Re: [Spice-devel] guest agent windows binaries

[Christophe Fergeau]

On Wed, Feb 12, 2014 at 11:08:42AM +1000, Lindsay Mathieson wrote:
 Is there somewhere I can download the windows guest agent binaries
 (vdagent, vdservice) from, so I can manually install them on Windows
 8?
 
 The link on the Spice Download page
   http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2
 is only to source, not to binaries

http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw32-spice-vdagent-0.7.2-1.fc19.noarch.rpm
and
http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw64-spice-vdagent-0.7.2-1.fc19.noarch.rpm
contain the binaries:

$ rpm -qpl ./mingw64-spice-vdagent-0.7.2-1.fc19.noarch.rpm
/usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdagent.exe
/usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdservice.exe

 
 
 Also - could someone *please* clarify the current and future status of
 windows 8/2012 drivers and guest tools? its extremely confusing trying
 to track down this information.

No QXL driver, the virtio-win drivers (located at
https://alt.fedoraproject.org/pub/alt/virtio-win/latest/images/bin/ ) seem
to have win8 versions, and the agent should be working as well on win8.
I've never tried these though.

Christophe


pgpWvXkugfp5y.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] Can't subscribe to Spice-devel

[shan wang]

Hi,

When I subscribe to Spice-devel mail list,  I encounter follow issue:

url: http://lists.freedesktop.org/mailman/listinfo/spice-devel

error info:
Bug in Mailman version 2.1.13

We're sorry, we hit a bug!

Please inform the webmaster for this site of this problem. Printing of
traceback and other system information has been explicitly inhibited, but
the webmaster can find this information in the Mailman error logs.


Thanks and Regards,

shan wang
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] Can't subscribe to Spice-devel

[Christophe Fergeau]

Hey,

On Wed, Feb 12, 2014 at 10:36:13AM +0800, shan wang wrote:
 Hi,
 
 When I subscribe to Spice-devel mail list,  I encounter follow issue:
 
 url: http://lists.freedesktop.org/mailman/listinfo/spice-devel
 
 error info:
 Bug in Mailman version 2.1.13

I replied offlist already, but I hit this issue when trying, and this seems
to be a known bug https://bugs.freedesktop.org/show_bug.cgi?id=74749 :(

Christophe


pgphHN0cfDLWj.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 03/13] spice-proxy: parse user and pass

[Christophe Fergeau]

On Tue, Feb 11, 2014 at 06:15:15PM +0100, Marc-André Lureau wrote:
 I am not fond of super-extra-small changes, I tend to make things that
 belong together in the same patch. This one could be splitted in 5
 patches or more, I don't see the point. All I want is a reasonable set
 of related changes that don't break things.

Well, this makes review more complicated. I was focused on reviewing a
change parsing user and password, so I tried to interpret this change as
something helping with that. When it did not seem useful for parsing
username and password, I was still stuck with am I missing something, or
is it just a parsing improvement?
If this small change were to be buggy in some corner cases a few years from
now, we'd be misled again if git bisect points us at a commit adding
user/password parsing rather than at a small commit doing some minor
cleanup. And we'll also won't know anything about the intent of this change
(was it a small and obvious cleanup? was it required for some reason? what
was that reason?).
So while it may make things more convenient now for the person doing the
commit, it's in my opinion (well not just me) better to avoid having silent
changes in commits.

 So should I split it or you agree it's fine as part of parsing changes?

I initially was planning to tell you to just go with it, and now I just
convinced myself that it's better to split it /o\

Christophe


pgpeTDDg3ZTkj.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] Can't subscribe to Spice-devel

[Christophe Fergeau]

Hey,

On Wed, Feb 12, 2014 at 10:36:13AM +0800, shan wang wrote:
 When I subscribe to Spice-devel mail list,  I encounter follow issue:
 
 url: http://lists.freedesktop.org/mailman/listinfo/spice-devel

After asking on IRC, it seems they had to disable subscriptions to mailing
list using the web interface as they were getting a lot of spam
subscriptions. Subscribing by using the email interface should still be
possible (see http://www.list.org/mailman-member/node13.html ).
The email address should be spice-devel-join  lists.freedesktop.org

Hope that helps,

Christophe


pgpaMDdrEedIA.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 03/13] spice-proxy: parse user and pass

[Marc-André Lureau]

On Wed, Feb 12, 2014 at 10:39 AM, Christophe Fergeau
cferg...@redhat.com wrote:
 On Tue, Feb 11, 2014 at 06:15:15PM +0100, Marc-André Lureau wrote:
 I am not fond of super-extra-small changes, I tend to make things that
 belong together in the same patch. This one could be splitted in 5
 patches or more, I don't see the point. All I want is a reasonable set
 of related changes that don't break things.

 Well, this makes review more complicated. I was focused on reviewing a
 change parsing user and password, so I tried to interpret this change as
 something helping with that. When it did not seem useful for parsing
 username and password, I was still stuck with am I missing something, or
 is it just a parsing improvement?

ok

 If this small change were to be buggy in some corner cases a few years from
 now, we'd be misled again if git bisect points us at a commit adding
 user/password parsing rather than at a small commit doing some minor
 cleanup. And we'll also won't know anything about the intent of this change
 (was it a small and obvious cleanup? was it required for some reason? what
 was that reason?).

In general, I agree, but I consider this to be simple, non-essential
and hopefully replacable in a near future with a real GUri parser (or
equivalent, like SoupUri)

 So while it may make things more convenient now for the person doing the
 commit, it's in my opinion (well not just me) better to avoid having silent
 changes in commits.
 So should I split it or you agree it's fine as part of parsing changes?

 I initially was planning to tell you to just go with it, and now I just
 convinced myself that it's better to split it /o\

I'll split this change off then.



-- 
Marc-André Lureau
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCHv2 spice-gtk 00/14] Add https + basic auth proxy support

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Hi,

The following series allows to connect to an HTTPS proxy, with Basic
authentication.

It also allows clients to ask credentials when required. This will
need some client side support, and a small patch is ready for
virt-viewer.

In v2:
- address Christophe's comments from v1 review

Marc-André Lureau (14):
  compat: add strtok_r fallback
  proxy: split uri with : in only 2 parts
  proxy: add user and pass properties
  proxy: parse user and pass from uri
  http-proxy: specify Basic scheme
  spice-proxy: parse https protocol
  http-proxy: add https proxy
  Fill g_proxy_address_new() with protocol, user and password
  openssl: learn to handle a new kind of BIO based on GIOStream
  channel: simplify has error code
  channel: talk to giostream instead of gsocket
  Make SpiceURI a public API
  channel: add spice_channel_get_error()
  session: add spice_session_get_proxy_uri()

 configure.ac |   2 +-
 doc/reference/Makefile.am|   1 +
 doc/reference/spice-gtk-docs.xml |   1 +
 doc/reference/spice-gtk-sections.txt |  28 +++
 gtk/Makefile.am  |   9 +-
 gtk/bio-gio.c| 135 +
 gtk/bio-gio.h|  34 
 gtk/bio-gsocket.c| 111 ---
 gtk/bio-gsocket.h|  30 ---
 gtk/glib-compat.c|  34 
 gtk/glib-compat.h|   8 +
 gtk/map-file |  14 ++
 gtk/spice-channel-priv.h |   5 +-
 gtk/spice-channel.c  | 137 -
 gtk/spice-channel.h  |   2 +
 gtk/spice-client.h   |   1 +
 gtk/spice-glib-sym-file  |  14 ++
 gtk/spice-proxy.c| 270 --
 gtk/spice-proxy.h|  60 --
 gtk/spice-session-priv.h |   5 +-
 gtk/spice-session.c  |  50 +++--
 gtk/spice-session.h  |   2 +
 gtk/spice-uri-priv.h |  30 +++
 gtk/spice-uri.c  | 360 +++
 gtk/spice-uri.h  |  52 +
 gtk/spicy.c  |   6 +
 gtk/wocky-http-proxy.c   | 154 +--
 gtk/wocky-http-proxy.h   |  14 ++
 28 files changed, 1006 insertions(+), 563 deletions(-)
 create mode 100644 gtk/bio-gio.c
 create mode 100644 gtk/bio-gio.h
 delete mode 100644 gtk/bio-gsocket.c
 delete mode 100644 gtk/bio-gsocket.h
 delete mode 100644 gtk/spice-proxy.c
 delete mode 100644 gtk/spice-proxy.h
 create mode 100644 gtk/spice-uri-priv.h
 create mode 100644 gtk/spice-uri.c
 create mode 100644 gtk/spice-uri.h

-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 01/14] compat: add strtok_r fallback

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

The following Spice proxy URI parsing code makes use of it, but it is
not available on Windows

Origin:
http://git.videolan.org/gitweb.cgi/vlc.git/?p=vlc.git;a=blob;f=compat/strtok_r.c
---
 configure.ac  |  2 +-
 gtk/glib-compat.c | 34 ++
 gtk/glib-compat.h |  8 
 3 files changed, 43 insertions(+), 1 deletion(-)

diff --git a/configure.ac b/configure.ac
index 4dcd81c..3670863 100644
--- a/configure.ac
+++ b/configure.ac
@@ -235,7 +235,7 @@ else
 EXTERNAL_PNP_IDS=$with_pnp_ids_path
 fi
 
-AC_CHECK_FUNCS(clearenv)
+AC_CHECK_FUNCS(clearenv strtok_r)
 
 PKG_CHECK_MODULES(GLIB2, glib-2.0 = 2.22)
 AC_SUBST(GLIB2_CFLAGS)
diff --git a/gtk/glib-compat.c b/gtk/glib-compat.c
index 9ffadbb..5714173 100644
--- a/gtk/glib-compat.c
+++ b/gtk/glib-compat.c
@@ -1,5 +1,8 @@
 /* -*- Mode: C; c-basic-offset: 4; indent-tabs-mode: nil -*- */
 /*
+   Copyright (C) 2012-2014 Red Hat, Inc.
+   Copyright © 1998-2009 VLC authors and VideoLAN
+
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
@@ -14,6 +17,7 @@
License along with this library; if not, see http://www.gnu.org/licenses/.
 */
 
+#include string.h
 #include glib-compat.h
 
 #if !GLIB_CHECK_VERSION(2,26,0)
@@ -105,3 +109,33 @@ guint64 g_get_monotonic_time(void)
 return (((gint64) tv.tv_sec) * 100) + tv.tv_usec;
 }
 #endif
+
+#ifndef HAVE_STRTOK_R
+G_GNUC_INTERNAL
+char *strtok_r(char *s, const char *delim, char **save_ptr)
+{
+char *token;
+
+if (s == NULL)
+s = *save_ptr;
+
+/* Scan leading delimiters. */
+s += strspn (s, delim);
+if (*s == '\0')
+return NULL;
+
+/* Find the end of the token. */
+token = s;
+s = strpbrk (token, delim);
+if (s == NULL)
+/* This token finishes the string. */
+*save_ptr = strchr (token, '\0');
+else
+{
+/* Terminate the token and make *SAVE_PTR point past it. */
+*s = '\0';
+*save_ptr = s + 1;
+}
+return token;
+}
+#endif
diff --git a/gtk/glib-compat.h b/gtk/glib-compat.h
index c30a735..62580e2 100644
--- a/gtk/glib-compat.h
+++ b/gtk/glib-compat.h
@@ -1,5 +1,8 @@
 /* -*- Mode: C; c-basic-offset: 4; indent-tabs-mode: nil -*- */
 /*
+   Copyright (C) 2012-2014 Red Hat, Inc.
+   Copyright © 1998-2009 VLC authors and VideoLAN
+
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
@@ -18,6 +21,7 @@
 
 #include glib-object.h
 #include gio/gio.h
+#include config.h
 
 #if !GLIB_CHECK_VERSION(2,26,0)
 #define G_DEFINE_BOXED_TYPE(TypeName, type_name, copy_func, free_func) 
G_DEFINE_BOXED_TYPE_WITH_CODE (TypeName, type_name, copy_func, free_func, {})
@@ -133,4 +137,8 @@ GType spice_main_context_get_type (void) G_GNUC_CONST;
 guint64 g_get_monotonic_time(void);
 #endif
 
+#ifndef HAVE_STRTOK_R
+char* strtok_r(char *s, const char *delim, char **save_ptr);
+#endif
+
 #endif /* GLIB_COMPAT_H */
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 02/14] proxy: split uri with : in only 2 parts

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

We want just host:port here.
---
 gtk/spice-proxy.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/gtk/spice-proxy.c b/gtk/spice-proxy.c
index bc4037e..f338213 100644
--- a/gtk/spice-proxy.c
+++ b/gtk/spice-proxy.c
@@ -74,7 +74,8 @@ gboolean spice_proxy_parse(SpiceProxy *self, const gchar 
*proxyuri, GError **err
 spice_proxy_set_protocol(self, http);
 spice_proxy_set_port(self, 3128);
 
-gchar **proxyv = g_strsplit(uri, :, 0);
+/* max 2 parts, host:port */
+gchar **proxyv = g_strsplit(uri, :, 2);
 const gchar *proxy_port = NULL;
 
 if (proxyv[0] == NULL || strlen(proxyv[0]) == 0) {
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 05/14] http-proxy: specify Basic scheme

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

Or Squid will fail with:
WARNING: Unsupported or unconfigured/inactive proxy-auth scheme
---
 gtk/wocky-http-proxy.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gtk/wocky-http-proxy.c b/gtk/wocky-http-proxy.c
index d73990b..11f557e 100644
--- a/gtk/wocky-http-proxy.c
+++ b/gtk/wocky-http-proxy.c
@@ -98,7 +98,7 @@ create_request (GProxyAddress *proxy_address, gboolean 
*has_cred)
   base64_cred = g_base64_encode ((guchar *) cred, strlen (cred));
   g_free (cred);
   g_string_append_printf (request,
-  Proxy-Authorization: %s\r\n,
+  Proxy-Authorization: Basic %s\r\n,
   base64_cred);
   g_free (base64_cred);
 }
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 06/14] spice-proxy: parse https protocol

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

---
 gtk/spice-proxy.c | 14 ++
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/gtk/spice-proxy.c b/gtk/spice-proxy.c
index 35eec67..685c120 100644
--- a/gtk/spice-proxy.c
+++ b/gtk/spice-proxy.c
@@ -65,9 +65,17 @@ gboolean spice_proxy_parse(SpiceProxy *self, const gchar 
*proxyuri, GError **err
 uri = dup = g_strdup(proxyuri);
 /* FIXME: use GUri when it is ready... only support http atm */
 /* the code is voluntarily not parsing thoroughly the uri */
-if (g_ascii_strncasecmp(http://;, uri, 7) == 0)
+if (g_ascii_strncasecmp(http://;, uri, 7) == 0) {
 uri += 7;
-
+spice_proxy_set_protocol(self, http);
+spice_proxy_set_port(self, 3128);
+} else if (g_ascii_strncasecmp(https://;, uri, 8) == 0) {
+uri += 8;
+spice_proxy_set_protocol(self, https);
+spice_proxy_set_port(self, 3129);
+} else {
+return FALSE;
+}
 /* remove trailing slash */
 len = strlen(uri);
 for (; len  0; len--)
@@ -76,8 +84,6 @@ gboolean spice_proxy_parse(SpiceProxy *self, const gchar 
*proxyuri, GError **err
 else
 break;
 
-spice_proxy_set_protocol(self, http);
-spice_proxy_set_port(self, 3128);
 
 /* yes, that parser is bad, we need GUri... */
 if (strstr(uri, @)) {
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 04/14] proxy: parse user and pass from uri

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

---
 gtk/spice-proxy.c | 13 +
 1 file changed, 13 insertions(+)

diff --git a/gtk/spice-proxy.c b/gtk/spice-proxy.c
index 7711e83..35eec67 100644
--- a/gtk/spice-proxy.c
+++ b/gtk/spice-proxy.c
@@ -19,6 +19,7 @@
 #include stdlib.h
 #include string.h
 
+#include glib-compat.h
 #include spice-client.h
 #include spice-proxy.h
 
@@ -78,6 +79,18 @@ gboolean spice_proxy_parse(SpiceProxy *self, const gchar 
*proxyuri, GError **err
 spice_proxy_set_protocol(self, http);
 spice_proxy_set_port(self, 3128);
 
+/* yes, that parser is bad, we need GUri... */
+if (strstr(uri, @)) {
+gchar *saveptr, *saveptr2;
+gchar *next = strstr(uri, @) + 1;
+gchar *auth = strtok_r(uri, @, saveptr);
+const gchar *user = strtok_r(auth, :, saveptr2);
+const gchar *pass = strtok_r(NULL, :, saveptr2);
+spice_proxy_set_user(self, user);
+spice_proxy_set_password(self, pass);
+uri = next;
+}
+
 /* max 2 parts, host:port */
 gchar **proxyv = g_strsplit(uri, :, 2);
 const gchar *proxy_port = NULL;
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 09/14] openssl: learn to handle a new kind of BIO based on GIOStream

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

Although reusing BIO_new_socket() once again is a hack, it seems
to be the easiest way... The proper solution is certainly to start
using GTls instead, but that will require a glib 2.28 dep bump.
---
 gtk/Makefile.am |   4 +-
 gtk/bio-gio.c   | 135 
 gtk/bio-gio.h   |  34 +
 gtk/bio-gsocket.c   | 111 --
 gtk/bio-gsocket.h   |  30 
 gtk/spice-channel.c |   2 +-
 6 files changed, 172 insertions(+), 144 deletions(-)
 create mode 100644 gtk/bio-gio.c
 create mode 100644 gtk/bio-gio.h
 delete mode 100644 gtk/bio-gsocket.c
 delete mode 100644 gtk/bio-gsocket.h

diff --git a/gtk/Makefile.am b/gtk/Makefile.am
index a1a1e79..0740e96 100644
--- a/gtk/Makefile.am
+++ b/gtk/Makefile.am
@@ -212,8 +212,8 @@ USB_ACL_HELPER_SRCS =
 endif
 
 libspice_client_glib_2_0_la_SOURCES =  \
-   bio-gsocket.c   \
-   bio-gsocket.h   \
+   bio-gio.c   \
+   bio-gio.h   \
glib-compat.c   \
glib-compat.h   \
spice-audio.c   \
diff --git a/gtk/bio-gio.c b/gtk/bio-gio.c
new file mode 100644
index 000..22d58b6
--- /dev/null
+++ b/gtk/bio-gio.c
@@ -0,0 +1,135 @@
+/* -*- Mode: C; c-basic-offset: 4; indent-tabs-mode: nil -*- */
+/*
+   Copyright (C) 2012 Red Hat, Inc.
+
+   This library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   This library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with this library; if not, see http://www.gnu.org/licenses/.
+*/
+
+#include string.h
+#include glib.h
+
+#include spice-util.h
+#include bio-gio.h
+
+typedef struct bio_gsocket_method {
+BIO_METHOD method;
+#if GLIB_CHECK_VERSION(2, 28, 0)
+GIOStream *stream;
+#else
+GSocket *gsocket;
+#endif
+} bio_gsocket_method;
+
+#define BIO_GET_GSOCKET(bio)  (((bio_gsocket_method*)bio-method)-gsocket)
+#define BIO_GET_ISTREAM(bio)  
(g_io_stream_get_input_stream(((bio_gsocket_method*)bio-method)-stream))
+#define BIO_GET_OSTREAM(bio)  
(g_io_stream_get_output_stream(((bio_gsocket_method*)bio-method)-stream))
+
+static int bio_gio_write(BIO *bio, const char *in, int inl)
+{
+gssize ret;
+GError *error = NULL;
+
+#if GLIB_CHECK_VERSION(2, 28, 0)
+ret = 
g_pollable_output_stream_write_nonblocking(G_POLLABLE_OUTPUT_STREAM(BIO_GET_OSTREAM(bio)),
+#else
+ret = g_socket_send(BIO_GET_GSOCKET(bio),
+#endif
+in, inl, NULL, error);
+BIO_clear_retry_flags(bio);
+
+if (g_error_matches(error, G_IO_ERROR, G_IO_ERROR_WOULD_BLOCK))
+BIO_set_retry_write(bio);
+if (error != NULL) {
+g_warning(%s, error-message);
+g_clear_error(error);
+}
+
+return ret;
+}
+
+static int bio_gio_read(BIO *bio, char *out, int outl)
+{
+gssize ret;
+GError *error = NULL;
+
+#if GLIB_CHECK_VERSION(2, 28, 0)
+ret = 
g_pollable_input_stream_read_nonblocking(G_POLLABLE_INPUT_STREAM(BIO_GET_ISTREAM(bio)),
+#else
+ret = g_socket_receive(BIO_GET_GSOCKET(bio),
+#endif
+   out, outl, NULL, error);
+BIO_clear_retry_flags(bio);
+
+if (g_error_matches(error, G_IO_ERROR, G_IO_ERROR_WOULD_BLOCK))
+BIO_set_retry_read(bio);
+else if (error != NULL)
+g_warning(%s, error-message);
+
+g_clear_error(error);
+
+return ret;
+}
+
+static int bio_gio_destroy(BIO *bio)
+{
+if (bio == NULL || bio-method == NULL)
+return 0;
+
+SPICE_DEBUG(bio gsocket destroy);
+g_free(bio-method);
+bio-method = NULL;;
+
+return 1;
+}
+
+static int bio_gio_puts(BIO *bio, const char *str)
+{
+int n, ret;
+
+n = strlen(str);
+ret = bio_gio_write(bio, str, n);
+
+return ret;
+}
+
+G_GNUC_INTERNAL
+#if GLIB_CHECK_VERSION(2, 28, 0)
+BIO* bio_new_giostream(GIOStream *stream)
+{
+// TODO: make an actual new BIO type, or just switch to GTls already...
+BIO *bio = BIO_new_socket(-1, BIO_NOCLOSE);
+#else
+BIO* bio_new_gsocket(GSocket *gsocket)
+{
+BIO *bio = BIO_new_socket(g_socket_get_fd(gsocket), BIO_NOCLOSE);
+#endif
+
+bio_gsocket_method *bio_method = g_new(bio_gsocket_method, 1);
+bio_method-method = *bio-method;
+#if GLIB_CHECK_VERSION(2, 28, 0)
+bio_method-stream = 

[Spice-devel] [PATCH spice-gtk 07/14] http-proxy: add https proxy

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

This will require glib 2.28 for GTls support, atm
---
 gtk/spice-session.c|   3 +
 gtk/wocky-http-proxy.c | 152 ++---
 gtk/wocky-http-proxy.h |  14 +
 3 files changed, 150 insertions(+), 19 deletions(-)

diff --git a/gtk/spice-session.c b/gtk/spice-session.c
index 5d1c1c1..475306a 100644
--- a/gtk/spice-session.c
+++ b/gtk/spice-session.c
@@ -637,6 +637,9 @@ static void spice_session_class_init(SpiceSessionClass 
*klass)
 #if GLIB_CHECK_VERSION(2, 26, 0)
 _wocky_http_proxy_get_type();
 #endif
+#if GLIB_CHECK_VERSION(2, 28, 0)
+_wocky_https_proxy_get_type();
+#endif
 
 gobject_class-dispose  = spice_session_dispose;
 gobject_class-finalize = spice_session_finalize;
diff --git a/gtk/wocky-http-proxy.c b/gtk/wocky-http-proxy.c
index 11f557e..feb9816 100644
--- a/gtk/wocky-http-proxy.c
+++ b/gtk/wocky-http-proxy.c
@@ -1,7 +1,9 @@
  /* wocky-http-proxy.c: Source for WockyHttpProxy
  *
  * Copyright (C) 2010 Collabora, Ltd.
+ * Copyright (C) 2014 Red Hat, Inc.
  * @author Nicolas Dufresne nicolas.dufre...@collabora.co.uk
+ * @author Marc-André Lureau marcandre.lur...@redhat.com
  *
  * This library is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Lesser General Public
@@ -180,10 +182,34 @@ wocky_http_proxy_connect (GProxy *proxy,
 {
   GInputStream *in;
   GOutputStream *out;
-  GDataInputStream *data_in;
-  gchar *buffer;
+  GDataInputStream *data_in = NULL;
+  gchar *buffer = NULL;
   gboolean has_cred;
 
+#if GLIB_CHECK_VERSION(2, 28, 0)
+  if (WOCKY_IS_HTTPS_PROXY (proxy))
+{
+  GIOStream *tlsconn;
+
+  tlsconn = g_tls_client_connection_new (io_stream,
+ 
G_SOCKET_CONNECTABLE(proxy_address),
+ error);
+  if (!tlsconn)
+  goto error;
+
+  GTlsCertificateFlags tls_validation_flags = 
G_TLS_CERTIFICATE_VALIDATE_ALL;
+#ifdef DEBUG
+  tls_validation_flags = ~(G_TLS_CERTIFICATE_UNKNOWN_CA | 
G_TLS_CERTIFICATE_BAD_IDENTITY);
+#endif
+  g_tls_client_connection_set_validation_flags (G_TLS_CLIENT_CONNECTION 
(tlsconn),
+tls_validation_flags);
+  if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), 
cancellable, error))
+  goto error;
+
+  io_stream = tlsconn;
+}
+#endif
+
   in = g_io_stream_get_input_stream (io_stream);
   out = g_io_stream_get_output_stream (io_stream);
 
@@ -291,6 +317,41 @@ do_write (GAsyncReadyCallback callback, ConnectAsyncData 
*data)
 }
 
 static void
+stream_connected (ConnectAsyncData *data,
+  GIOStream *io_stream)
+{
+  GInputStream *in;
+
+  data-io_stream = g_object_ref (io_stream);
+  in = g_io_stream_get_input_stream (io_stream);
+  data-data_in = g_data_input_stream_new (in);
+  g_filter_input_stream_set_close_base_stream (G_FILTER_INPUT_STREAM 
(data-data_in),
+   FALSE);
+
+  do_write (request_write_cb, data);
+}
+
+#if GLIB_CHECK_VERSION(2, 28, 0)
+static void
+handshake_completed (GObject *source_object,
+ GAsyncResult *res,
+ gpointer user_data)
+{
+  GTlsConnection *conn = G_TLS_CONNECTION (source_object);
+  ConnectAsyncData *data = user_data;
+  GError *error = NULL;
+
+  if (!g_tls_connection_handshake_finish (conn, res, error))
+{
+  complete_async_from_error (data, error);
+  return;
+}
+
+  stream_connected (data, G_IO_STREAM (conn));
+}
+#endif
+
+static void
 wocky_http_proxy_connect_async (GProxy *proxy,
 GIOStream *io_stream,
 GProxyAddress *proxy_address,
@@ -300,34 +361,55 @@ wocky_http_proxy_connect_async (GProxy *proxy,
 {
   GSimpleAsyncResult *simple;
   ConnectAsyncData *data;
-  GInputStream *in;
 
   simple = g_simple_async_result_new (G_OBJECT (proxy),
-  callback, user_data,
-  wocky_http_proxy_connect_async);
+  callback, user_data,
+  wocky_http_proxy_connect_async);
 
   data = g_slice_new0 (ConnectAsyncData);
-
-  data-simple = simple;
-  data-io_stream = g_object_ref (io_stream);
-
   if (cancellable != NULL)
 data-cancellable = g_object_ref (cancellable);
-
-  in = g_io_stream_get_input_stream (io_stream);
-
-  data-data_in = g_data_input_stream_new (in);
-  g_filter_input_stream_set_close_base_stream (G_FILTER_INPUT_STREAM 
(data-data_in),
-  FALSE);
-
-  g_simple_async_result_set_op_res_gpointer (simple, data,
-  (GDestroyNotify) free_connect_data);
+  data-simple = simple;
 
   data-buffer = create_request (proxy_address, data-has_cred);
   data-length = strlen (data-buffer);
   data-offset = 0;
 
-  do_write (request_write_cb, data);
+  g_simple_async_result_set_op_res_gpointer (simple, data,
+ (GDestroyNotify) 

[Spice-devel] [PATCH spice-gtk 08/14] Fill g_proxy_address_new() with protocol, user and password

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

This way, the call might eventually support more proxy and
authentication.
---
 gtk/spice-session.c | 7 +--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/gtk/spice-session.c b/gtk/spice-session.c
index 475306a..975fd7b 100644
--- a/gtk/spice-session.c
+++ b/gtk/spice-session.c
@@ -1724,8 +1724,11 @@ static void proxy_lookup_ready(GObject *source_object, 
GAsyncResult *result,
 
 for (it = addresses; it != NULL; it = it-next) {
 address = g_proxy_address_new(G_INET_ADDRESS(it-data),
-  spice_proxy_get_port(open_host-proxy), 
http,
-  s-host, open_host-port, NULL, NULL);
+  spice_proxy_get_port(open_host-proxy),
+  
spice_proxy_get_protocol(open_host-proxy),
+  s-host, open_host-port,
+  spice_proxy_get_user(open_host-proxy),
+  
spice_proxy_get_password(open_host-proxy));
 if (address != NULL)
 break;
 }
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 10/14] channel: simplify has error code

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

Get rid of a superflous g_socket_condition_check().
---
 gtk/spice-channel.c | 28 +++-
 1 file changed, 11 insertions(+), 17 deletions(-)

diff --git a/gtk/spice-channel.c b/gtk/spice-channel.c
index d47f19f..7bcb0b8 100644
--- a/gtk/spice-channel.c
+++ b/gtk/spice-channel.c
@@ -2099,29 +2099,23 @@ static gboolean wait_migration(gpointer data)
 static gboolean spice_channel_iterate(SpiceChannel *channel)
 {
 SpiceChannelPrivate *c = channel-priv;
-GIOCondition ret;
 
 if (c-state == SPICE_CHANNEL_STATE_MIGRATING 
 !g_coroutine_condition_wait(c-coroutine, wait_migration, channel))
 CHANNEL_DEBUG(channel, migration wait cancelled);
 
-if (c-has_error) {
-CHANNEL_DEBUG(channel, channel has error, breaking loop);
-return FALSE;
-}
-
 /* flush any pending write and read */
-SPICE_CHANNEL_GET_CLASS(channel)-iterate_write(channel);
-SPICE_CHANNEL_GET_CLASS(channel)-iterate_read(channel);
-
-ret = g_socket_condition_check(c-sock, G_IO_IN | G_IO_ERR | G_IO_HUP);
-if (c-state  SPICE_CHANNEL_STATE_CONNECTING 
-ret  (G_IO_ERR|G_IO_HUP)) {
-SPICE_DEBUG(got socket error: %d, ret);
-emit_main_context(channel, SPICE_CHANNEL_EVENT,
-  c-state == SPICE_CHANNEL_STATE_READY ?
-  SPICE_CHANNEL_ERROR_IO : SPICE_CHANNEL_ERROR_LINK);
-c-has_error = TRUE;
+if (!c-has_error)
+SPICE_CHANNEL_GET_CLASS(channel)-iterate_write(channel);
+if (!c-has_error)
+SPICE_CHANNEL_GET_CLASS(channel)-iterate_read(channel);
+
+if (c-has_error) {
+CHANNEL_DEBUG(channel, channel got error);
+if (c-state  SPICE_CHANNEL_STATE_CONNECTING)
+emit_main_context(channel, SPICE_CHANNEL_EVENT,
+  c-state == SPICE_CHANNEL_STATE_READY ?
+  SPICE_CHANNEL_ERROR_IO : 
SPICE_CHANNEL_ERROR_LINK);
 return FALSE;
 }
 
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 11/14] channel: talk to giostream instead of gsocket

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

---
 gtk/spice-channel-priv.h |  4 ++-
 gtk/spice-channel.c  | 81 +++-
 2 files changed, 56 insertions(+), 29 deletions(-)

diff --git a/gtk/spice-channel-priv.h b/gtk/spice-channel-priv.h
index 35704ea..00893c3 100644
--- a/gtk/spice-channel-priv.h
+++ b/gtk/spice-channel-priv.h
@@ -80,8 +80,10 @@ struct _SpiceChannelPrivate {
 SSL_CTX *ctx;
 SSL *ssl;
 SpiceOpenSSLVerify  *sslverify;
-GSocket *sock;
+GSocket *socket;
 GSocketConnection   *conn;
+GInputStream*in;
+GOutputStream   *out;
 
 #if HAVE_SASL
 sasl_conn_t *sasl_conn;
diff --git a/gtk/spice-channel.c b/gtk/spice-channel.c
index 7bcb0b8..b1e049d 100644
--- a/gtk/spice-channel.c
+++ b/gtk/spice-channel.c
@@ -763,7 +763,8 @@ static void spice_channel_flush_wire(SpiceChannel *channel,
 GIOCondition cond;
 
 while (offset  datalen) {
-int ret;
+gssize ret;
+GError *error = NULL;
 
 if (c-has_error) return;
 
@@ -779,9 +780,12 @@ static void spice_channel_flush_wire(SpiceChannel *channel,
 ret = -1;
 }
 } else {
-GError *error = NULL;
-ret = g_socket_send(c-sock, ptr+offset, datalen-offset,
-NULL, error);
+#if GLIB_CHECK_VERSION(2, 28, 0)
+ret = 
g_pollable_output_stream_write_nonblocking(G_POLLABLE_OUTPUT_STREAM(c-out),
+ ptr+offset, 
datalen-offset, NULL, error);
+#else
+ret = g_socket_send(c-socket, ptr+offset, datalen-offset, NULL, 
error);
+#endif
 if (ret  0) {
 if (g_error_matches(error, G_IO_ERROR, 
G_IO_ERROR_WOULD_BLOCK)) {
 cond = G_IO_OUT;
@@ -794,7 +798,8 @@ static void spice_channel_flush_wire(SpiceChannel *channel,
 }
 if (ret == -1) {
 if (cond != 0) {
-g_coroutine_socket_wait(c-coroutine, c-sock, cond);
+// TODO: should use 
g_pollable_input/output_stream_create_source() in 2.28 ?
+g_coroutine_socket_wait(c-coroutine, c-socket, cond);
 continue;
 } else {
 CHANNEL_DEBUG(channel, Closing the channel: 
spice_channel_flush %d, errno);
@@ -888,7 +893,7 @@ static void spice_channel_write_msg(SpiceChannel *channel, 
SpiceMsgOut *out)
 static int spice_channel_read_wire(SpiceChannel *channel, void *data, size_t 
len)
 {
 SpiceChannelPrivate *c = channel-priv;
-int ret;
+gssize ret;
 GIOCondition cond;
 
 reread:
@@ -908,7 +913,13 @@ reread:
 }
 } else {
 GError *error = NULL;
-ret = g_socket_receive(c-sock, data, len, NULL, error);
+#if GLIB_CHECK_VERSION(2, 28, 0)
+ret = 
g_pollable_input_stream_read_nonblocking(G_POLLABLE_INPUT_STREAM(c-in),
+   data, len, NULL, 
error);
+#else
+ret = g_socket_receive(c-socket,
+   data, len, NULL, error);
+#endif
 if (ret  0) {
 if (g_error_matches(error, G_IO_ERROR, G_IO_ERROR_WOULD_BLOCK)) {
 cond = G_IO_IN;
@@ -922,7 +933,8 @@ reread:
 
 if (ret == -1) {
 if (cond != 0) {
-g_coroutine_socket_wait(c-coroutine, c-sock, cond);
+// TODO: should use g_pollable_input/output_stream_create_source() 
?
+g_coroutine_socket_wait(c-coroutine, c-socket, cond);
 goto reread;
 } else {
 c-has_error = TRUE;
@@ -1351,7 +1363,7 @@ static gboolean 
spice_channel_perform_auth_sasl(SpiceChannel *channel)
 }
 
 /* Get local address in form  IPADDR:PORT */
-addr = g_socket_get_local_address(c-sock, NULL);
+addr = g_socket_get_local_address(c-socket, NULL);
 if (!addr) {
 g_critical(failed to get local address);
 goto error;
@@ -1363,7 +1375,7 @@ static gboolean 
spice_channel_perform_auth_sasl(SpiceChannel *channel)
 g_clear_object(addr);
 
 /* Get remote address in form  IPADDR:PORT */
-addr = g_socket_get_remote_address(c-sock, NULL);
+addr = g_socket_get_remote_address(c-socket, NULL);
 if (!addr) {
 g_critical(failed to get peer address);
 goto error;
@@ -2062,14 +2074,17 @@ static void spice_channel_iterate_read(SpiceChannel 
*channel)
 {
 SpiceChannelPrivate *c = channel-priv;
 
-g_coroutine_socket_wait(c-coroutine, c-sock, G_IO_IN);
+g_coroutine_socket_wait(c-coroutine, c-socket, G_IO_IN);
 
 /* treat all incoming data (block on message completion) */
 while (!c-has_error 
c-state != SPICE_CHANNEL_STATE_MIGRATING 
-   g_socket_condition_check(c-sock, G_IO_IN)  G_IO_IN) {
-
-do
+#if 

[Spice-devel] [PATCH spice-gtk 13/14] channel: add spice_channel_get_error()

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

Add a function to retrieve the last GError from a channel, this may be
useful to provide additional error details to the client.
---
 doc/reference/spice-gtk-sections.txt |  1 +
 gtk/map-file |  1 +
 gtk/spice-channel-priv.h |  1 +
 gtk/spice-channel.c  | 26 --
 gtk/spice-channel.h  |  2 ++
 gtk/spice-glib-sym-file  |  1 +
 gtk/spice-session-priv.h |  2 +-
 gtk/spice-session.c  |  6 +++---
 gtk/spicy.c  |  6 ++
 9 files changed, 40 insertions(+), 6 deletions(-)

diff --git a/doc/reference/spice-gtk-sections.txt 
b/doc/reference/spice-gtk-sections.txt
index 411ca0e..9f0cf67 100644
--- a/doc/reference/spice-gtk-sections.txt
+++ b/doc/reference/spice-gtk-sections.txt
@@ -102,6 +102,7 @@ spice_channel_string_to_type
 spice_channel_set_capability
 spice_channel_flush_async
 spice_channel_flush_finish
+spice_channel_get_error
 SUBSECTION Standard
 SPICE_TYPE_CHANNEL_EVENT
 spice_channel_event_get_type
diff --git a/gtk/map-file b/gtk/map-file
index 82a10eb..0067960 100644
--- a/gtk/map-file
+++ b/gtk/map-file
@@ -7,6 +7,7 @@ spice_channel_connect;
 spice_channel_destroy;
 spice_channel_disconnect;
 spice_channel_event_get_type;
+spice_channel_get_error;
 spice_channel_get_type;
 spice_channel_new;
 spice_channel_open_fd;
diff --git a/gtk/spice-channel-priv.h b/gtk/spice-channel-priv.h
index 00893c3..fc15c0e 100644
--- a/gtk/spice-channel-priv.h
+++ b/gtk/spice-channel-priv.h
@@ -136,6 +136,7 @@ struct _SpiceChannelPrivate {
 GSList  *flushing;
 
 gbooleandisable_channel_msg;
+GError  *error;
 };
 
 SpiceMsgIn *spice_msg_in_new(SpiceChannel *channel);
diff --git a/gtk/spice-channel.c b/gtk/spice-channel.c
index b1e049d..13992c1 100644
--- a/gtk/spice-channel.c
+++ b/gtk/spice-channel.c
@@ -154,6 +154,8 @@ static void spice_channel_dispose(GObject *gobject)
  c-session = NULL;
 }
 
+g_clear_error(c-error);
+
 /* Chain up to the parent class */
 if (G_OBJECT_CLASS(spice_channel_parent_class)-dispose)
 G_OBJECT_CLASS(spice_channel_parent_class)-dispose(gobject);
@@ -2220,6 +,25 @@ static int spice_channel_load_ca(SpiceChannel *channel)
 return count;
 }
 
+/**
+ * spice_channel_get_error:
+ * @channel:
+ *
+ * Retrieves the #GError currently set on channel, if the #SpiceChannel
+ * is in error state.
+ *
+ * Returns: the pointer to the error, or NULL
+ * Since: 0.24
+ **/
+const GError* spice_channel_get_error(SpiceChannel *self)
+{
+SpiceChannelPrivate *c;
+
+g_return_val_if_fail(SPICE_IS_CHANNEL(self), NULL);
+c = self-priv;
+
+return c-error;
+}
 
 /* coroutine context */
 static void *spice_channel_coroutine(void *data)
@@ -2257,15 +2278,16 @@ static void *spice_channel_coroutine(void *data)
 
 
 reconnect:
-c-conn = spice_session_channel_open_host(c-session, channel, c-tls);
+c-conn = spice_session_channel_open_host(c-session, channel, c-tls, 
c-error);
 if (c-conn == NULL) {
-if (!c-tls) {
+if (!c-error  !c-tls) {
 CHANNEL_DEBUG(channel, trying with TLS port);
 c-tls = true; /* FIXME: does that really work with provided fd */
 goto reconnect;
 } else {
 CHANNEL_DEBUG(channel, Connect error);
 emit_main_context(channel, SPICE_CHANNEL_EVENT, 
SPICE_CHANNEL_ERROR_CONNECT);
+g_clear_error(c-error);
 goto cleanup;
 }
 }
diff --git a/gtk/spice-channel.h b/gtk/spice-channel.h
index 705dddf..1c303b4 100644
--- a/gtk/spice-channel.h
+++ b/gtk/spice-channel.h
@@ -123,6 +123,8 @@ void spice_channel_set_capability(SpiceChannel *channel, 
guint32 cap);
 const gchar* spice_channel_type_to_string(gint type);
 gint spice_channel_string_to_type(const gchar *str);
 
+const GError* spice_channel_get_error(SpiceChannel *channel);
+
 G_END_DECLS
 
 #endif /* __SPICE_CLIENT_CHANNEL_H__ */
diff --git a/gtk/spice-glib-sym-file b/gtk/spice-glib-sym-file
index 3dc709a..9d5001a 100644
--- a/gtk/spice-glib-sym-file
+++ b/gtk/spice-glib-sym-file
@@ -7,6 +7,7 @@ spice_channel_disconnect
 spice_channel_event_get_type
 spice_channel_flush_async
 spice_channel_flush_finish
+spice_channel_get_error
 spice_channel_get_type
 spice_channel_new
 spice_channel_open_fd
diff --git a/gtk/spice-session-priv.h b/gtk/spice-session-priv.h
index 85a1dad..cf9f9d1 100644
--- a/gtk/spice-session-priv.h
+++ b/gtk/spice-session-priv.h
@@ -119,7 +119,7 @@ int spice_session_get_connection_id(SpiceSession *session);
 gboolean spice_session_get_client_provided_socket(SpiceSession *session);
 
 GSocketConnection* spice_session_channel_open_host(SpiceSession *session, 
SpiceChannel *channel,
-   gboolean *use_tls);
+ 

[Spice-devel] [PATCH spice-gtk 12/14] Make SpiceURI a public API

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

Generalize a little bit SpiceProxy to allow easy URI manipulation by
clients.
---
 doc/reference/Makefile.am|   1 +
 doc/reference/spice-gtk-docs.xml |   1 +
 doc/reference/spice-gtk-sections.txt |  26 +++
 gtk/Makefile.am  |   5 +-
 gtk/map-file |  12 ++
 gtk/spice-client.h   |   1 +
 gtk/spice-glib-sym-file  |  12 ++
 gtk/spice-proxy.c| 362 ---
 gtk/spice-proxy.h|  64 ---
 gtk/spice-session-priv.h |   3 +-
 gtk/spice-session.c  |  28 +--
 gtk/spice-session.h  |   1 +
 gtk/spice-uri-priv.h |  30 +++
 gtk/spice-uri.c  | 360 ++
 gtk/spice-uri.h  |  52 +
 15 files changed, 514 insertions(+), 444 deletions(-)
 delete mode 100644 gtk/spice-proxy.c
 delete mode 100644 gtk/spice-proxy.h
 create mode 100644 gtk/spice-uri-priv.h
 create mode 100644 gtk/spice-uri.c
 create mode 100644 gtk/spice-uri.h

diff --git a/doc/reference/Makefile.am b/doc/reference/Makefile.am
index a0a856c..76c7d34 100644
--- a/doc/reference/Makefile.am
+++ b/doc/reference/Makefile.am
@@ -44,6 +44,7 @@ IGNORE_HFILES=\
spice-marshal.h \
spice-pulse.h   \
spice-session-priv.h\
+   spice-uri-priv.h\
spice-util-priv.h   \
spice-widget-priv.h \
usb-acl-helper.h\
diff --git a/doc/reference/spice-gtk-docs.xml b/doc/reference/spice-gtk-docs.xml
index 4a9a3cf..d2c1a2b 100644
--- a/doc/reference/spice-gtk-docs.xml
+++ b/doc/reference/spice-gtk-docs.xml
@@ -52,6 +52,7 @@
   xi:include href=xml/smartcard-manager.xml/
   xi:include href=xml/usb-device-manager.xml/
   xi:include href=xml/spice-util.xml/
+  xi:include href=xml/spice-uri.xml/
 /chapter
 
   /part
diff --git a/doc/reference/spice-gtk-sections.txt 
b/doc/reference/spice-gtk-sections.txt
index 8d61aa9..411ca0e 100644
--- a/doc/reference/spice-gtk-sections.txt
+++ b/doc/reference/spice-gtk-sections.txt
@@ -429,3 +429,29 @@ SPICE_PORT_CHANNEL_GET_CLASS
 SpicePortChannelPrivate
 /SECTION
 
+SECTION
+FILEspice-uri/FILE
+spice_uri_get_scheme
+spice_uri_set_scheme
+spice_uri_get_hostname
+spice_uri_set_hostname
+spice_uri_get_port
+spice_uri_set_port
+spice_uri_get_user
+spice_uri_set_user
+spice_uri_get_password
+spice_uri_set_password
+spice_uri_to_string
+SpiceURIClass
+SpiceURI
+SUBSECTION Standard
+SPICE_IS_URI
+SPICE_IS_URI_CLASS
+SPICE_TYPE_URI
+SPICE_URI
+SPICE_URI_CLASS
+SPICE_URI_GET_CLASS
+spice_uri_get_type
+SUBSECTION Private
+SpiceURIPrivate
+/SECTION
diff --git a/gtk/Makefile.am b/gtk/Makefile.am
index 0740e96..dc09d8c 100644
--- a/gtk/Makefile.am
+++ b/gtk/Makefile.am
@@ -251,8 +251,8 @@ libspice_client_glib_2_0_la_SOURCES =   
\
channel-usbredir-priv.h \
smartcard-manager.c \
smartcard-manager-priv.h\
-   spice-proxy.c   \
-   spice-proxy.h   \
+   spice-uri.c \
+   spice-uri-priv.h\
usb-device-manager.c\
usb-device-manager-priv.h   \
usbutil.c   \
@@ -287,6 +287,7 @@ libspice_client_glibincludedir = 
$(includedir)/spice-client-glib-2.0
 libspice_client_glibinclude_HEADERS =  \
spice-audio.h   \
spice-client.h  \
+   spice-uri.h \
spice-types.h   \
spice-session.h \
spice-channel.h \
diff --git a/gtk/map-file b/gtk/map-file
index b6c184d..82a10eb 100644
--- a/gtk/map-file
+++ b/gtk/map-file
@@ -119,6 +119,18 @@ spice_util_get_version_string;
 spice_util_set_debug;
 spice_uuid_to_string;
 spice_webdav_channel_get_type;
+spice_uri_get_hostname;
+spice_uri_get_password;
+spice_uri_get_port;
+spice_uri_get_scheme;
+spice_uri_get_type;
+spice_uri_get_user;
+spice_uri_set_hostname;
+spice_uri_set_password;
+spice_uri_set_port;
+spice_uri_set_scheme;
+spice_uri_set_user;
+spice_uri_to_string;
 local:
 *;
 };
diff --git a/gtk/spice-client.h b/gtk/spice-client.h
index f5c8aa7..1f80419 100644
--- a/gtk/spice-client.h
+++ b/gtk/spice-client.h
@@ -31,6 +31,7 @@
 #include spice-session.h
 #include spice-channel.h
 #include spice-option.h
+#include spice-uri.h
 
 #include channel-main.h
 #include channel-display.h
diff --git a/gtk/spice-glib-sym-file b/gtk/spice-glib-sym-file

[Spice-devel] [PATCH spice-gtk 14/14] session: add spice_session_get_proxy_uri()

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

Learn to return the currently configured proxy, to allow
client to tweak parameters, such as username and password.
---
 doc/reference/spice-gtk-sections.txt |  1 +
 gtk/map-file |  1 +
 gtk/spice-glib-sym-file  |  1 +
 gtk/spice-session.c  | 14 ++
 gtk/spice-session.h  |  1 +
 5 files changed, 18 insertions(+)

diff --git a/doc/reference/spice-gtk-sections.txt 
b/doc/reference/spice-gtk-sections.txt
index 9f0cf67..08b1b4e 100644
--- a/doc/reference/spice-gtk-sections.txt
+++ b/doc/reference/spice-gtk-sections.txt
@@ -28,6 +28,7 @@ spice_session_disconnect
 spice_session_get_channels
 spice_session_get_read_only
 spice_session_has_channel_type
+spice_session_get_proxy_uri
 SUBSECTION
 SpiceSessionMigration
 SpiceSessionVerify
diff --git a/gtk/map-file b/gtk/map-file
index 0067960..2a46159 100644
--- a/gtk/map-file
+++ b/gtk/map-file
@@ -90,6 +90,7 @@ spice_session_migration_get_type;
 spice_session_new;
 spice_session_open_fd;
 spice_session_verify_get_type;
+spice_session_get_proxy_uri;
 spice_set_session_option;
 spice_smartcard_channel_get_type;
 spice_smartcard_manager_get;
diff --git a/gtk/spice-glib-sym-file b/gtk/spice-glib-sym-file
index 9d5001a..20f3885 100644
--- a/gtk/spice-glib-sym-file
+++ b/gtk/spice-glib-sym-file
@@ -106,3 +106,4 @@ spice_uri_set_port
 spice_uri_set_scheme
 spice_uri_set_user
 spice_uri_to_string
+spice_session_get_proxy_uri
diff --git a/gtk/spice-session.c b/gtk/spice-session.c
index 5e76b5f..ea32cf7 100644
--- a/gtk/spice-session.c
+++ b/gtk/spice-session.c
@@ -2190,3 +2190,17 @@ guint32 spice_session_get_playback_latency(SpiceSession 
*session)
 return 0;
 }
 }
+
+/**
+ * spice_session_get_proxy_uri:
+ * @session: a #SpiceSession
+ *
+ * Returns: (transfer none): the session proxy #SpiceURI or %NULL.
+ * Since: 0.24
+ **/
+SpiceURI *spice_session_get_proxy_uri(SpiceSession *session)
+{
+SpiceSessionPrivate *s = SPICE_SESSION_GET_PRIVATE(session);
+
+return s-proxy;
+}
diff --git a/gtk/spice-session.h b/gtk/spice-session.h
index 4ea645e..665c609 100644
--- a/gtk/spice-session.h
+++ b/gtk/spice-session.h
@@ -93,6 +93,7 @@ void spice_session_disconnect(SpiceSession *session);
 GList *spice_session_get_channels(SpiceSession *session);
 gboolean spice_session_has_channel_type(SpiceSession *session, gint type);
 gboolean spice_session_get_read_only(SpiceSession *session);
+SpiceURI *spice_session_get_proxy_uri(SpiceSession *session);
 
 G_END_DECLS
 
-- 
1.8.4.2

___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] [PATCH spice-gtk 03/14] proxy: add user and pass properties

[Marc-André Lureau]

From: Marc-André Lureau marcandre.lur...@redhat.com

---
 gtk/spice-proxy.c | 72 +++
 gtk/spice-proxy.h |  4 
 2 files changed, 76 insertions(+)

diff --git a/gtk/spice-proxy.c b/gtk/spice-proxy.c
index f338213..7711e83 100644
--- a/gtk/spice-proxy.c
+++ b/gtk/spice-proxy.c
@@ -26,6 +26,8 @@ struct _SpiceProxyPrivate {
 gchar *protocol;
 gchar *hostname;
 guint port;
+gchar *user;
+gchar *password;
 };
 
 #define SPICE_PROXY_GET_PRIVATE(o) (G_TYPE_INSTANCE_GET_PRIVATE((o), 
SPICE_TYPE_PROXY, SpiceProxyPrivate))
@@ -35,6 +37,8 @@ G_DEFINE_TYPE(SpiceProxy, spice_proxy, G_TYPE_OBJECT);
 enum  {
 SPICE_PROXY_DUMMY_PROPERTY,
 SPICE_PROXY_PROTOCOL,
+SPICE_PROXY_USER,
+SPICE_PROXY_PASSWORD,
 SPICE_PROXY_HOSTNAME,
 SPICE_PROXY_PORT
 };
@@ -173,6 +177,12 @@ static void spice_proxy_get_property(GObject *object, 
guint property_id,
 case SPICE_PROXY_PORT:
 g_value_set_uint(value, spice_proxy_get_port(self));
 break;
+case SPICE_PROXY_USER:
+g_value_set_string(value, spice_proxy_get_user(self));
+break;
+case SPICE_PROXY_PASSWORD:
+g_value_set_string(value, spice_proxy_get_password(self));
+break;
 default:
 G_OBJECT_WARN_INVALID_PROPERTY_ID(object, property_id, pspec);
 break;
@@ -193,6 +203,12 @@ static void spice_proxy_set_property(GObject *object, 
guint property_id,
 case SPICE_PROXY_HOSTNAME:
 spice_proxy_set_hostname(self, g_value_get_string(value));
 break;
+case SPICE_PROXY_USER:
+spice_proxy_set_user(self, g_value_get_string(value));
+break;
+case SPICE_PROXY_PASSWORD:
+spice_proxy_set_password(self, g_value_get_string(value));
+break;
 case SPICE_PROXY_PORT:
 spice_proxy_set_port(self, g_value_get_uint(value));
 break;
@@ -209,6 +225,8 @@ static void spice_proxy_finalize(GObject* obj)
 self = G_TYPE_CHECK_INSTANCE_CAST(obj, SPICE_TYPE_PROXY, SpiceProxy);
 g_free(self-priv-protocol);
 g_free(self-priv-hostname);
+g_free(self-priv-user);
+g_free(self-priv-password);
 
 G_OBJECT_CLASS (spice_proxy_parent_class)-finalize (obj);
 }
@@ -254,6 +272,24 @@ static void spice_proxy_class_init(SpiceProxyClass *klass)
0, G_MAXUINT, 0,
G_PARAM_STATIC_STRINGS |
G_PARAM_READWRITE));
+
+g_object_class_install_property(G_OBJECT_CLASS (klass),
+SPICE_PROXY_USER,
+g_param_spec_string (user,
+ user,
+ user,
+ NULL,
+ 
G_PARAM_STATIC_STRINGS |
+ G_PARAM_READWRITE));
+
+g_object_class_install_property(G_OBJECT_CLASS (klass),
+SPICE_PROXY_PASSWORD,
+g_param_spec_string (password,
+ password,
+ password,
+ NULL,
+ 
G_PARAM_STATIC_STRINGS |
+ G_PARAM_READWRITE));
 }
 
 G_GNUC_INTERNAL
@@ -269,3 +305,39 @@ gchar* spice_proxy_to_string(SpiceProxy* self)
 
 return g_strdup_printf(%s://%s:%u, p-protocol, p-hostname, p-port);
 }
+
+G_GNUC_INTERNAL
+const gchar* spice_proxy_get_user(SpiceProxy *self)
+{
+g_return_val_if_fail(SPICE_IS_PROXY(self), NULL);
+return self-priv-user;
+}
+
+
+G_GNUC_INTERNAL
+void spice_proxy_set_user(SpiceProxy *self, const gchar *value)
+{
+g_return_if_fail(SPICE_IS_PROXY(self));
+
+g_free(self-priv-user);
+self-priv-user = g_strdup(value);
+g_object_notify((GObject *)self, user);
+}
+
+G_GNUC_INTERNAL
+const gchar* spice_proxy_get_password(SpiceProxy *self)
+{
+g_return_val_if_fail(SPICE_IS_PROXY(self), NULL);
+return self-priv-password;
+}
+
+
+G_GNUC_INTERNAL
+void spice_proxy_set_password(SpiceProxy *self, const gchar *value)
+{
+g_return_if_fail(SPICE_IS_PROXY(self));
+
+g_free(self-priv-password);
+self-priv-password = g_strdup(value);
+g_object_notify((GObject *)self, password);
+}
diff --git a/gtk/spice-proxy.h b/gtk/spice-proxy.h
index 1e7b6d7..e74053b 100644
--- a/gtk/spice-proxy.h
+++ b/gtk/spice-proxy.h
@@ -54,6 +54,10 @@ void spice_proxy_set_hostname(SpiceProxy* self, const gchar* 
value);
 guint spice_proxy_get_port(SpiceProxy* self);
 void spice_proxy_set_port(SpiceProxy* self, guint port);
 gchar 

Re: [Spice-devel] guest agent windows binaries

[Lindsay Mathieson]

On Wed, 12 Feb 2014 09:16:04 AM Christophe Fergeau wrote:
 http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw3
 2-spice-vdagent-0.7.2-1.fc19.noarch.rpm and
 http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw6
 4-spice-vdagent-0.7.2-1.fc19.noarch.rpm contain the binaries:
 
 $ rpm -qpl ./mingw64-spice-vdagent-0.7.2-1.fc19.noarch.rpm
 /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdagent.exe
 /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdservice.exe

Why pack windows binaries in a rpm that can't be unpacked on windows?
Thanks, but I'll try replicating the binaries and service entries from a win 7 
installation.


 
  
  
 
  Also - could someone *please* clarify the current and future status of
  windows 8/2012 drivers and guest tools? its extremely confusing trying
  to track down this information.
 
 No QXL driver, the virtio-win drivers (located at
 https://alt.fedoraproject.org/pub/alt/virtio-win/latest/images/bin/ ) seem
 to have win8 versions

Couldn't get them to install myself.

 , and the agent should be working as well on win8.
 I've never tried these though.

I believe I have been labouring under the misapprehension that the qxl driver 
was necessary for the vdagent and all its goodies (display resizing, hot keys 
etc) to work, but that is not the case?

It will need the virtio driver though won't it?

I'll rebundle vdagent for win8/2012 and see how it goes.

Thanks,

-- 
Lindsay

signature.asc
Description: This is a digitally signed message part.
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 06/13] http-proxy: add https proxy

[Christophe Fergeau]

On Tue, Feb 11, 2014 at 12:37:53PM -0500, Marc-André Lureau wrote:
static void
wocky_http_proxy_init (WockyHttpProxy *proxy)
   @@ -180,10 +182,34 @@ wocky_http_proxy_connect (GProxy *proxy,
{
  GInputStream *in;
  GOutputStream *out;
   -  GDataInputStream *data_in;
   -  gchar *buffer;
   +  GDataInputStream *data_in = NULL;
   +  gchar *buffer = NULL;
  gboolean has_cred;

   +#if GLIB_CHECK_VERSION(2, 28, 0)
   +  if (WOCKY_IS_HTTPS_PROXY (proxy))
  
  Having runtime type checks in a interface virtual method is not very nice,
  having a wocky_https_proxy_iface_init would allow to call directly the
  right implementation without having to resort to runtime type checking.
 
 I decided to use this consciously, as doing it differently added much
 more boilerplate for no clear benefit.

Gave it a quick try (only compile-tested), and the resulting code is not
particularly bad. Having both inheritance and runtime type checks is imo
quite ugly. I can send a follow-up patch after your series land if it's likely
to be accepted.


diff --git a/gtk/wocky-http-proxy.c b/gtk/wocky-http-proxy.c
index 25f2af5..3eb6931 100644
--- a/gtk/wocky-http-proxy.c
+++ b/gtk/wocky-http-proxy.c
@@ -186,30 +186,6 @@ wocky_http_proxy_connect (GProxy *proxy,
   gchar *buffer = NULL;
   gboolean has_cred;
 
-#if GLIB_CHECK_VERSION(2, 28, 0)
-  if (WOCKY_IS_HTTPS_PROXY (proxy))
-{
-  GIOStream *tlsconn;
-
-  tlsconn = g_tls_client_connection_new (io_stream,
- 
G_SOCKET_CONNECTABLE(proxy_address),
- error);
-  if (!tlsconn)
-  goto error;
-
-  GTlsCertificateFlags tls_validation_flags = 
G_TLS_CERTIFICATE_VALIDATE_ALL;
-#ifdef DEBUG
-  tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
G_TLS_CERTIFICATE_BAD_IDENTITY;
-#endif
-  g_tls_client_connection_set_validation_flags (G_TLS_CLIENT_CONNECTION 
(tlsconn),
-tls_validation_flags);
-  if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), 
cancellable, error))
-  goto error;
-
-  io_stream = tlsconn;
-}
-#endif
-
   in = g_io_stream_get_input_stream (io_stream);
   out = g_io_stream_get_output_stream (io_stream);
 
@@ -251,6 +227,41 @@ error:
   return NULL;
 }
 
+#if GLIB_CHECK_VERSION(2, 28, 0)
+static GIOStream *
+wocky_https_proxy_connect (GProxy *proxy,
+   GIOStream *io_stream,
+   GProxyAddress *proxy_address,
+   GCancellable *cancellable,
+   GError **error)
+{
+{
+GIOStream *tlsconn;
+
+tlsconn = g_tls_client_connection_new (io_stream,
+   
G_SOCKET_CONNECTABLE(proxy_address),
+   error);
+if (!tlsconn)
+goto error;
+
+GTlsCertificateFlags tls_validation_flags = 
G_TLS_CERTIFICATE_VALIDATE_ALL;
+#ifdef DEBUG
+tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
G_TLS_CERTIFICATE_BAD_IDENTITY;
+#endif
+g_tls_client_connection_set_validation_flags (G_TLS_CLIENT_CONNECTION 
(tlsconn),
+  tls_validation_flags);
+if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), 
cancellable, error))
+goto error;
+
+io_stream = tlsconn;
+}
+return wocky_http_proxy_connect(proxy, io_stream, proxy_address, 
cancellable, error);
+
+error:
+return NULL;
+}
+#endif
+
 
 typedef struct
 {
@@ -351,8 +362,9 @@ handshake_completed (GObject *source_object,
 }
 #endif
 
-static void
-wocky_http_proxy_connect_async (GProxy *proxy,
+static ConnectAsyncData *
+wocky_http_proxy_connect_async_data_new (
+GProxy *proxy,
 GIOStream *io_stream,
 GProxyAddress *proxy_address,
 GCancellable *cancellable,
@@ -364,7 +376,7 @@ wocky_http_proxy_connect_async (GProxy *proxy,
 
   simple = g_simple_async_result_new (G_OBJECT (proxy),
   callback, user_data,
-  wocky_http_proxy_connect_async);
+  
wocky_http_proxy_connect_async_data_new);
 
   data = g_slice_new0 (ConnectAsyncData);
   if (cancellable != NULL)
@@ -378,12 +390,30 @@ wocky_http_proxy_connect_async (GProxy *proxy,
   g_simple_async_result_set_op_res_gpointer (simple, data,
  (GDestroyNotify) 
free_connect_data);
 
+  return data;
+}
+
+
 #if GLIB_CHECK_VERSION(2, 28, 0)
-  if (WOCKY_IS_HTTPS_PROXY (proxy))
+static void
+wocky_https_proxy_connect_async (GProxy *proxy,
+ GIOStream *io_stream,
+ GProxyAddress *proxy_address,
+ 

Re: [Spice-devel] guest agent windows binaries

[Marc-André Lureau]

- Original Message -
 On Wed, 12 Feb 2014 09:16:04 AM Christophe Fergeau wrote:
  http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw3
  2-spice-vdagent-0.7.2-1.fc19.noarch.rpm and
  http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw6
  4-spice-vdagent-0.7.2-1.fc19.noarch.rpm contain the binaries:
  
  $ rpm -qpl ./mingw64-spice-vdagent-0.7.2-1.fc19.noarch.rpm
  /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdagent.exe
  /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdservice.exe
 
 Why pack windows binaries in a rpm that can't be unpacked on windows?
 Thanks, but I'll try replicating the binaries and service entries from a win
 7
 installation.

I'll work on making an MSI installer for it, that shouldn't be difficult.
thanks for the reminder

 
  
   
   
  
   Also - could someone *please* clarify the current and future status of
   windows 8/2012 drivers and guest tools? its extremely confusing trying
   to track down this information.
  
  No QXL driver, the virtio-win drivers (located at
  https://alt.fedoraproject.org/pub/alt/virtio-win/latest/images/bin/ ) seem
  to have win8 versions
 
 Couldn't get them to install myself.
 
  , and the agent should be working as well on win8.
  I've never tried these though.
 
 I believe I have been labouring under the misapprehension that the qxl driver
 was necessary for the vdagent and all its goodies (display resizing, hot keys
 etc) to work, but that is not the case?
 
 It will need the virtio driver though won't it?
 
 I'll rebundle vdagent for win8/2012 and see how it goes.
 
 Thanks,
 
 --
 Lindsay
 ___
 Spice-devel mailing list
 Spice-devel@lists.freedesktop.org
 http://lists.freedesktop.org/mailman/listinfo/spice-devel
 
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 06/13] http-proxy: add https proxy

[Marc-André Lureau]

On Wed, Feb 12, 2014 at 11:48 AM, Christophe Fergeau
cferg...@redhat.com wrote:
 On Tue, Feb 11, 2014 at 12:37:53PM -0500, Marc-André Lureau wrote:
static void
wocky_http_proxy_init (WockyHttpProxy *proxy)
   @@ -180,10 +182,34 @@ wocky_http_proxy_connect (GProxy *proxy,
{
  GInputStream *in;
  GOutputStream *out;
   -  GDataInputStream *data_in;
   -  gchar *buffer;
   +  GDataInputStream *data_in = NULL;
   +  gchar *buffer = NULL;
  gboolean has_cred;
  
   +#if GLIB_CHECK_VERSION(2, 28, 0)
   +  if (WOCKY_IS_HTTPS_PROXY (proxy))
 
  Having runtime type checks in a interface virtual method is not very nice,
  having a wocky_https_proxy_iface_init would allow to call directly the
  right implementation without having to resort to runtime type checking.

 I decided to use this consciously, as doing it differently added much
 more boilerplate for no clear benefit.

 Gave it a quick try (only compile-tested), and the resulting code is not
 particularly bad. Having both inheritance and runtime type checks is imo
 quite ugly. I can send a follow-up patch after your series land if it's likely
 to be accepted.

Sure, I'll review then (the patch below looks incomplete from a brief
view). I still prefer a runtime type check for this final case.



 diff --git a/gtk/wocky-http-proxy.c b/gtk/wocky-http-proxy.c
 index 25f2af5..3eb6931 100644
 --- a/gtk/wocky-http-proxy.c
 +++ b/gtk/wocky-http-proxy.c
 @@ -186,30 +186,6 @@ wocky_http_proxy_connect (GProxy *proxy,
gchar *buffer = NULL;
gboolean has_cred;

 -#if GLIB_CHECK_VERSION(2, 28, 0)
 -  if (WOCKY_IS_HTTPS_PROXY (proxy))
 -{
 -  GIOStream *tlsconn;
 -
 -  tlsconn = g_tls_client_connection_new (io_stream,
 - 
 G_SOCKET_CONNECTABLE(proxy_address),
 - error);
 -  if (!tlsconn)
 -  goto error;
 -
 -  GTlsCertificateFlags tls_validation_flags = 
 G_TLS_CERTIFICATE_VALIDATE_ALL;
 -#ifdef DEBUG
 -  tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
 G_TLS_CERTIFICATE_BAD_IDENTITY;
 -#endif
 -  g_tls_client_connection_set_validation_flags (G_TLS_CLIENT_CONNECTION 
 (tlsconn),
 -tls_validation_flags);
 -  if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), 
 cancellable, error))
 -  goto error;
 -
 -  io_stream = tlsconn;
 -}
 -#endif
 -
in = g_io_stream_get_input_stream (io_stream);
out = g_io_stream_get_output_stream (io_stream);

 @@ -251,6 +227,41 @@ error:
return NULL;
  }

 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +static GIOStream *
 +wocky_https_proxy_connect (GProxy *proxy,
 +   GIOStream *io_stream,
 +   GProxyAddress *proxy_address,
 +   GCancellable *cancellable,
 +   GError **error)
 +{
 +{
 +GIOStream *tlsconn;
 +
 +tlsconn = g_tls_client_connection_new (io_stream,
 +   
 G_SOCKET_CONNECTABLE(proxy_address),
 +   error);
 +if (!tlsconn)
 +goto error;
 +
 +GTlsCertificateFlags tls_validation_flags = 
 G_TLS_CERTIFICATE_VALIDATE_ALL;
 +#ifdef DEBUG
 +tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
 G_TLS_CERTIFICATE_BAD_IDENTITY;
 +#endif
 +g_tls_client_connection_set_validation_flags 
 (G_TLS_CLIENT_CONNECTION (tlsconn),
 +  tls_validation_flags);
 +if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), 
 cancellable, error))
 +goto error;
 +
 +io_stream = tlsconn;
 +}
 +return wocky_http_proxy_connect(proxy, io_stream, proxy_address, 
 cancellable, error);
 +
 +error:
 +return NULL;
 +}
 +#endif
 +

  typedef struct
  {
 @@ -351,8 +362,9 @@ handshake_completed (GObject *source_object,
  }
  #endif

 -static void
 -wocky_http_proxy_connect_async (GProxy *proxy,
 +static ConnectAsyncData *
 +wocky_http_proxy_connect_async_data_new (
 +GProxy *proxy,
  GIOStream *io_stream,
  GProxyAddress *proxy_address,
  GCancellable *cancellable,
 @@ -364,7 +376,7 @@ wocky_http_proxy_connect_async (GProxy *proxy,

simple = g_simple_async_result_new (G_OBJECT (proxy),
callback, user_data,
 -  wocky_http_proxy_connect_async);
 +  
 wocky_http_proxy_connect_async_data_new);

data = g_slice_new0 (ConnectAsyncData);
if (cancellable != NULL)
 @@ -378,12 +390,30 @@ wocky_http_proxy_connect_async (GProxy *proxy,
g_simple_async_result_set_op_res_gpointer (simple, data,
   (GDestroyNotify) 
 free_connect_data);

 +  

Re: [Spice-devel] [spice-gtk 06/13] http-proxy: add https proxy

[Christophe Fergeau]

On Mon, Feb 03, 2014 at 07:02:37PM +0100, Marc-André Lureau wrote:
 From: Marc-André Lureau marcandre.lur...@redhat.com
 
 This will require glib 2.28 for GTls support, atm
 ---
  gtk/spice-session.c|   3 +
  gtk/wocky-http-proxy.c | 166 
 +
  gtk/wocky-http-proxy.h |  14 +
  3 files changed, 157 insertions(+), 26 deletions(-)
 
 diff --git a/gtk/spice-session.c b/gtk/spice-session.c
 index ae14a1f..6ac397c 100644
 --- a/gtk/spice-session.c
 +++ b/gtk/spice-session.c
 @@ -636,6 +636,9 @@ static void spice_session_class_init(SpiceSessionClass 
 *klass)
  #if GLIB_CHECK_VERSION(2, 26, 0)
  _wocky_http_proxy_get_type();
  #endif
 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +_wocky_https_proxy_get_type();
 +#endif
  
  gobject_class-dispose  = spice_session_dispose;
  gobject_class-finalize = spice_session_finalize;
 diff --git a/gtk/wocky-http-proxy.c b/gtk/wocky-http-proxy.c
 index 7210859..25f2af5 100644
 --- a/gtk/wocky-http-proxy.c
 +++ b/gtk/wocky-http-proxy.c
 @@ -1,7 +1,9 @@
   /* wocky-http-proxy.c: Source for WockyHttpProxy
   *
   * Copyright (C) 2010 Collabora, Ltd.
 + * Copyright (C) 2014 Red Hat, Inc.
   * @author Nicolas Dufresne nicolas.dufre...@collabora.co.uk
 + * @author Marc-André Lureau marcandre.lur...@redhat.com
   *
   * This library is free software; you can redistribute it and/or
   * modify it under the terms of the GNU Lesser General Public
 @@ -40,13 +42,13 @@ static void wocky_http_proxy_iface_init (GProxyInterface 
 *proxy_iface);
  
  #define wocky_http_proxy_get_type _wocky_http_proxy_get_type
  G_DEFINE_TYPE_WITH_CODE (WockyHttpProxy, wocky_http_proxy, G_TYPE_OBJECT,
 -G_IMPLEMENT_INTERFACE (G_TYPE_PROXY,
 -  wocky_http_proxy_iface_init)
 -g_io_extension_point_set_required_type (
 -  g_io_extension_point_register (G_PROXY_EXTENSION_POINT_NAME),
 -  G_TYPE_PROXY);
 -g_io_extension_point_implement (G_PROXY_EXTENSION_POINT_NAME,
 -  g_define_type_id, http, 0))
 +  G_IMPLEMENT_INTERFACE (G_TYPE_PROXY,
 +wocky_http_proxy_iface_init)
 +  g_io_extension_point_set_required_type (
 +g_io_extension_point_register (G_PROXY_EXTENSION_POINT_NAME),
 +G_TYPE_PROXY);
 +  g_io_extension_point_implement (G_PROXY_EXTENSION_POINT_NAME,
 +g_define_type_id, http, 0))
  
  static void
  wocky_http_proxy_init (WockyHttpProxy *proxy)
 @@ -180,10 +182,34 @@ wocky_http_proxy_connect (GProxy *proxy,
  {
GInputStream *in;
GOutputStream *out;
 -  GDataInputStream *data_in;
 -  gchar *buffer;
 +  GDataInputStream *data_in = NULL;
 +  gchar *buffer = NULL;
gboolean has_cred;
  
 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +  if (WOCKY_IS_HTTPS_PROXY (proxy))
 +{
 +  GIOStream *tlsconn;
 +
 +  tlsconn = g_tls_client_connection_new (io_stream,
 + 
 G_SOCKET_CONNECTABLE(proxy_address),
 + error);
 +  if (!tlsconn)
 +  goto error;
 +
 +  GTlsCertificateFlags tls_validation_flags = 
 G_TLS_CERTIFICATE_VALIDATE_ALL;
 +#ifdef DEBUG
 +  tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
 G_TLS_CERTIFICATE_BAD_IDENTITY;
 +#endif
 +  g_tls_client_connection_set_validation_flags (G_TLS_CLIENT_CONNECTION 
 (tlsconn),
 +tls_validation_flags);
 +  if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), 
 cancellable, error))
 +  goto error;
 +
 +  io_stream = tlsconn;

Missed that initially, but tlsconn is leaked in error cases, and I think we
will leak a reference to it in success case too. My understanding is that
we don't own the io_stream passed as an argument, so we need to ref it in
the http case, but in the https case, we already own a ref on the io_stream
(which is tlsconn), so we ref it one too many.

 +
 +static void
  wocky_http_proxy_connect_async (GProxy *proxy,
  GIOStream *io_stream,
  GProxyAddress *proxy_address,
 @@ -300,34 +361,55 @@ wocky_http_proxy_connect_async (GProxy *proxy,
 +
 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +  if (WOCKY_IS_HTTPS_PROXY (proxy))
 +{
 +  GError *error = NULL;
 +  GIOStream *tlsconn;
 +
 +  tlsconn = g_tls_client_connection_new (io_stream,
 + 
 G_SOCKET_CONNECTABLE(proxy_address),
 + error);
 +  if (error)
 +{
 +  complete_async_from_error (data, error);

tlsconn is leaked here (dunno if it's guaranteed to be NULL when error is
set).

 +  return;
 +}
 +
 +  g_return_if_fail (tlsconn != NULL);
 +
 +  GTlsCertificateFlags tls_validation_flags = 
 G_TLS_CERTIFICATE_VALIDATE_ALL;
 +#ifdef DEBUG
 +  tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
 G_TLS_CERTIFICATE_BAD_IDENTITY;
 +#endif
 +  g_tls_client_connection_set_validation_flags (G_TLS_CLIENT_CONNECTION 
 (tlsconn),
 +   

Re: [Spice-devel] guest agent windows binaries

[Christophe Fergeau]

On Wed, Feb 12, 2014 at 08:48:44PM +1000, Lindsay Mathieson wrote:
 On Wed, 12 Feb 2014 09:16:04 AM Christophe Fergeau wrote:
  http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw3
  2-spice-vdagent-0.7.2-1.fc19.noarch.rpm and
  http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw6
  4-spice-vdagent-0.7.2-1.fc19.noarch.rpm contain the binaries:
  
  $ rpm -qpl ./mingw64-spice-vdagent-0.7.2-1.fc19.noarch.rpm
  /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdagent.exe
  /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdservice.exe
 
 Why pack windows binaries in a rpm that can't be unpacked on windows?

Because that's how I get them from from the fedora build system, because I
haven't automated much of the generation of the spice guest tools
installer, because that's already a lot of manual steps, ...

Christophe


pgpX_EDxpt2Vh.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 06/13] http-proxy: add https proxy

[Marc-André Lureau]

On Wed, Feb 12, 2014 at 11:59 AM, Christophe Fergeau
cferg...@redhat.com wrote:
 On Mon, Feb 03, 2014 at 07:02:37PM +0100, Marc-André Lureau wrote:
 From: Marc-André Lureau marcandre.lur...@redhat.com

 This will require glib 2.28 for GTls support, atm
 ---
  gtk/spice-session.c|   3 +
  gtk/wocky-http-proxy.c | 166 
 +
  gtk/wocky-http-proxy.h |  14 +
  3 files changed, 157 insertions(+), 26 deletions(-)

 diff --git a/gtk/spice-session.c b/gtk/spice-session.c
 index ae14a1f..6ac397c 100644
 --- a/gtk/spice-session.c
 +++ b/gtk/spice-session.c
 @@ -636,6 +636,9 @@ static void spice_session_class_init(SpiceSessionClass 
 *klass)
  #if GLIB_CHECK_VERSION(2, 26, 0)
  _wocky_http_proxy_get_type();
  #endif
 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +_wocky_https_proxy_get_type();
 +#endif

  gobject_class-dispose  = spice_session_dispose;
  gobject_class-finalize = spice_session_finalize;
 diff --git a/gtk/wocky-http-proxy.c b/gtk/wocky-http-proxy.c
 index 7210859..25f2af5 100644
 --- a/gtk/wocky-http-proxy.c
 +++ b/gtk/wocky-http-proxy.c
 @@ -1,7 +1,9 @@
   /* wocky-http-proxy.c: Source for WockyHttpProxy
   *
   * Copyright (C) 2010 Collabora, Ltd.
 + * Copyright (C) 2014 Red Hat, Inc.
   * @author Nicolas Dufresne nicolas.dufre...@collabora.co.uk
 + * @author Marc-André Lureau marcandre.lur...@redhat.com
   *
   * This library is free software; you can redistribute it and/or
   * modify it under the terms of the GNU Lesser General Public
 @@ -40,13 +42,13 @@ static void wocky_http_proxy_iface_init (GProxyInterface 
 *proxy_iface);

  #define wocky_http_proxy_get_type _wocky_http_proxy_get_type
  G_DEFINE_TYPE_WITH_CODE (WockyHttpProxy, wocky_http_proxy, G_TYPE_OBJECT,
 -G_IMPLEMENT_INTERFACE (G_TYPE_PROXY,
 -  wocky_http_proxy_iface_init)
 -g_io_extension_point_set_required_type (
 -  g_io_extension_point_register (G_PROXY_EXTENSION_POINT_NAME),
 -  G_TYPE_PROXY);
 -g_io_extension_point_implement (G_PROXY_EXTENSION_POINT_NAME,
 -  g_define_type_id, http, 0))
 +  G_IMPLEMENT_INTERFACE (G_TYPE_PROXY,
 +wocky_http_proxy_iface_init)
 +  g_io_extension_point_set_required_type (
 +g_io_extension_point_register (G_PROXY_EXTENSION_POINT_NAME),
 +G_TYPE_PROXY);
 +  g_io_extension_point_implement (G_PROXY_EXTENSION_POINT_NAME,
 +g_define_type_id, http, 0))

  static void
  wocky_http_proxy_init (WockyHttpProxy *proxy)
 @@ -180,10 +182,34 @@ wocky_http_proxy_connect (GProxy *proxy,
  {
GInputStream *in;
GOutputStream *out;
 -  GDataInputStream *data_in;
 -  gchar *buffer;
 +  GDataInputStream *data_in = NULL;
 +  gchar *buffer = NULL;
gboolean has_cred;

 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +  if (WOCKY_IS_HTTPS_PROXY (proxy))
 +{
 +  GIOStream *tlsconn;
 +
 +  tlsconn = g_tls_client_connection_new (io_stream,
 + 
 G_SOCKET_CONNECTABLE(proxy_address),
 + error);
 +  if (!tlsconn)
 +  goto error;
 +
 +  GTlsCertificateFlags tls_validation_flags = 
 G_TLS_CERTIFICATE_VALIDATE_ALL;
 +#ifdef DEBUG
 +  tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
 G_TLS_CERTIFICATE_BAD_IDENTITY;
 +#endif
 +  g_tls_client_connection_set_validation_flags (G_TLS_CLIENT_CONNECTION 
 (tlsconn),
 +tls_validation_flags);
 +  if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), 
 cancellable, error))
 +  goto error;
 +
 +  io_stream = tlsconn;

 Missed that initially, but tlsconn is leaked in error cases, and I think we
 will leak a reference to it in success case too. My understanding is that

if error, function returns NULL

 we don't own the io_stream passed as an argument, so we need to ref it in
 the http case, but in the https case, we already own a ref on the io_stream
 (which is tlsconn), so we ref it one too many.

Where do you see we ref something?
I remember I did check with valgrind and debugging than all references
where correct. But I could be wrong. Please give further details.


 +
 +static void
  wocky_http_proxy_connect_async (GProxy *proxy,
  GIOStream *io_stream,
  GProxyAddress *proxy_address,
 @@ -300,34 +361,55 @@ wocky_http_proxy_connect_async (GProxy *proxy,
 +
 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +  if (WOCKY_IS_HTTPS_PROXY (proxy))
 +{
 +  GError *error = NULL;
 +  GIOStream *tlsconn;
 +
 +  tlsconn = g_tls_client_connection_new (io_stream,
 + 
 G_SOCKET_CONNECTABLE(proxy_address),
 + error);
 +  if (error)
 +{
 +  complete_async_from_error (data, error);

 tlsconn is leaked here (dunno if it's guaranteed to be NULL when error is
 set).

Yes it is guaratee, I can add a warning and  g_clear_object if you want.

 +  return;
 +}
 +
 +  

Re: [Spice-devel] guest agent windows binaries

[Lindsay Mathieson]

On Wed, 12 Feb 2014 12:00:45 PM Christophe Fergeau wrote:
 Because that's how I get them from from the fedora build system, because I
 haven't automated much of the generation of the spice guest tools
 installer, because that's already a lot of manual steps, ...


Fair enough :)

thanks.
-- 
Lindsay

signature.asc
Description: This is a digitally signed message part.
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] guest agent windows binaries

[Marian Krcmarik]

- Original Message -
 From: Lindsay Mathieson lindsay.mathie...@gmail.com
 To: spice-devel@lists.freedesktop.org
 Sent: Wednesday, February 12, 2014 11:48:44 AM
 Subject: Re: [Spice-devel] guest agent windows binaries
 
 On Wed, 12 Feb 2014 09:16:04 AM Christophe Fergeau wrote:
  http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw3
  2-spice-vdagent-0.7.2-1.fc19.noarch.rpm and
  http://www.spice-space.org/download/windows/vdagent/vdagent-win-0.7.2/mingw6
  4-spice-vdagent-0.7.2-1.fc19.noarch.rpm contain the binaries:
  
  $ rpm -qpl ./mingw64-spice-vdagent-0.7.2-1.fc19.noarch.rpm
  /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdagent.exe
  /usr/x86_64-w64-mingw32/sys-root/mingw/bin/vdservice.exe
 
 Why pack windows binaries in a rpm that can't be unpacked on windows?
RPM is still a file archive :) All you would need is just good file archiver 
(for example open sourced 7-zip for Windows).
 Thanks, but I'll try replicating the binaries and service entries from a win
 7
 installation.
 
 
  
   
   
  
   Also - could someone *please* clarify the current and future status of
   windows 8/2012 drivers and guest tools? its extremely confusing trying
   to track down this information.
  
  No QXL driver, the virtio-win drivers (located at
  https://alt.fedoraproject.org/pub/alt/virtio-win/latest/images/bin/ ) seem
  to have win8 versions
 
 Couldn't get them to install myself.
 
  , and the agent should be working as well on win8.
  I've never tried these though.
 
 I believe I have been labouring under the misapprehension that the qxl driver
 was necessary for the vdagent and all its goodies (display resizing, hot keys
 etc) to work, but that is not the case?
 
 It will need the virtio driver though won't it?
 
 I'll rebundle vdagent for win8/2012 and see how it goes.
 
 Thanks,
 
 --
 Lindsay
 ___
 Spice-devel mailing list
 Spice-devel@lists.freedesktop.org
 http://lists.freedesktop.org/mailman/listinfo/spice-devel
 
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 08/13] openssl: learn to handle a new kind of BIO based on GIOStream

[Christophe Fergeau]

On Mon, Feb 03, 2014 at 07:02:39PM +0100, Marc-André Lureau wrote:
 From: Marc-André Lureau marcandre.lur...@redhat.com
 
 Although reusing BIO_new_socket() once again is a hack, it seems
 to be the easiest way... The proper solution is certainly to start
 using GTls instead, but that will require a glib 2.28 dep bump.

Ah, much easier to review with git show -M (rename detection). 
 +
 +G_GNUC_INTERNAL
 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +BIO* bio_new_giostream(GIOStream *stream)
 +{
 +// TODO: make an actual new BIO type, or just switch to GTls already...
 +BIO *bio = BIO_new_socket(-1, BIO_NOCLOSE);
 +#else
 +BIO* bio_new_gsocket(GSocket *gsocket)
 +{
 +BIO *bio = BIO_new_socket(g_socket_get_fd(gsocket), BIO_NOCLOSE);
 +#endif

This #if block is not quite readable, I probably would have split that in a
helper function with the common code + 2 #ifdef'ed
bio_new_giostream/bio_new_gsocket calling that helper.

BIO_new_socket() will not mind having its fd set to -1 ?

ACK.

Christophe


pgptDOVvN1nce.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 08/13] openssl: learn to handle a new kind of BIO based on GIOStream

[Marc-André Lureau]

On Wed, Feb 12, 2014 at 12:21 PM, Christophe Fergeau
cferg...@redhat.com wrote:
 +G_GNUC_INTERNAL
 +#if GLIB_CHECK_VERSION(2, 28, 0)
 +BIO* bio_new_giostream(GIOStream *stream)
 +{
 +// TODO: make an actual new BIO type, or just switch to GTls already...
 +BIO *bio = BIO_new_socket(-1, BIO_NOCLOSE);
 +#else
 +BIO* bio_new_gsocket(GSocket *gsocket)
 +{
 +BIO *bio = BIO_new_socket(g_socket_get_fd(gsocket), BIO_NOCLOSE);
 +#endif

 This #if block is not quite readable, I probably would have split that in a
 helper function with the common code + 2 #ifdef'ed
 bio_new_giostream/bio_new_gsocket calling that helper.

My idea is more to get the code in the (2,28) block ready, so that
else #else can be removed without more code churn.

 BIO_new_socket() will not mind having its fd set to -1 ?

nope, but this is all just quite hacky tbh (the whole BIO thingy)..
Hopefully we can move to 2.28 soon.


 ACK.

 Christophe



-- 
Marc-André Lureau
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 10/13] channel: simplify has error code

[Christophe Fergeau]

ACK

Christophe

On Mon, Feb 03, 2014 at 07:02:41PM +0100, Marc-André Lureau wrote:
 From: Marc-André Lureau marcandre.lur...@redhat.com
 
 Get rid of a superflous g_socket_condition_check().
 ---
  gtk/spice-channel.c | 28 +++-
  1 file changed, 11 insertions(+), 17 deletions(-)
 
 diff --git a/gtk/spice-channel.c b/gtk/spice-channel.c
 index 7459a49..87bc471 100644
 --- a/gtk/spice-channel.c
 +++ b/gtk/spice-channel.c
 @@ -2105,29 +2105,23 @@ static gboolean wait_migration(gpointer data)
  static gboolean spice_channel_iterate(SpiceChannel *channel)
  {
  SpiceChannelPrivate *c = channel-priv;
 -GIOCondition ret;
  
  if (c-state == SPICE_CHANNEL_STATE_MIGRATING 
  !g_coroutine_condition_wait(c-coroutine, wait_migration, channel))
  CHANNEL_DEBUG(channel, migration wait cancelled);
  
 -if (c-has_error) {
 -CHANNEL_DEBUG(channel, channel has error, breaking loop);
 -return FALSE;
 -}
 -
  /* flush any pending write and read */
 -SPICE_CHANNEL_GET_CLASS(channel)-iterate_write(channel);
 -SPICE_CHANNEL_GET_CLASS(channel)-iterate_read(channel);
 -
 -ret = g_socket_condition_check(c-sock, G_IO_IN | G_IO_ERR | G_IO_HUP);
 -if (c-state  SPICE_CHANNEL_STATE_CONNECTING 
 -ret  (G_IO_ERR|G_IO_HUP)) {
 -SPICE_DEBUG(got socket error: %d, ret);
 -emit_main_context(channel, SPICE_CHANNEL_EVENT,
 -  c-state == SPICE_CHANNEL_STATE_READY ?
 -  SPICE_CHANNEL_ERROR_IO : SPICE_CHANNEL_ERROR_LINK);
 -c-has_error = TRUE;
 +if (!c-has_error)
 +SPICE_CHANNEL_GET_CLASS(channel)-iterate_write(channel);
 +if (!c-has_error)
 +SPICE_CHANNEL_GET_CLASS(channel)-iterate_read(channel);
 +
 +if (c-has_error) {
 +CHANNEL_DEBUG(channel, channel got error);
 +if (c-state  SPICE_CHANNEL_STATE_CONNECTING)
 +emit_main_context(channel, SPICE_CHANNEL_EVENT,
 +  c-state == SPICE_CHANNEL_STATE_READY ?
 +  SPICE_CHANNEL_ERROR_IO : 
 SPICE_CHANNEL_ERROR_LINK);
  return FALSE;
  }
  
 -- 
 1.8.4.2
 
 ___
 Spice-devel mailing list
 Spice-devel@lists.freedesktop.org
 http://lists.freedesktop.org/mailman/listinfo/spice-devel


pgp1QuhDkZwF9.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 06/13] http-proxy: add https proxy

[Christophe Fergeau]

On Wed, Feb 12, 2014 at 12:04:25PM +0100, Marc-André Lureau wrote:
 On Wed, Feb 12, 2014 at 11:59 AM, Christophe Fergeau
 cferg...@redhat.com wrote:
  Missed that initially, but tlsconn is leaked in error cases, and I think we
  will leak a reference to it in success case too. My understanding is that
 
 if error, function returns NULL

tlsconn = g_tls_client_connection_new (io_stream,
   G_SOCKET_CONNECTABLE(proxy_address),
   error);

if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), cancellable, 
error))
goto error;

error:
  if (data_in != NULL)
g_object_unref (data_in);

  g_free (buffer);
  return NULL;
}

tlsconn will be leaked.


 
  we don't own the io_stream passed as an argument, so we need to ref it in
  the http case, but in the https case, we already own a ref on the io_stream
  (which is tlsconn), so we ref it one too many.
 
 Where do you see we ref something?
 I remember I did check with valgrind and debugging than all references
 where correct. But I could be wrong. Please give further details.
 

  tlsconn = g_tls_client_connection_new (io_stream,
 G_SOCKET_CONNECTABLE(proxy_address),
 error);
  io_stream = tlsconn;

  return g_object_ref (io_stream);

I think we are taking an extra ref on tlsconn here (ref which is probably
needed in the !tls case).
  tlsconn is leaked here (dunno if it's guaranteed to be NULL when error is
  set).
 
 Yes it is guaratee, I can add a warning and  g_clear_object if you want.

a g_warn_if_fail() cannot hurt as it's not explicit in the API doc (or
fixing the API doc would work too ;)

 
  +  return;
  +}
  +
  +  g_return_if_fail (tlsconn != NULL);
  +
  +  GTlsCertificateFlags tls_validation_flags = 
  G_TLS_CERTIFICATE_VALIDATE_ALL;
  +#ifdef DEBUG
  +  tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
  G_TLS_CERTIFICATE_BAD_IDENTITY;
  +#endif
  +  g_tls_client_connection_set_validation_flags 
  (G_TLS_CLIENT_CONNECTION (tlsconn),
  +tls_validation_flags);
  +  g_tls_connection_handshake_async (G_TLS_CONNECTION (tlsconn),
  +G_PRIORITY_DEFAULT, cancellable,
  +handshake_completed, data);
 
  A g_object_unref(tlsconn) is probably needed here.
 
 
 no, the async is running and we need to keep the ref.

Ah, my bad, I thought glib async functions were ref'ing their arg as
needed, but I'm wrong from a quick glance at gio/glib-networking.

Christophe


pgp4ZKPzRACCw.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 11/13] Make SpiceURI a public API

[Christophe Fergeau]

On Mon, Feb 03, 2014 at 07:02:42PM +0100, Marc-André Lureau wrote:
 From: Marc-André Lureau marcandre.lur...@redhat.com
 
 Generalize a little bit SpiceProxy to allow easy URI manipulation by
 clients.
 ---
  doc/reference/Makefile.am|   1 +
  doc/reference/spice-gtk-docs.xml |   1 +
  doc/reference/spice-gtk-sections.txt |  26 ++
  gtk/Makefile.am  |   5 +-
  gtk/map-file |  12 +
  gtk/spice-client.h   |   1 +
  gtk/spice-glib-sym-file  |  12 +
  gtk/spice-proxy.c| 357 --
  gtk/spice-proxy.h|  64 -
  gtk/spice-session-priv.h |   3 +-
  gtk/spice-session.c  |  28 +--
  gtk/spice-session.h  |   1 +
  gtk/spice-uri-priv.h |  30 +++
  gtk/spice-uri.c  | 473 
 +++
  gtk/spice-uri.h  |  52 
  15 files changed, 627 insertions(+), 439 deletions(-)
  delete mode 100644 gtk/spice-proxy.c
  delete mode 100644 gtk/spice-proxy.h
  create mode 100644 gtk/spice-uri-priv.h
  create mode 100644 gtk/spice-uri.c
  create mode 100644 gtk/spice-uri.h
 
 diff --git a/doc/reference/Makefile.am b/doc/reference/Makefile.am
 index a0a856c..76c7d34 100644
 --- a/doc/reference/Makefile.am
 +++ b/doc/reference/Makefile.am
 @@ -44,6 +44,7 @@ IGNORE_HFILES=  \
   spice-marshal.h \
   spice-pulse.h   \
   spice-session-priv.h\
 + spice-uri-priv.h\
   spice-util-priv.h   \
   spice-widget-priv.h \
   usb-acl-helper.h\
 diff --git a/doc/reference/spice-gtk-docs.xml 
 b/doc/reference/spice-gtk-docs.xml
 index 4a9a3cf..d2c1a2b 100644
 --- a/doc/reference/spice-gtk-docs.xml
 +++ b/doc/reference/spice-gtk-docs.xml
 @@ -52,6 +52,7 @@
xi:include href=xml/smartcard-manager.xml/
xi:include href=xml/usb-device-manager.xml/
xi:include href=xml/spice-util.xml/
 +  xi:include href=xml/spice-uri.xml/
  /chapter
  
/part
 diff --git a/doc/reference/spice-gtk-sections.txt 
 b/doc/reference/spice-gtk-sections.txt
 index 8d61aa9..411ca0e 100644
 --- a/doc/reference/spice-gtk-sections.txt
 +++ b/doc/reference/spice-gtk-sections.txt
 @@ -429,3 +429,29 @@ SPICE_PORT_CHANNEL_GET_CLASS
  SpicePortChannelPrivate
  /SECTION
  
 +SECTION
 +FILEspice-uri/FILE
 +spice_uri_get_scheme
 +spice_uri_set_scheme
 +spice_uri_get_hostname
 +spice_uri_set_hostname
 +spice_uri_get_port
 +spice_uri_set_port
 +spice_uri_get_user
 +spice_uri_set_user
 +spice_uri_get_password
 +spice_uri_set_password
 +spice_uri_to_string
 +SpiceURIClass
 +SpiceURI
 +SUBSECTION Standard
 +SPICE_IS_URI
 +SPICE_IS_URI_CLASS
 +SPICE_TYPE_URI
 +SPICE_URI
 +SPICE_URI_CLASS
 +SPICE_URI_GET_CLASS
 +spice_uri_get_type
 +SUBSECTION Private
 +SpiceURIPrivate
 +/SECTION
 diff --git a/gtk/Makefile.am b/gtk/Makefile.am
 index 840d129..45d6796 100644
 --- a/gtk/Makefile.am
 +++ b/gtk/Makefile.am
 @@ -248,8 +248,8 @@ libspice_client_glib_2_0_la_SOURCES = 
 \
   channel-usbredir-priv.h \
   smartcard-manager.c \
   smartcard-manager-priv.h\
 - spice-proxy.c   \
 - spice-proxy.h   \
 + spice-uri.c \
 + spice-uri-priv.h\
   usb-device-manager.c\
   usb-device-manager-priv.h   \
   usbutil.c   \
 @@ -282,6 +282,7 @@ libspice_client_glibincludedir = 
 $(includedir)/spice-client-glib-2.0
  libspice_client_glibinclude_HEADERS =\
   spice-audio.h   \
   spice-client.h  \
 + spice-uri.h \
   spice-types.h   \
   spice-session.h \
   spice-channel.h \
 diff --git a/gtk/map-file b/gtk/map-file
 index 368b44f..cf65db0 100644
 --- a/gtk/map-file
 +++ b/gtk/map-file
 @@ -118,6 +118,18 @@ spice_util_get_debug;
  spice_util_get_version_string;
  spice_util_set_debug;
  spice_uuid_to_string;
 +spice_uri_get_hostname;
 +spice_uri_get_password;
 +spice_uri_get_port;
 +spice_uri_get_scheme;
 +spice_uri_get_type;
 +spice_uri_get_user;
 +spice_uri_set_hostname;
 +spice_uri_set_password;
 +spice_uri_set_port;
 +spice_uri_set_scheme;
 +spice_uri_set_user;
 +spice_uri_to_string;
  local:
  *;
  };
 diff --git a/gtk/spice-client.h b/gtk/spice-client.h
 index 730d11a..975259a 100644
 --- a/gtk/spice-client.h
 +++ b/gtk/spice-client.h
 @@ -31,6 +31,7 @@
  #include spice-session.h
  #include spice-channel.h
  #include 

Re: [Spice-devel] guest agent windows binaries

[Lindsay Mathieson]

On 12 February 2014 21:08, Marian Krcmarik mkrcm...@redhat.com wrote:


 Why pack windows binaries in a rpm that can't be unpacked on windows?
 RPM is still a file archive :) All you would need is just good file archiver 
 (for example open sourced 7-zip for Windows).


Cool, thanks.

And I tried the 0.74 drivers again - all working this time (Windows
8.1), dunno what I was doing wrong before, but duh!

Will play with the service tomorrow,

thanks for all the help everyone.



-- 
Lindsay
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 12/13] channel: add spice_channel_get_error()

[Christophe Fergeau]

On Tue, Feb 11, 2014 at 01:08:55PM -0500, Marc-André Lureau wrote:
 
 
 - Original Message -
  On Tue, Feb 11, 2014 at 06:50:18PM +0100, Marc-André Lureau wrote:
@@ -2248,15 +2269,16 @@ static void *spice_channel_coroutine(void 
*data)
   
   
 reconnect:
-c-conn = spice_session_channel_open_host(c-session, channel,
c-tls);
+c-conn = spice_session_channel_open_host(c-session, channel,
c-tls, c-error);
 if (c-conn == NULL) {
-if (!c-tls) {
+if (!c-error  !c-tls) {
 CHANNEL_DEBUG(channel, trying with TLS port);
 c-tls = true; /* FIXME: does that really work with
 provided fd */
 goto reconnect;
 } else {
 CHANNEL_DEBUG(channel, Connect error);
 emit_main_context(channel, SPICE_CHANNEL_EVENT,
 SPICE_CHANNEL_ERROR_CONNECT);
+g_clear_error(c-error);
   
My understanding of this bit is that spice_channel_get_error() can only
be
used in a callback attached to SPICE_CHANNEL_EVENT as after that it will
be
cleared? Maybe this needs to be more explicit in the API doc for
spice_channel_get_error().
   
   It is meant to be usable at any point.

Currently, if you don't use it in this callback, all you will get is a NULL
error even if a connection error occurred before.

  
  Won't the g_clear_error() I quoted cause spice_channel_get_error() to
  return the correct error in the SPICE_CHANNEL_EVENT callback when there is
  a connection error, and then return NULL after the g_clear_error() has been
  called? Ie the error state of the channel will not be permanent after an
  error has happened, but will be silently reset by the library.
  Not saying this is bad per-se, at first I was worried that we could get in
  a permanent error state it's not possible to get out of.
 
 The API doesn't specify when the channel is in error state. When it is,
 you may get more information via this API, I don't see the need to make
 it more specific.

IThis does not strike me as a very good public API to export. From the
library user point of view, there will be this function that can give more
information about channel errors, but when an error occurred, the function
may return NULL when you call it as the library resets the error at
totally unspecified moments.
So it can be good enough as something internal, but imo we need to do
better at least in the API doc if it's something that we export.

Christophe


pgpUg64zriOyY.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] Win 8 Display Resolution

[Lindsay Mathieson]

Without a QXL driver isn't it limited to 1024x768 resolution?

-- 
Lindsay
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] Win 8 Display Resolution

[Daniel P. Berrange]

On Wed, Feb 12, 2014 at 10:13:18PM +1000, Lindsay Mathieson wrote:
 Without a QXL driver isn't it limited to 1024x768 resolution?

You should be able to go higher than that. IIUC even the cirrus driver
supports more than that, but if not then the 'vga' driver for QEMU
should. It is mostly determined by how much video RAM is exposed to
the guest.

QXL gives you other nice features like multi-monitor, dynamic mode
switching via the guest agent, better performance with spice, etc.

Daniel
-- 
|: http://berrange.com  -o-http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org  -o- http://virt-manager.org :|
|: http://autobuild.org   -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org   -o-   http://live.gnome.org/gtk-vnc :|
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [PATCH spice-gtk 04/14] proxy: parse user and pass from uri

[Christophe Fergeau]

On Wed, Feb 12, 2014 at 11:19:02AM +0100, Marc-André Lureau wrote:
 From: Marc-André Lureau marcandre.lur...@redhat.com
 
 ---
  gtk/spice-proxy.c | 13 +
  1 file changed, 13 insertions(+)
 
 diff --git a/gtk/spice-proxy.c b/gtk/spice-proxy.c
 index 7711e83..35eec67 100644
 --- a/gtk/spice-proxy.c
 +++ b/gtk/spice-proxy.c
 @@ -19,6 +19,7 @@
  #include stdlib.h
  #include string.h
  
 +#include glib-compat.h
  #include spice-client.h
  #include spice-proxy.h
  
 @@ -78,6 +79,18 @@ gboolean spice_proxy_parse(SpiceProxy *self, const gchar 
 *proxyuri, GError **err
  spice_proxy_set_protocol(self, http);
  spice_proxy_set_port(self, 3128);
  
 +/* yes, that parser is bad, we need GUri... */
 +if (strstr(uri, @)) {
 +gchar *saveptr, *saveptr2;
 +gchar *next = strstr(uri, @) + 1;

The strstr() call could probably be done only once, but it's fine enough
like this.

ACK.

Christophe

 +gchar *auth = strtok_r(uri, @, saveptr);
 +const gchar *user = strtok_r(auth, :, saveptr2);
 +const gchar *pass = strtok_r(NULL, :, saveptr2);
 +spice_proxy_set_user(self, user);
 +spice_proxy_set_password(self, pass);
 +uri = next;
 +}
 +
  /* max 2 parts, host:port */
  gchar **proxyv = g_strsplit(uri, :, 2);
  const gchar *proxy_port = NULL;
 -- 
 1.8.4.2
 
 ___
 Spice-devel mailing list
 Spice-devel@lists.freedesktop.org
 http://lists.freedesktop.org/mailman/listinfo/spice-devel


pgp09MTOZuDIQ.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [PATCH spice-gtk 11/14] channel: talk to giostream instead of gsocket

[Christophe Fergeau]

On Wed, Feb 12, 2014 at 11:19:09AM +0100, Marc-André Lureau wrote:
 From: Marc-André Lureau marcandre.lur...@redhat.com
 
 ---
  gtk/spice-channel-priv.h |  4 ++-
  gtk/spice-channel.c  | 81 
 +++-
  2 files changed, 56 insertions(+), 29 deletions(-)
 
 diff --git a/gtk/spice-channel-priv.h b/gtk/spice-channel-priv.h
 index 35704ea..00893c3 100644
 --- a/gtk/spice-channel-priv.h
 +++ b/gtk/spice-channel-priv.h
 @@ -80,8 +80,10 @@ struct _SpiceChannelPrivate {
  SSL_CTX *ctx;
  SSL *ssl;
  SpiceOpenSSLVerify  *sslverify;
 -GSocket *sock;
 +GSocket *socket;

NACK (this was not in v1 :-/ )

Christophe


pgpMcq1wjaGlR.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [PATCH spice-gtk] Add SpiceVMC GIOStream

[Christophe Fergeau]

On Tue, Feb 11, 2014 at 06:52:49PM +0100, Marc-André Lureau wrote:
 This allows to use conveniently GIOStream APIs without caring about
 coroutine and Spice messages details.
 ---
  gtk/Makefile.am |   2 +
  gtk/vmcstream.c | 532 
 
  gtk/vmcstream.h |  81 +
  3 files changed, 615 insertions(+)
  create mode 100644 gtk/vmcstream.c
  create mode 100644 gtk/vmcstream.h
 
 diff --git a/gtk/Makefile.am b/gtk/Makefile.am
 index 62afd36..7ceb22f 100644
 --- a/gtk/Makefile.am
 +++ b/gtk/Makefile.am
 @@ -255,6 +255,8 @@ libspice_client_glib_2_0_la_SOURCES = 
 \
   usbutil.c   \
   usbutil.h   \
   $(USB_ACL_HELPER_SRCS)  \
 + vmcstream.c \
 + vmcstream.h \
   \
   decode.h\
   decode-glz.c\
 diff --git a/gtk/vmcstream.c b/gtk/vmcstream.c
 new file mode 100644
 index 000..f92f8b7
 --- /dev/null
 +++ b/gtk/vmcstream.c
 @@ -0,0 +1,532 @@
 +/* -*- Mode: C; c-basic-offset: 4; indent-tabs-mode: nil -*- */
 +/*
 +  Copyright (C) 2013 Red Hat, Inc.
 +
 +  This library is free software; you can redistribute it and/or
 +  modify it under the terms of the GNU Lesser General Public
 +  License as published by the Free Software Foundation; either
 +  version 2.1 of the License, or (at your option) any later version.
 +
 +  This library is distributed in the hope that it will be useful,
 +  but WITHOUT ANY WARRANTY; without even the implied warranty of
 +  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 +  Lesser General Public License for more details.
 +
 +  You should have received a copy of the GNU Lesser General Public
 +  License along with this library; if not, see 
 http://www.gnu.org/licenses/.
 +*/
 +#include string.h
 +
 +#include vmcstream.h
 +#include spice-channel-priv.h
 +#include gio-coroutine.h
 +
 +struct _SpiceVmcInputStream
 +{
 +GInputStream parent_instance;
 +GSimpleAsyncResult *result;
 +struct coroutine *coroutine;
 +
 +SpiceChannel *channel;
 +gboolean all;
 +guint8 *buffer;
 +gsize count;
 +gsize pos;
 +
 +GCancellable *cancellable;
 +gulong cancel_id;
 +};
 +
 +struct _SpiceVmcInputStreamClass
 +{
 +GInputStreamClass parent_class;
 +};
 +
 +static gssize   spice_vmc_input_stream_read(GInputStream
 *stream,
 +void
 *buffer,
 +gsize
 count,
 +GCancellable
 *cancellable,
 +GError 
 **error);
 +static void spice_vmc_input_stream_read_async  (GInputStream
 *stream,
 +void
 *buffer,
 +gsize
 count,
 +int  
 io_priority,
 +GCancellable
 *cancellable,
 +GAsyncReadyCallback  
 callback,
 +gpointer 
 user_data);
 +static gssize   spice_vmc_input_stream_read_finish (GInputStream
 *stream,
 +GAsyncResult
 *result,
 +GError 
 **error);
 +static gssize   spice_vmc_input_stream_skip(GInputStream
 *stream,
 +gsize
 count,
 +GCancellable
 *cancellable,
 +GError 
 **error);
 +static gboolean spice_vmc_input_stream_close   (GInputStream
 *stream,
 +GCancellable
 *cancellable,
 +GError 
 **error);
 +
 +G_DEFINE_TYPE(SpiceVmcInputStream, spice_vmc_input_stream, 
 G_TYPE_INPUT_STREAM)
 +
 +
 +static void
 +spice_vmc_input_stream_class_init(SpiceVmcInputStreamClass *klass)
 +{
 +GInputStreamClass *istream_class;
 +
 +istream_class = G_INPUT_STREAM_CLASS(klass);
 +istream_class-read_fn = spice_vmc_input_stream_read;
 +istream_class-read_async = spice_vmc_input_stream_read_async;
 +istream_class-read_finish = spice_vmc_input_stream_read_finish;
 +istream_class-skip = spice_vmc_input_stream_skip;
 +

Re: [Spice-devel] [PATCH spice-gtk 12/14] Make SpiceURI a public API

[Christophe Fergeau]

On Wed, Feb 12, 2014 at 11:19:10AM +0100, Marc-André Lureau wrote:
 +G_GNUC_INTERNAL
 +void spice_uri_set_user(SpiceURI *self, const gchar *value)
 +{
 +g_return_if_fail(SPICE_IS_URI(self));
 +
 +g_free(self-user);
 +self-user = g_strdup(value);
 +g_object_notify((GObject *)self, user);
 +}
 +
 +G_GNUC_INTERNAL
 +const gchar* spice_uri_get_password(SpiceURI *self)
 +{
 +g_return_val_if_fail(SPICE_IS_URI(self), NULL);
 +return self-password;
 +}

This whole file looks like an older version (G_GNUC_INTERNAL was not there
in v1, and there was API documentation before each func).

Christophe


pgp7AmTtsNJQ3.pgp
Description: PGP signature
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

Re: [Spice-devel] [spice-gtk 06/13] http-proxy: add https proxy

[Marc-André Lureau]

On Wed, Feb 12, 2014 at 12:46 PM, Christophe Fergeau
cferg...@redhat.com wrote:
 On Wed, Feb 12, 2014 at 12:04:25PM +0100, Marc-André Lureau wrote:
 On Wed, Feb 12, 2014 at 11:59 AM, Christophe Fergeau
 cferg...@redhat.com wrote:
  Missed that initially, but tlsconn is leaked in error cases, and I think we
  will leak a reference to it in success case too. My understanding is that

 if error, function returns NULL

 tlsconn = g_tls_client_connection_new (io_stream,
G_SOCKET_CONNECTABLE(proxy_address),
error);

 if (!g_tls_connection_handshake (G_TLS_CONNECTION (tlsconn), cancellable, 
 error))
 goto error;

 error:
   if (data_in != NULL)
 g_object_unref (data_in);

   g_free (buffer);
   return NULL;
 }

 tlsconn will be leaked.

ack




  we don't own the io_stream passed as an argument, so we need to ref it in
  the http case, but in the https case, we already own a ref on the io_stream
  (which is tlsconn), so we ref it one too many.

 Where do you see we ref something?
 I remember I did check with valgrind and debugging than all references
 where correct. But I could be wrong. Please give further details.


   tlsconn = g_tls_client_connection_new (io_stream,
  G_SOCKET_CONNECTABLE(proxy_address),
  error);
   io_stream = tlsconn;

   return g_object_ref (io_stream);

 I think we are taking an extra ref on tlsconn here (ref which is probably
 needed in the !tls case).
  tlsconn is leaked here (dunno if it's guaranteed to be NULL when error is
  set).
 
 Yes it is guaratee, I can add a warning and  g_clear_object if you want.

 a g_warn_if_fail() cannot hurt as it's not explicit in the API doc (or
 fixing the API doc would work too ;)

I'll check for return value instead.



  +  return;
  +}
  +
  +  g_return_if_fail (tlsconn != NULL);
  +
  +  GTlsCertificateFlags tls_validation_flags = 
  G_TLS_CERTIFICATE_VALIDATE_ALL;
  +#ifdef DEBUG
  +  tls_validation_flags -= G_TLS_CERTIFICATE_UNKNOWN_CA + 
  G_TLS_CERTIFICATE_BAD_IDENTITY;
  +#endif
  +  g_tls_client_connection_set_validation_flags 
  (G_TLS_CLIENT_CONNECTION (tlsconn),
  +
  tls_validation_flags);
  +  g_tls_connection_handshake_async (G_TLS_CONNECTION (tlsconn),
  +G_PRIORITY_DEFAULT, cancellable,
  +handshake_completed, data);
 
  A g_object_unref(tlsconn) is probably needed here.
 

 no, the async is running and we need to keep the ref.

 Ah, my bad, I thought glib async functions were ref'ing their arg as
 needed, but I'm wrong from a quick glance at gio/glib-networking.

 Christophe



-- 
Marc-André Lureau
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel

[Spice-devel] using scan code from hardware keyboards

[i iordanov]

Hey guys,

When my hardware keyboard is connected/paired with my Android device, key
events do contain a scan code. I would love to just send that scan code
through to the VM, however, the Android documentation is really
discouraging on this matter.

Can you guys take a look at the docs for the method which retrieves the
scan code, and tell me whether you would rely on the value returned by
getScanCode() given what's written?

http://developer.android.com/reference/android/view/KeyEvent.html#getScanCode%28%29

In addition, my hardware keyboard has an English layout, and I have no way
of testing what a hardware keyboard with a non-English layout generates
when one presses a key which is not in the ASCII table. The reason it's
interesting is that Android key events may contain a key code, and  scan
code. The documented key codes for Android are all within the ASCII table.
When a soft-keyboard KeyEvent carries unicode data outside the ASCII table,
its key code == 0, and its action is not DOWN or UP, but MULTIPLE. So,
will (for example) a German hardware keyboard on which u-umlaut is pressed
generate an event with action == ACTION_MULTIPLE and a key code == 0, or
will it generate an event with action == DOWN/UP and key code set to
something inaccurate? Does anybody own an Android and a hardware keyboard
with a non-English layout to help me test this?

As things stand, if I don't trust the scan code the keyboard generates, I
cannot use my English layout hardware keyboard to type any language other
than English because the scan codes are simulated based on the unicode
character the event contains. On my English keyboard, the event inevitably
contains unicode characters that fall within the ASCII table, as you can
imagine :).

Your thoughts and suggestions are welcome.

Thanks!
iordan

-- 
The conscious mind has only one thread of execution.
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel