-filter.h (SOUP_IS_MESSAGE_FILTER_CLASS):
fix. noted by "cascardo" on libsoup-list.
svn path=/trunk/; revision=926
Jonathon Jongsma [Fri, 1 Jun 2007 20:41:47 +0000 (20:41 +0000)]
add G_BEGIN_DECLS / G_END_DECLS to all installed headers so that libsoup
2007-05-16 Jonathon Jongsma <jjongsma@gnome.org>
* libsoup/*.h: add G_BEGIN_DECLS / G_END_DECLS to all installed
headers so that libsoup can be used from C++ programs. #438776
svn path=/trunk/; revision=925
Dan Winship [Mon, 16 Apr 2007 17:40:25 +0000 (17:40 +0000)]
Make a real SoupSSLCredentials type rather than just using gpointer
* libsoup/soup-ssl.h: Make a real SoupSSLCredentials type rather
than just using gpointer
* libsoup/soup-server.c (SoupServerPrivate): use it
* libsoup/soup-session.c (SoupSessionPrivate): use it
* libsoup/soup-gnutls.c: Use it, and consistently use "creds"
rather than "cred" as the abbreviation for "credentials".
* docs/reference/libsoup-sections.txt:
* docs/reference/tmpl/soup-misc.sgml:
* docs/reference/tmpl/soup-ssl.sgml: update
svn path=/trunk/; revision=924
Dan Winship [Fri, 30 Mar 2007 14:58:54 +0000 (14:58 +0000)]
Implement this by sending the message (synchronously) in another thread
* libsoup/soup-session-sync.c (queue_message): Implement this by
sending the message (synchronously) in another thread and then
queueing the callback back in the main thread.
* libsoup/soup-session.c (soup_session_queue_message): update docs
to be more explicit about what thread the callback occurs in
svn path=/trunk/; revision=923
Dan Winship [Sat, 17 Mar 2007 18:51:28 +0000 (18:51 +0000)]
get/set auth/proxy_auth info for a message.
* libsoup/soup-message.c (soup_message_set_auth)
(soup_message_get_auth, soup_message_set_proxy_auth)
(soup_message_get_proxy_auth): get/set auth/proxy_auth info for a
message.
* libsoup/soup-session.c (add_auth): Use soup_message_set_auth and
soup_message_set_proxy_auth.
(update_auth_internal): Call soup_message_get_auth or
soup_message_get_proxy_auth to determine the message's prior auth,
rather than calling lookup_auth() again, since it isn't guaranteed
to return the same thing now as it did when the message was
originally sent. Fixes erroneous 401s when queuing multiple
messages at once to an as-yet-unauthenticated-to server. #271540
* libsoup/soup-session-async.c (queue_message): don't run the
queue right away, do it at idle time. Otherwise in some cases
(especially errors), the message callbacks could be invoked before
queue_message returns.
* tests/auth-test.c: add a regression test for #271540.
svn path=/trunk/; revision=922
Dan Winship [Sat, 17 Mar 2007 17:55:30 +0000 (17:55 +0000)]
require glib 2.12. check for timegm().
* configure.in: require glib 2.12. check for timegm().
* libsoup/soup-date.c (soup_mktime_utc): Use timegm if available.
(soup_date_iso8601_parse): use g_time_val_from_iso8601. #337010,
patch from Emmanuele Bassi.
* libsoup/soup-types.h: remove local copy of
G_GNUC_NULL_TERMINATED since we now depend on a new-enough copy of
glib.
* libsoup/soup-misc.c (soup_base64_encode_close)
(soup_base64_encode_step, soup_base64_encode)
(soup_base64_decode_step): Make these just be wrappers around the
glib base64 methods. (For now; eventually they'll just go away.)
* libsoup/soup-auth-basic.c (authenticate):
* libsoup/soup-auth-digest.c (authenticate):
* libsoup/soup-connection-ntlm.c (soup_ntlm_parse_challenge)
(soup_ntlm_response):
* libsoup/soup-server-auth.c (soup_server_auth_new):
* libsoup/soup-soap-message.c (soup_soap_message_write_base64):
* libsoup/soup-xmlrpc-message.c
(soup_xmlrpc_message_write_base64):
* libsoup/soup-xmlrpc-response.c (soup_xmlrpc_value_get_base64):
Use glib base64 methods
svn path=/trunk/; revision=921
Dan Winship [Fri, 16 Mar 2007 21:02:17 +0000 (21:02 +0000)]
update the handling of CONNECT: it has no response body by default, but
* libsoup/soup-message.c (soup_message_get_response_encoding):
update the handling of CONNECT: it has no response body by
default, but does have a body if its headers say so.
* tests/proxy-test.c: test libsoup's behavior when talking to
proxies.
* tests/httpd.conf.in: Load mod_proxy and mod_ssl, and add
sections configuring them, for proxy-test
* configure.in: update the apache-module-dir-finding code to deal
with the fact that some modules (eg, mod_ssl) might only be in the
mpm-specific module dir, while others (eg, mod_php5) might only be
in the generic module dir.
svn path=/trunk/; revision=920
Dan Winship [Fri, 16 Mar 2007 18:07:57 +0000 (18:07 +0000)]
add <sys/socket.h> for solaris. #418919, from Wang Xin
svn path=/trunk/; revision=919
Dan Winship [Tue, 13 Mar 2007 01:00:54 +0000 (01:00 +0000)]
add $(LIBGNUTLS_CFLAGS) for ssl-test. #417617, patch from Elijah Newren.
* tests/Makefile.am (INCLUDES): add $(LIBGNUTLS_CFLAGS) for
ssl-test. #417617, patch from Elijah Newren.
svn path=/trunk/; revision=917
Dan Winship [Tue, 13 Mar 2007 00:58:28 +0000 (00:58 +0000)]
if soup_connection_connect_sync() returns SOUP_STATUS_TRY_AGAIN, then try
* libsoup/soup-session-sync.c (wait_for_connection): if
soup_connection_connect_sync() returns SOUP_STATUS_TRY_AGAIN, then
try again. (Duh.) Fixes SSL-via-proxy-when-using-synchronous-I/O-
where-the-proxy-closes-the-connection-when-returning-407. (Reported
by Varadhan.)
* tests/get.c: Rewrite to use soup_session_send_message rather
than soup_session_queue_message, and add a "-s" flag to use
SoupSessionSync rather than SoupSessionAsync (so we can test bugs
in the sync code paths).
svn path=/trunk/; revision=916
Dan Winship [Thu, 8 Mar 2007 21:11:00 +0000 (21:11 +0000)]
don't return G_IO_STATUS_AGAIN if we're doing blocking I/O; just keep
* libsoup/soup-gnutls.c (do_handshake): don't return
G_IO_STATUS_AGAIN if we're doing blocking I/O; just keep retrying
until the handshake is complete.
(soup_gnutls_read, soup_gnutls_write): if we get
GNUTLS_E_REHANDSHAKE, call do_handshake() immediately rather than
returning G_IO_STATUS_AGAIN; if the socket is blocking then
G_IO_STATUS_AGAIN is wrong, and if the socket is non-blocking, we
might already need to return SOUP_SSL_ERROR_HANDSHAKE_NEEDS_WRITE
or SOUP_SSL_ERROR_HANDSHAKE_NEEDS_READ.
#415402, based on a patch from Jacob Berkman.
* tests/ssl-test.c: basic ssl test. In particular, tests that
rehandshake requests are handled correctly during both synchronous
and asynchronous I/O. Might eventually test other stuff too...
* configure.in:
* tests/Makefile.am: updates for ssl-test
svn path=/trunk/; revision=914
Dan Winship [Mon, 19 Feb 2007 23:31:38 +0000 (23:31 +0000)]
Get gcrypt libs/cflags.
* configure.in: Get gcrypt libs/cflags.
* libsoup/Makefile.am (INCLUDES, libsoup_2_2_la_LIBADD): add
gcrypt flags. Patch from "Cygwin Ports Maintainer", #384498
svn path=/trunk/; revision=913
Dan Winship [Mon, 12 Feb 2007 15:35:04 +0000 (15:35 +0000)]
Fixed soup_headers_parse_status_line() so WebDAV response parsing will
* Fixed soup_headers_parse_status_line() so WebDAV response
parsing will work again. [406997]
* Fixed a bug in the header-parsing regression test that
caused the test to fail sometimes, even though the actual
header-parsing code was fine.
svn path=/trunk/; revision=910
Dan Winship [Mon, 12 Feb 2007 15:14:08 +0000 (15:14 +0000)]
Fix this to handle "\0"-terminated status lines (eg, from WebDAV
* libsoup/soup-headers.c (soup_headers_parse_status_line): Fix
this to handle "\0"-terminated status lines (eg, from WebDAV
responses), like the docs say it does. #406997
(soup_headers_parse): Balance that out by rejecting internal "\0"s
here.
(soup_headers_parse_request, soup_headers_parse_response): Update
docs to warn that @dest may be modified even on error. (This was
always true, it just wasn't documented.)
svn path=/trunk/; revision=909
Dan Winship [Tue, 16 Jan 2007 20:54:02 +0000 (20:54 +0000)]
initialize "errors" to 0. duh. Pointed out by Michael Wolf.
* tests/header-parsing.c (do_request_tests, do_response_tests):
initialize "errors" to 0. duh. Pointed out by Michael Wolf.
svn path=/trunk/; revision=908
Dan Winship [Mon, 8 Jan 2007 21:20:07 +0000 (21:20 +0000)]
2.2.99
* configure.in: 2.2.99
* NEWS: update
svn path=/trunk/; revision=906
Dan Winship [Sat, 6 Jan 2007 19:24:44 +0000 (19:24 +0000)]
Rewrite this to be easier to understand and more correct, and make the
* libsoup/soup-headers.c (soup_headers_parse): Rewrite this to be
easier to understand and more correct, and make the "str" param
const rather than overwriting it during parsing.
(soup_headers_parse_request, soup_headers_parse_response):
Likewise, make "str" param const. Fix the doc comment to describe
the correct constraint on str. Make the parsing slightly more
lenient as per sections 4.1 and 19.3 of RFC 2616.
* tests/header-parsing.c: new regression test, for Request-Line,
Status-Line, and message-header parsing.
Inspired by #391970 (crash in SoupServer when certain invalid
requests are received).
svn path=/trunk/; revision=905
Dan Winship [Tue, 5 Dec 2006 19:03:27 +0000 (19:03 +0000)]
Free ssl_creds. Pointed out by Chris Austin.
* libsoup/soup-session.c (finalize): Free ssl_creds. Pointed out
by Chris Austin.
Dan Winship [Tue, 5 Dec 2006 18:37:16 +0000 (18:37 +0000)]
Remove the calls to soup_message_io_stop() here; the corresponding calls
* libsoup/soup-message.c (soup_message_set_uri): Remove the calls
to soup_message_io_stop() here; the corresponding calls were
needed back in the SoupContext days, but they are wrong now and
cause async-redirects-to-other-hosts to fail. #382251. Also
clarify docs with respect to soup_session_requeue_message().
* libsoup/soup-message-io.c (soup_message_io_stop): Clarify docs
Dan Winship [Mon, 20 Nov 2006 21:52:58 +0000 (21:52 +0000)]
2.2.98
* configure.in: 2.2.98
* NEWS: update
Dan Winship [Mon, 20 Nov 2006 21:47:04 +0000 (21:47 +0000)]
Patch from Andrew W. Nosenko:
* libsoup/soup-message-client-io.c (parse_response_headers): Avoid
memory leak when parse_response_headers() is called on a message
that has a 'reason_phrase' already for some reason.
* libsoup/soup-gnutls.c (soup_gnutls_free): Avoid memory leak:
hostname was not freed.
(soup_ssl_wrap_iochannel): Avoid memory leak: SoupGNUTLSChannel
'chan' was not freed in case of initialization error. Avoid double
close of the "real" (plain, non-ssl) channel FD.
* libsoup/soup-socket.c (soup_socket_start_proxy_ssl): Avoid
memory leak: the "real" (plain, non-ssl) GIOChannel was never
"finally" unreffed (one more *_ref() than *_unref()) in case of
ssl-wrapping.
Dan Winship [Mon, 20 Nov 2006 21:38:18 +0000 (21:38 +0000)]
Add some missing "static"s. Patch from Matthias Clasen, #376387
* libsoup/soup-connection-ntlm.c (send_request):
* libsoup/soup-session-sync.c (queue_message):
* libsoup/soup-status.c (reason_phrases): Add some missing
"static"s. Patch from Matthias Clasen, #376387
Dan Winship [Mon, 20 Nov 2006 21:34:53 +0000 (21:34 +0000)]
<value>foo</value> should mean the same thing as
* libsoup/soup-xmlrpc-response.c (soup_xmlrpc_value_get_type)
(soup_xmlrpc_value_get_string): <value>foo</value> should mean the
same thing as <value><string>foo</string></value>. Pointed out by
Todd Kulesza. #364490
Dan Winship [Mon, 6 Nov 2006 20:55:24 +0000 (20:55 +0000)]
Bump version to 2.2.97. Bump AGE and CURRENT for addition of
* configure.in: Bump version to 2.2.97. Bump AGE and CURRENT for
addition of soup_xml_real_node.
* NEWS: update
Dan Winship [Mon, 6 Nov 2006 20:35:28 +0000 (20:35 +0000)]
new method to find a "real" (ie, not comment or whitespace) xml node
* libsoup/soup-misc.c (soup_xml_real_node): new method to find a
"real" (ie, not comment or whitespace) xml node
* libsoup/soup-soap-response.c (parse_parameters)
(soup_soap_response_from_string)
(soup_soap_parameter_get_first_child)
(soup_soap_parameter_get_next_child): Use soup_xml_real_node.
Based on a patch from Andrew W. Nosenko.
* libsoup/soup-xmlrpc-message.c (soup_xmlrpc_message_from_string):
don't call xmlKeepBlanksDefault, which changes libxml's behavior
globally! Instead, use soup_xml_real_node() when traversing the
xml tree.
* libsoup/soup-xmlrpc-response.c
(soup_xmlrpc_response_from_string): don't call
xmlKeepBlanksDefault.
(exactly_one_child): rewrite in terms of soup_xml_real_node()
(which means it handles comments now as well)
(soup_xmlrpc_value_get_struct)
(soup_xmlrpc_value_array_get_iterator)
(soup_xmlrpc_value_array_iterator_prev)
(soup_xmlrpc_value_array_iterator_next): Use soup_xml_real_node.
Dan Winship [Sun, 5 Nov 2006 22:01:59 +0000 (22:01 +0000)]
misc minor updates
Dan Winship [Sun, 5 Nov 2006 22:01:27 +0000 (22:01 +0000)]
document this (in particular, point out that str+len must point to exactly
* libsoup/soup-headers.c (soup_headers_parse_request): document
this (in particular, point out that str+len must point to exactly
the right place). Allow req_method and req_path to be NULL.
(soup_headers_parse_status_line, soup_headers_parse_response):
document. Also, change "status_phrase" argument to "reason_phrase"
to match the spec. Inspired by #339889.
Dan Winship [Fri, 3 Nov 2006 18:32:16 +0000 (18:32 +0000)]
Fix an off by one pointed out by Andrew W. Nosenko.
* libsoup/soup-uri.c (soup_uri_get_protocol): Fix an off by one
pointed out by Andrew W. Nosenko.
Dan Winship [Fri, 3 Nov 2006 18:31:08 +0000 (18:31 +0000)]
fix lots of warnings. Partially from patches from Andrew W. Nosenko, and
* libsoup/*.c: fix lots of warnings. Partially from patches from
Andrew W. Nosenko, and also some fixes from libsoup-pre214-branch.
Dan Winship [Fri, 3 Nov 2006 18:12:15 +0000 (18:12 +0000)]
remove gnutls stuff (we're using pkg-config now)
* acinclude.m4: remove gnutls stuff (we're using pkg-config now)
Dan Winship [Fri, 3 Nov 2006 17:58:33 +0000 (17:58 +0000)]
uninstall the pkgconfig file. Based on #356809 from Matthew Barnes.
* Makefile.am (uninstall-local): uninstall the pkgconfig file.
Based on #356809 from Matthew Barnes.
Dan Winship [Fri, 3 Nov 2006 17:54:02 +0000 (17:54 +0000)]
Fix leaks pointed out by Paolo Borelli. #351500
* libsoup/soup-server.c (get_property): Fix leaks pointed out by
Paolo Borelli. #351500
Dan Winship [Fri, 3 Nov 2006 17:36:23 +0000 (17:36 +0000)]
Use pkgconfig to find gnutls. Remove old static linking stuff that was
* configure.in: Use pkgconfig to find gnutls. Remove old static
linking stuff that was only needed for rcd.
* libsoup.pc.in: Use Requires rather than putting xml/ssl
dependencies directly into Libs/Cflags. From Mikhail Zabaluev.
#343340.
Dan Winship [Fri, 3 Nov 2006 16:46:38 +0000 (16:46 +0000)]
fix build on cygwin. From "Cygwin Ports maintainer", #321827.
* libsoup/Makefile.am (libsoup_2_2_la_LDFLAGS): fix build on
cygwin. From "Cygwin Ports maintainer", #321827.
Dan Winship [Fri, 3 Nov 2006 16:46:15 +0000 (16:46 +0000)]
update
Dan Winship [Mon, 24 Jul 2006 22:14:56 +0000 (22:14 +0000)]
2.2.96. bump AGE and CURRENT for new API
* configure.in: 2.2.96. bump AGE and CURRENT for new API
* NEWS: update
* libsoup/soup-xmlrpc-message.c (soup_xmlrpc_message_from_string):
New, from Fernando Herrera, bug 348532.
Dan Winship [Mon, 24 Jul 2006 22:05:57 +0000 (22:05 +0000)]
New, from Fernando Herrera, bug 348532.
* libsoup/soup-xmlrpc-message.c (soup_xmlrpc_message_from_string):
New, from Fernando Herrera, bug 348532.
Dan Winship [Fri, 21 Jul 2006 18:55:46 +0000 (18:55 +0000)]
if the constructed auth doesn't have a realm, it's invalid, as per RFC
* libsoup/soup-auth.c (soup_auth_new_from_header_list): if the
constructed auth doesn't have a realm, it's invalid, as per RFC
2617. Based on a patch from Nate Nielsen on libsoup-list.
(soup_auth_get_realm): remove "if available" from docs; all auths
always have a realm.
Dan Winship [Fri, 21 Jul 2006 18:06:47 +0000 (18:06 +0000)]
update to mention mailing list and bugzilla
* README: update to mention mailing list and bugzilla
* HACKING: kill this since there's nothing here that isn't either
obvious, or redundant with the README
Dan Winship [Fri, 21 Jul 2006 18:05:53 +0000 (18:05 +0000)]
If the server handler set a Content-Length header on the message, don't
* libsoup/soup-message-server-io.c (get_response_headers): If the
server handler set a Content-Length header on the message, don't
add a second one. (Preserves compatibility with an old hacky way
that people might have been handling HEAD from SoupServer.)
Dan Winship [Fri, 21 Jul 2006 16:36:56 +0000 (16:36 +0000)]
initialize encoding to SOUP_TRANSFER_CONTENT_LENGTH rather than
* libsoup/soup-server-message.c (soup_server_message_init):
initialize encoding to SOUP_TRANSFER_CONTENT_LENGTH rather than
SOUP_TRANSFER_UNKNOWN, since SOUP_TRANSFER_UNKNOWN has never
actually worked here, and so there was an undocumented requirement
that you manually set the encoding on every response
(which SoupServer itself was not doing on internal errors).
Problem pointed out by Dennis Jacobfeuerborn on libsoup-list.
(soup_server_message_set_encoding): reject the new
SoupTransferEncoding values, for compatibility
* libsoup/soup-message.h (SoupTransferEncoding): Clarify that
SOUP_TRANSFER_UNKNOWN is essentially an error value, since in the
public API, it always has been, due to bugs. Add some new values,
currently just for internal use: SOUP_TRANSFER_NONE (for cases
like HEAD which never have a body), SOUP_TRANSFER_EOF (to replace
SOUP_TRANSFER_UNKNOWN), and SOUP_TRANSFER_BYTERANGES (which isn't
actually implemented yet).
* libsoup/soup-message.c (soup_message_get_request_encoding,
soup_message_get_response_encoding): figure out the body encoding
being used by the request/response, including all the tricky
cases like HEAD/1xx/etc.
(soup_message_is_keepalive): if the response encoding is
SOUP_TRANSFER_EOF, then the message isn't keepalive.
* libsoup/soup-message-client-io.c (parse_response_headers): use
soup_message_get_response_encoding.
* libsoup/soup-message-server-io.c (parse_request_headers): use
soup_message_get_request_encoding.
(get_response_headers): use both soup_server_message_get_encoding
and soup_message_get_response_encoding, to properly distinguish
between the wire encoding and the alleged-by-headers encoding
(which differ for HEAD, etc).
* libsoup/soup-message-io.c (io_error, read_body_chunk):
s/SOUP_TRANSFER_UNKNOWN/SOUP_TRANSFER_EOF/.
(io_body_state): if encoding is SOUP_TRANSFER_NONE, jump right to
SOUP_MESSAGE_IO_STATE_FINISHING.
* libsoup/soup-server.c (request_finished): Check
soup_socket_is_connected() *before* soup_message_is_keepalive(),
since the message will be invalid if the client unexpectedly
dropped the connection.
* tests/simple-httpd.c (server_callback): handle HEAD requests.
Remove no-longer-necessary soup_server_message_set_encoding()
call.
* tests/get.c: add -d (debug) flag to print headers, and -h flag
to do a HEAD rather than GET
Dan Winship [Thu, 13 Jul 2006 13:44:02 +0000 (13:44 +0000)]
sync to xmlrpc changes
Dan Winship [Mon, 10 Jul 2006 16:26:20 +0000 (16:26 +0000)]
2.2.95.1, and bump SOUP_AGE/SOUP_CURRENT this time. Pointed out by Daniel
* configure.in: 2.2.95.1, and bump SOUP_AGE/SOUP_CURRENT this
time. Pointed out by Daniel Holbach.
* configure.in: 2.2.95
* NEWS: update
Dan Winship [Mon, 19 Jun 2006 20:23:12 +0000 (20:23 +0000)]
don't build xmlrpc-test unless we have apache/php/xmlrpc-epi. Fixes the
* tests/Makefile.am (noinst_PROGRAMS): don't build xmlrpc-test
unless we have apache/php/xmlrpc-epi. Fixes the build. #345342
* configure.in: fix some quoting
Dan Winship [Mon, 19 Jun 2006 16:57:47 +0000 (16:57 +0000)]
add tests for apache mod_php5 and xmlrpc-epi-php
* configure.in: add tests for apache mod_php5 and xmlrpc-epi-php
* tests/xmlrpc-test.c: XML-RPC regression test
* tests/xmlrpc-server.php: PHP server for xmlrpc-test
* tests/httpd.conf.in: add php stuff
* tests/apache-wrapper.c (apache_cleanup): Use "graceful-stop"
rather than "stop", so that it stops listening on the socket
before exiting, so that we can immediately start another apache
(eg, in "make check").
* libsoup/soup-date.c (soup_mktime_utc): Fix a bug in leap-year
counting.
* libsoup/soup-xmlrpc-message.c
(soup_xmlrpc_message_write_datetime): rename from
"..._write_time", to make it consistent with the XML-RPC type name
and the corresponding SoupXmlrpcResponse method. Also, fix it to
use the same ISO 8601 format as the spec, and use the right value
for the seconds field.
(soup_xmlrpc_message_write_base64): Change the buf arg to a
gconstpointer rather than a const char *.
* libsoup/soup-xmlrpc-response.c (soup_xmlrpc_value_get_base64):
Return a GByteArray containing the decoded data, rather than
the base64-encoded string.
(soup_xmlrpc_value_dump_internal): Update for that (and don't
leak it).
(soup_xmlrpc_value_array_get_iterator,
soup_xmlrpc_value_array_iterator_get_value): Make these actually
work.
Dan Winship [Mon, 12 Jun 2006 19:36:27 +0000 (19:36 +0000)]
2.2.94
* configure.in: 2.2.94
* NEWS: update
Dan Winship [Mon, 12 Jun 2006 19:36:14 +0000 (19:36 +0000)]
make auth-test distcheck
Dan Winship [Mon, 12 Jun 2006 19:35:45 +0000 (19:35 +0000)]
dist client-howto.xml and server-howto.xml
Dan Winship [Mon, 12 Jun 2006 19:35:12 +0000 (19:35 +0000)]
s/SOAP/XMLRPC/ in one place where it was missed long ago
Dan Winship [Mon, 12 Jun 2006 19:19:19 +0000 (19:19 +0000)]
New client and server API tutorials.
* docs/reference/client-howto.xml:
* docs/reference/server-howto.xml: New client and server API
tutorials.
* docs/reference/*: reorganize, regenerate, fill in some missing
pieces, etc
* libsoup/soup-connection.c (soup_connection_new): document the
varargs param
* libsoup/soup-date.h: sync prototypes to declarations for gtk-doc.
* libsoup/soup-headers.c (soup_headers_parse_response): fix typo
in doc.
Dan Winship [Mon, 12 Jun 2006 15:19:29 +0000 (15:19 +0000)]
Further fixes from Brent Smith. #344458. (soup_xmlrpc_value_get_boolean):
* libsoup/soup-xmlrpc-response.c (soup_xmlrpc_value_get_int,
soup_xmlrpc_value_get_double): Further fixes from Brent Smith.
#344458.
(soup_xmlrpc_value_get_boolean): Similar fix, plus actually set
the output parameter and make the return value match the other
get_* methods.
Dan Winship [Fri, 9 Jun 2006 16:28:13 +0000 (16:28 +0000)]
Add tests for apache, output tests/httpd.conf
* configure.in: Add tests for apache, output tests/httpd.conf
* tests/htdigest:
* tests/htpasswd:
* tests/httpd.conf.in: Apache 2.2 config files for auth-test
* tests/apache-wrapper.c (apache_init, apache_cleanup): functions
to start/stop apache
* tests/auth-test.c: Use apache-wrapper functions to start a local
apache process to test authentication against, since the auth-test
tree at developer.ximian.com went missing a long time ago. #311825
* tests/Makefile.am (auth_test_SOURCES): use apache-wrapper.c
(TESTS): include auth-test if HAVE_APACHE.
* libsoup/soup-session.c (lookup_auth): Fix this in the case of a
URI pointing to a directory rather than a file.
Dan Winship [Thu, 8 Jun 2006 12:40:19 +0000 (12:40 +0000)]
Check return value of strtol/g_ascii_strtod correctly. #344222, patch from
* libsoup/soup-xmlrpc-response.c (soup_xmlrpc_value_get_int,
soup_xmlrpc_value_get_double, soup_xmlrpc_value_get_boolean):
Check return value of strtol/g_ascii_strtod correctly. #344222,
patch from Brent Smith.
Dan Winship [Wed, 7 Jun 2006 19:42:21 +0000 (19:42 +0000)]
record whether or not the response was a fault.
* libsoup/soup-xmlrpc-response.c
(soup_xmlrpc_response_from_string): record whether or not the
response was a fault.
(soup_xmlrpc_response_is_fault): test that. #343973, patch from
Brent Smith.
Dan Winship [Mon, 29 May 2006 19:52:02 +0000 (19:52 +0000)]
2.2.93
* configure.in: 2.2.93
* NEWS: update
Dan Winship [Mon, 29 May 2006 19:41:14 +0000 (19:41 +0000)]
add a new state "FINISHING" which means "done I/O, but not yet done
* libsoup/soup-message-io.c (SoupMessageIOState): add a new state
"FINISHING" which means "done I/O, but not yet done processing and
cleanup" before "DONE" (which now always means "completely done").
(soup_message_io_stop): disconnect the socket if the read state is
"< FINISHING", not "!= DONE".
(io_error): on an EOF-that-signals-end-of-data, set state to
FINISHING and run io_read().
(io_read, io_write): remove the g_return_if_fails from before.
s/DONE/FINISHING/ in most places. In the FINISHING handler, stop
listening for the readable/writable signal (eg, so we don't end up
reading a following pipelined request), and set the state to DONE.
(soup_message_io_unpause): Only reconnect the readable/writable
signals if the io state isn't DONE. Guard the calls to
io_read/io_write better so that it's safe to call this even after
they are both DONE, since it may be easier for us to test that
than for the caller to.
Fixes 334469, 342640, and another bug caused by the earlier
workaround to 334469. Based on patches and analysis from William
Jon McCann and Armin Bauer.
* tests/simple-proxy.c (main): add g_thread_init (NULL) to make
this work again. (Pointed out by Alex Larsson)
Dan Winship [Fri, 26 May 2006 21:24:58 +0000 (21:24 +0000)]
#include <sys/time.h> for struct timeval. #342048
* libsoup/soup-socket.c: #include <sys/time.h> for struct timeval.
#342048
Dan Winship [Fri, 26 May 2006 18:14:37 +0000 (18:14 +0000)]
Start SSL after CONNECTing! Doh. Part of bnc #174255.
* libsoup/soup-connection.c (soup_connection_connect_sync): Start
SSL after CONNECTing! Doh. Part of bnc #174255.
(SoupConnectionMode): new enum for the three types of
SoupConnection (direct, proxy, tunnel).
(set_property): set priv->mode according to proxy_uri and
conn_uri.
(socket_connect_result, soup_connection_connect_sync): use
priv->mode to decide whether or not to tunnel.
(send_request): Only pass TRUE for is_proxy to
soup_message_send_request if mode is PROXY, not if it's TUNNEL.
(Also part of bnc #174255).
Dan Winship [Fri, 26 May 2006 14:40:11 +0000 (14:40 +0000)]
tests if IO is currently in progress on a message.
* libsoup/soup-message-io.c (soup_message_io_in_progress): tests
if IO is currently in progress on a message.
* libsoup/soup-session-async.c (run_queue): don't process messages
that are io_in_progress. #342545, fix based on analysis from Wang
Xin. (In the future we may want to re-fix this by adding a
REQUEUED message status separate from QUEUED.)
Dan Winship [Mon, 22 May 2006 14:47:30 +0000 (14:47 +0000)]
Don't free the hosts while holding host_lock; that's not allowed and can
* libsoup/soup-session.c (cleanup_hosts): Don't free the hosts
while holding host_lock; that's not allowed and can cause
deadlock. #309867. Based on a patch from Veerapuram Varadhan for
part of bnc #174255.
Dan Winship [Mon, 10 Apr 2006 18:09:59 +0000 (18:09 +0000)]
bump version to 2.2.92
* configure.in: bump version to 2.2.92
* NEWS: update
Dan Winship [Mon, 10 Apr 2006 18:09:01 +0000 (18:09 +0000)]
g_return_if_fail if these get called after the IO is done. This isn't
* libsoup/soup-message-io.c (io_write, io_read): g_return_if_fail
if these get called after the IO is done. This isn't supposed to
happen, but apparently does. Workaround for #334469.
* libsoup/soup-auth-digest.c (qop_types, algorithm_types):
NULL-terminate these so we don't crash when trying to parse an
invalid value. (Flip side of the previous #328615 patch.)
Dan Winship [Mon, 3 Apr 2006 02:58:02 +0000 (02:58 +0000)]
Write out correct digest algorithm value. #328615.
* libsoup/soup-server-auth.c (soup_server_auth_context_challenge):
Write out correct digest algorithm value. #328615.