From rhowell at novell.com Mon Aug 4 14:23:13 2008 From: rhowell at novell.com (Rusty Howell) Date: Mon, 04 Aug 2008 12:23:13 -0600 Subject: [Moonlight-list] Moonlight status Message-ID: <48974911.1040600@novell.com> Miguel wanted to know what the major issues are right now with the Moonlight 1.0. Here are some of the biggest issues. Out dated Silverlight.js: This affects a number of sites. As per Geoff's suggestion, I've tried to find contact info on each site and sent them a message about it. Although I'm skeptical that the email will actually make it to a person who a) cares and b) can do something about it. We may just have to remove those sites from the test run. If they update, then we can add them back in. Performance / slow sites: Silverlight world map https://bugzilla.novell.com/show_bug.cgi?id=375236 Photo Gallery Wall https://bugzilla.novell.com/show_bug.cgi?id=375283 Page Turn https://bugzilla.novell.com/show_bug.cgi?id=383907 Glyphs https://bugzilla.novell.com/show_bug.cgi?id=395160 Video Tutorials https://bugzilla.novell.com/show_bug.cgi?id=410638 Kit 3D https://bugzilla.novell.com/show_bug.cgi?id=412302 Ink Journal https://bugzilla.novell.com/show_bug.cgi?id=407815 Mouse bugs: https://bugzilla.novell.com/show_bug.cgi?id=412986 https://bugzilla.novell.com/show_bug.cgi?id=411949 https://bugzilla.novell.com/show_bug.cgi?id=401406 https://bugzilla.novell.com/show_bug.cgi?id=364990 https://bugzilla.novell.com/show_bug.cgi?id=410619 Rendering: https://bugzilla.novell.com/show_bug.cgi?id=345978 https://bugzilla.novell.com/show_bug.cgi?id=352072 https://bugzilla.novell.com/show_bug.cgi?id=340730 https://bugzilla.novell.com/show_bug.cgi?id=346316 https://bugzilla.novell.com/show_bug.cgi?id=346082 From rhowell at novell.com Mon Aug 4 19:40:22 2008 From: rhowell at novell.com (Rusty Howell) Date: Mon, 04 Aug 2008 17:40:22 -0600 Subject: [Moonlight-list] Moonlight performance tests In-Reply-To: <67ed228e0807311426q4ca32bbbpd24e3fd71d138842@mail.gmail.com> References: <48920CF9.6050007@novell.com> <67ed228e0807311426q4ca32bbbpd24e3fd71d138842@mail.gmail.com> Message-ID: <48979366.60001@novell.com> The box is located at moon.mono.lab.novell.com. You'll need VPN to access it. Sorry :( MDK wrote: > On Thu, Jul 31, 2008 at 9:05 PM, Rusty Howell wrote: > >> Hey guys, >> >> Our new performance testing machine has arrived. Do we have a specific >> plan for performance tests already? What kind of setup are we going to >> need and how are the tests going to work? >> > > Hey, > > I'm looking a bit into this. First we need some reliable way to > measure performance (FPS is not 100% accurate here), second -- some > interface to probe/access it from the test software. I'm doing some > work now on the animation-screenshooting, and I'm starting to think > this might be a good starting point (in a nutshell: let's measure how > much time it takes to generate all frames of a given > animation/testcase). > > Will update on something more concrete soon, > > From ceronman at gmail.com Mon Aug 4 20:05:58 2008 From: ceronman at gmail.com (=?ISO-8859-1?Q?Manuel_Alejandro_Cer=F3n_Estrada?=) Date: Mon, 4 Aug 2008 19:05:58 -0500 Subject: [Moonlight-list] Unable to run GtkSilver apps (MultipleSurfacesPerAppDomain again) Message-ID: <796874fc0808041705s39feb5j753edcf348ffab4e@mail.gmail.com> Hi Guys. When I use DependencyObject.GetValue() I get this exception: "Multiple surfaces per domain is enabled, so calling SurfaceInDomain/PluginInDomain is wrong." Apparently, this is because NativeMethods.dependency_object_get_value calls to Mono.Xaml.XamlLoader.SurfaceInDomain, and this property can't be called from GtkSilver apps (or any MultipleSurfacesPerAppDomain app) For a test, run any of the files of this bug: https://bugzilla.novell.com/show_bug.cgi?id=386468 I'm using moon revision 109627. Thanks in advance. Manuel. From rolflists at ya.com Tue Aug 5 10:57:24 2008 From: rolflists at ya.com (Rolf Bjarne Kvinge) Date: Tue, 5 Aug 2008 16:57:24 +0200 Subject: [Moonlight-list] Unable to run GtkSilver apps (MultipleSurfacesPerAppDomain again) In-Reply-To: <796874fc0808041705s39feb5j753edcf348ffab4e@mail.gmail.com> References: <796874fc0808041705s39feb5j753edcf348ffab4e@mail.gmail.com> Message-ID: <006001c8f70b$9536b2f0$bfa418d0$@com> Hi, This should now be fixed. Rolf > -----Original Message----- > From: moonlight-list-bounces at lists.ximian.com [mailto:moonlight-list- > bounces at lists.ximian.com] On Behalf Of Manuel Alejandro Cer?n Estrada > Sent: martes, 05 de agosto de 2008 2:06 > To: moonlight-list at lists.ximian.com > Subject: [Moonlight-list] Unable to run GtkSilver apps > (MultipleSurfacesPerAppDomain again) > > Hi Guys. > > When I use DependencyObject.GetValue() I get this exception: > > "Multiple surfaces per domain is enabled, so calling > SurfaceInDomain/PluginInDomain is wrong." > > Apparently, this is because NativeMethods.dependency_object_get_value > calls to Mono.Xaml.XamlLoader.SurfaceInDomain, and this property can't > be called from GtkSilver apps (or any MultipleSurfacesPerAppDomain > app) > > For a test, run any of the files of this bug: > https://bugzilla.novell.com/show_bug.cgi?id=386468 > > I'm using moon revision 109627. > > Thanks in advance. > > Manuel. > _______________________________________________ > Moonlight-list mailing list > Moonlight-list at lists.ximian.com > http://lists.ximian.com/mailman/listinfo/moonlight-list From rhowell at novell.com Wed Aug 6 19:35:48 2008 From: rhowell at novell.com (Rusty Howell) Date: Wed, 06 Aug 2008 17:35:48 -0600 Subject: [Moonlight-list] Tagging Moonlight for 0.8 release Message-ID: <489A3554.7000806@novell.com> Hey all, I'm going to tag Moonlight on Thursday morning around 10 am MST and do a release of Moonlight 0.8 (covering SL 1.0 profile ) for the download page. Thoughts, comments, objections? Rusty From miguel at ximian.com Thu Aug 7 15:59:07 2008 From: miguel at ximian.com (Miguel de Icaza) Date: Thu, 07 Aug 2008 15:59:07 -0400 Subject: [Moonlight-list] Tagging Moonlight for 0.8 release In-Reply-To: <489A3554.7000806@novell.com> References: <489A3554.7000806@novell.com> Message-ID: <1218139147.9870.52.camel@linux-1p3a.site> Hey! > Thoughts, comments, objections? This sounds like a good idea; Hopefully we have a stable release at that point. Before we release, we should do a QA pass to ensure that the code at that point works, considering that trunk right now is getting a lot of 2.0 bits. From rolflists at ya.com Fri Aug 8 14:46:40 2008 From: rolflists at ya.com (Rolf Bjarne Kvinge) Date: Fri, 8 Aug 2008 20:46:40 +0200 Subject: [Moonlight-list] DependencyProperty registration Message-ID: <012401c8f987$1bfff6c0$53ffe440$@com> Hi, Just a little heads-up: The process of creating new (unmanaged) DependencyProperties has changed a bit (hopefully to the better, less code to write now J) So, if you ever want to create one (or more), please read the documentation in moon/tools/generators/HOWTO and/or README (or c&p any of the existing DPs J) Rolf -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ximian.com/pipermail/moonlight-list/attachments/20080808/de65be78/attachment-0001.html From michaldominik at gmail.com Mon Aug 11 13:03:52 2008 From: michaldominik at gmail.com (MDK) Date: Mon, 11 Aug 2008 19:03:52 +0200 Subject: [Moonlight-list] Moonlight Performance Suite Message-ID: <67ed228e0808111003m2ee5d2d1qba24a93db44f19e4@mail.gmail.com> Hey, I was working last week on the performance suite for Moonlight and it's almost ready for deployment. The kind of output it generates: http://www.blondeisthenight.com/perf-report/ The main idea behind this tool is to make it easy for us to spot performance regressions and have a general understanding of how the code changes affect performance -- we currently have some performance-related bugs in the Bugzilla which have "I think it used to be fine in the past..." kind of comments. Not good. The general idea of benchmarking the perf is to measure the time it takes to render a given sample "as fast as possible" using the manual time source. In this example report above there are only 3 tests, but obviously we need much more of them. Some info about the tool: * Written in C# (plus small C tool to run the tests) * Uses sqlite simple relational database for storage * Separate runner/report generator tools * Uses drtlist (similiar to the test harness) for input and test configuration We obviously need to pair this tool with Monobuild (?) for building the revisions and spawing the benchmarking process. Comments welcome, -- Michael From rhowell at novell.com Fri Aug 15 16:09:58 2008 From: rhowell at novell.com (Rusty Howell) Date: Fri, 15 Aug 2008 14:09:58 -0600 Subject: [Moonlight-list] Moonlight 0.8 Released Message-ID: <48A5E296.9090400@novell.com> Hey folks, We just released the next version of Moonlight! You can download the Firfefox plugin from http://go-mono.com/moonlight or get the sources from http://ftp.novell.com/pub/mono/sources/moon/ Install it and give it a try! If you come across a bug, let us know about it ! You can also search our current bug list to see if your bug has already been found. Rusty Howell Moonlight QA -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ximian.com/pipermail/moonlight-list/attachments/20080815/a3fae7ea/attachment.html From ajorgensen at novell.com Fri Aug 15 18:55:36 2008 From: ajorgensen at novell.com (Andrew Jorgensen) Date: Fri, 15 Aug 2008 16:55:36 -0600 Subject: [Moonlight-list] package names Message-ID: <48A5B5080200004000038A49@lucius.provo.novell.com> I'm preparing to submit moonlight to openSUSE for inclusion in 11.1 and SLE11. We need to pin down package names and contents. As it stands now I have the following: libmoon0: .so and the 2.1 and 3.0 assemblies libmoon-devel: headers and moon.pc libmoon-tools: mopen, xamlg, svg2xaml, etc. libmoon-sharp: gtksilver (including .pc and docs) libmoon-examples: example apps, including desklets moonlight-plugin: browser plugin files At the very least I want to consider renaming libmoon-sharp libmoon-tools and libmoon-examples. Perhaps moonlight-*, and maybe libmoon-sharp should be gtksilver or something like it? I've promised Aaron Bockover I'd try to get this submitted by Monday. I'd really like to have the names and content correct when I submit this. Thanks, Andrew Jorgensen From toshok at gmail.com Sun Aug 17 14:22:11 2008 From: toshok at gmail.com (Chris Toshok) Date: Sun, 17 Aug 2008 11:22:11 -0700 Subject: [Moonlight-list] new configure support for disabling browser/desktop assemblies independently Message-ID: I replaced the --with-mono configure arg with --with-managed. The possible values (--with-managed=$value) are: - desktop: attempts to configure in support for just desktop assemblies (the "3.0" profile. GtkSilver is here) - browser: attempts to configure in support for just browser assemblies (the "2.1" profile. This enables silverlight2 browser plugin support) - yes (the default when you use "--with-managed"): probes for dependencies for both. - no (specified either as "--with-managed=no" or "--without-managed"): disables managed code entirely. If you don't have the right dependencies installed for the 2 options (gtk-sharp-2.0 and rsvg-sharp-2.0 for desktop, mono >= 1.9 with smcs for browser), the support still won't be enabled, and if neither desktop nor browser dependencies are available, configure will still disable managed code entirely. It'll tell you what's going on in the final configuration output, though. for hackers: This also requires differentiating between Silverlight2 support in the src/ and plugin/ directories. In the case where moonlight is configured --with-managed=desktop, the plugin will be built only to support 1.0 (since we won't be compiling the 2.1 assemblies). So, any code that needs to be ifdefed in the plugin for 2.0-ness should use the new SL_PLUGIN_2_0 define. This will be defined iff the 2.1 assemblies are being compiled. The old SL_2_0 ifdef should still be used in src/, and will be defined if either 2.1 or 3.0 assemblies are being built. -toshok -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ximian.com/pipermail/moonlight-list/attachments/20080817/9a0c831b/attachment.html From toshok at gmail.com Sun Aug 17 18:06:31 2008 From: toshok at gmail.com (Chris Toshok) Date: Sun, 17 Aug 2008 15:06:31 -0700 Subject: [Moonlight-list] package names In-Reply-To: <48A5B5080200004000038A49@lucius.provo.novell.com> References: <48A5B5080200004000038A49@lucius.provo.novell.com> Message-ID: <8E1AE59B-67F7-4C06-BE84-E9956964A34A@gmail.com> Changing libmoon- to moonlight- sounds good to me. I think moonlight- sharp would be a good place for the 3.0 assemblies and gtksilver to be. The 2.1 assemblies could be packaged with the plugin stuff, right? Also, I'd really like to get rid of the name gtksilver, as well as silver.pc/silverdesktop.pc. Silver.pc is handled automatically by smcs anyway, and we should be using "moon" instead of "silver" on the desktop. Maybe gtkmoon? And gtksilver's namespace is Gtk.Moonlight. Maybe change that to Moonlight.Gtk.GtkMoon? -toshok On Aug 15, 2008, at 3:55 PM, "Andrew Jorgensen" wrote: > I'm preparing to submit moonlight to openSUSE for inclusion in 11.1 > and SLE11. We need to pin down package names and contents. As it > stands now I have the following: > > libmoon0: .so and the 2.1 and 3.0 assemblies > libmoon-devel: headers and moon.pc > libmoon-tools: mopen, xamlg, svg2xaml, etc. > libmoon-sharp: gtksilver (including .pc and docs) > libmoon-examples: example apps, including desklets > moonlight-plugin: browser plugin files > > At the very least I want to consider renaming libmoon-sharp libmoon- > tools and libmoon-examples. Perhaps moonlight-*, and maybe libmoon- > sharp should be gtksilver or something like it? > > I've promised Aaron Bockover I'd try to get this submitted by > Monday. I'd really like to have the names and content correct when > I submit this. > > Thanks, > Andrew Jorgensen > > _______________________________________________ > Moonlight-list mailing list > Moonlight-list at lists.ximian.com > http://lists.ximian.com/mailman/listinfo/moonlight-list From jaharper at novell.com Mon Aug 18 13:22:34 2008 From: jaharper at novell.com (Jackson Harper) Date: Mon, 18 Aug 2008 13:22:34 -0400 Subject: [Moonlight-list] [Mono-dev] Moonlight 0.8 - Where are they Release Notes In-Reply-To: <78577.31022.qm@web30804.mail.mud.yahoo.com> References: <78577.31022.qm@web30804.mail.mud.yahoo.com> Message-ID: <1219080154.11900.1.camel@linux-isls.site> On Mon, 2008-08-18 at 10:14 -0700, Daniel Morgan wrote: > As Eno pointed out, where are the release notes for Moonlight 0.8? > > Is Moonlight good enough to watch video from http://www.nbcolympics.com/ on Linux now? ?no. Most of our work has been focused on Silverlight 1.0 and that site uses a Silverlight 2.0 Beta. Jackson > --- On Fri, 8/15/08, Rusty Howell wrote: > > > From: Rusty Howell > > Subject: [Mono-list] Moonlight 0.8 Released > > To: moonlight-list at lists.ximian.com, "mono-list" , "Ximian Mono List" , "mono-devel" > > Date: Friday, August 15, 2008, 4:09 PM > > Hey folks, > > > > We just released the next version of Moonlight! > > > > You can download the Firfefox plugin from > > http://go-mono.com/moonlight > > or get the sources from > > http://ftp.novell.com/pub/mono/sources/moon/ > > > > Install it and give it a try! If you come across a bug, let > > us know > > about it > > ! > > You can also search our current bug > > > > > > list to see if your bug has already been found. > > > > Rusty Howell > > Moonlight QA > > _______________________________________________ > > Mono-list maillist - Mono-list at lists.ximian.com > > http://lists.ximian.com/mailman/listinfo/mono-list > > > > _______________________________________________ > Mono-devel-list mailing list > Mono-devel-list at lists.ximian.com > http://lists.ximian.com/mailman/listinfo/mono-devel-list From fejj at novell.com Mon Aug 18 13:22:58 2008 From: fejj at novell.com (Jeffrey Stedfast) Date: Mon, 18 Aug 2008 13:22:58 -0400 Subject: [Moonlight-list] [Mono-dev] Moonlight 0.8 - Where are they Release Notes In-Reply-To: <78577.31022.qm@web30804.mail.mud.yahoo.com> References: <78577.31022.qm@web30804.mail.mud.yahoo.com> Message-ID: <1219080178.21169.124.camel@tazmanian-devil.novell.com> On Mon, 2008-08-18 at 10:14 -0700, Daniel Morgan wrote: > As Eno pointed out, where are the release notes for Moonlight 0.8? > > Is Moonlight good enough to watch video from http://www.nbcolympics.com/ on Linux now? > Yea, we slipped up on updating the release notes file. Moonlight 0.8 only reliably works for Silverlight 1.0 sites so far. Silverlight 2.0 is still in Beta and so the API's are not yet frozen. Unfortunately, nbcolympics.com targets Silverlight 2.0 Beta and so Moonlight 0.8 will not work with it. Jeff From magowiz at gmail.com Mon Aug 18 16:37:49 2008 From: magowiz at gmail.com (Marcello Magaldi) Date: Mon, 18 Aug 2008 22:37:49 +0200 Subject: [Moonlight-list] libmoon.so: undefined symbol: XML_ParserCreateNS Message-ID: <9c037bdd0808181337v41e06014u31db6c5f8331b48@mail.gmail.com> Hi, I'm using gentoo and I have installed moonlight 0.8 using the ebuild provided in this bug report : http://bugs.gentoo.org/show_bug.cgi?id=234189, with that ebuild, after installed the novell-moonlight.xpi file in firefox (I'm using ff 2.0.0.16) if I open a silverlight page my firefox crashes with the error in subject : PluginXamlLoader::TryLoad, filename: /home/magowiz/.mozilla/firefox/ig4xtyjr.default/Cache/3A660E95d01, str: (null) /usr/lib/mozilla-firefox/firefox-bin: symbol lookup error: /home/magowiz/.mozilla/firefox/ig4xtyjr.default/extensions/moonlight at novell.com/plugins/moonlight/libmoon.so: undefined symbol: XML_ParserCreateNS firefox-bin exited with non-zero status (127) This error comes out on EVERY silverlight enabled page on my system I have mono 1.9.1 , which could be the issue? Best Regards Marcello Magaldi -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ximian.com/pipermail/moonlight-list/attachments/20080818/5f55fcf6/attachment-0001.html From ajorgensen at novell.com Mon Aug 18 17:26:44 2008 From: ajorgensen at novell.com (Andrew Jorgensen) Date: Mon, 18 Aug 2008 15:26:44 -0600 Subject: [Moonlight-list] libmoon.so: undefined symbol: XML_ParserCreateNS In-Reply-To: <48A994B40200004000038F5C@lucius.provo.novell.com> References: <48A98970020000BF00028CED@lucius.provo.novell.com> <48A994B40200004000038F5C@lucius.provo.novell.com> Message-ID: <48A994B40200004000038F5C@lucius.provo.novell.com> On Mon, 2008-08-18 at 22:37 +0200, Marcello Magaldi wrote: > Hi, > I'm using gentoo and I have installed moonlight 0.8 using the ebuild > provided in this bug report : > http://bugs.gentoo.org/show_bug.cgi?id=234189 , with that ebuild, > after installed the novell-moonlight.xpi file in firefox Using --enable-user-plugin for a local build is wrong. We should probably rename this option to --enable-firefox-plugin-installer or something. Doing both --enable-user-plugin and make install will have unexpected results. This is probably not related to your problem though. > This error comes out on EVERY silverlight enabled page > on my system I have mono 1.9.1 , which could be the issue? The Silverlight 2.0 profile is still very experimental. We do not recommend compiling against mono at this time. Again, though, not related to your problem I'd guess. It looks like the ebuild uses the system cairo and pixman, I don't know what the status of those are yet but we are not using this option. From jb at nurv.fr Tue Aug 19 06:47:10 2008 From: jb at nurv.fr (Jb Evain) Date: Tue, 19 Aug 2008 12:47:10 +0200 Subject: [Moonlight-list] WCF in Moonlight Message-ID: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> Hey all, Silverlight 2.0b2 contains three assemblies that we do not provide yet: * System.ComponentModel * System.ComponentModel.Web * System.Runtime.Serialization Currently, we have parts of those assemblies implemented in the olive module. We could either: 1) depend on olive, and tune the assemblies there. 2) move those assemblies to the mcs module, and tune them there. According to the discussion we had on IRC, it seems that 1) is a no go, as we don't want to release olive, and we rather avoid to add another dependency to moon. How to do 2) properly then? Should we simply move those assemblies to mcs, and develop there, or continue to develop in olive, and merge changes when we feel confident into mcs? I'd rather simply move them, but then we would have unstable assemblies in mcs. Should we apply the same policy we apply for Cecil, that is not exposing it to the compiler without using a pkg? Also, Atsushi, how complete is our implementation wrt what SL2 ships? -- Jb Evain From atsushi at ximian.com Tue Aug 19 08:13:57 2008 From: atsushi at ximian.com (Atsushi Eno) Date: Tue, 19 Aug 2008 21:13:57 +0900 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> Message-ID: <48AAB905.5060605@ximian.com> Hey JB, > Hey all, > > Silverlight 2.0b2 contains three assemblies that we do not provide yet: > > * System.ComponentModel > * System.ComponentModel.Web > * System.Runtime.Serialization You meant System.ServiceModel*. > Currently, we have parts of those assemblies implemented in the olive > module. We could either: > > 1) depend on olive, and tune the assemblies there. > 2) move those assemblies to the mcs module, and tune them there. > > According to the discussion we had on IRC, it seems that 1) is a no > go, as we don't want to release olive, and we rather avoid to add > another dependency to moon. > > How to do 2) properly then? Should we simply move those assemblies to > mcs, and develop there, or continue to develop in olive, and merge > changes when we feel confident into mcs? I'd rather simply move them, > but then we would have unstable assemblies in mcs. Should we apply the > same policy we apply for Cecil, that is not exposing it to the > compiler without using a pkg? Most generally, we have two choices: A) abandon existing premises (mcs => stable, olive => unstable) B) keep stability premises above. I rather prefer A) as it is closer to our fact. Stable assembly and unstable assembly can be distinguished from some explicit declaration e.g. in our documentation and/or website. Even if it comes to B), "stability" could be evaluated in two ways: 1) public API stability 2) implementation stability Since 2) has never been achieved in every assembly, and we have a lot of unsupported assemblies under mcs (e.g.System.EnterpriseServices.dll), I'd say "stability" is almost all about 1). API stability can be achieved relatively easily than implementation stability. > Also, Atsushi, how complete is our implementation wrt what SL2 ships? No idea. The last maintainers were not me but Mainsoft guys. Atsushi Eno From jb at nurv.fr Tue Aug 19 09:51:43 2008 From: jb at nurv.fr (Jb Evain) Date: Tue, 19 Aug 2008 15:51:43 +0200 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <48AAB905.5060605@ximian.com> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <48AAB905.5060605@ximian.com> Message-ID: <69f7d8470808190651l27dfa9abw3d0894752d48ab30@mail.gmail.com> Hey, On 8/19/08, Atsushi Eno wrote: > You meant System.ServiceModel*. Of course, thanks :) > Most generally, we have two choices: Another option I'm thinking about, is to have SL2 *only* versions of such assemblies in mcs. That is, the code still lives in olive, and is developed there, and we manually backport code to mcs, where those assemblies are built only for the net_2_1 profile. That would avoid to move all the WCF assemblies to mcs, such as System.IdentityModel* that will be needed for the non plugin version of System.ServiceModel. On the other hand, am thinking that if we start moving code into mcs, we might as well move all the required code for all usages, so I have mixed feelings about that option. -- Jb Evain From atsushi at ximian.com Tue Aug 19 12:39:37 2008 From: atsushi at ximian.com (Atsushi Eno) Date: Wed, 20 Aug 2008 01:39:37 +0900 Subject: [Moonlight-list] [Mono-dev] WCF in Moonlight In-Reply-To: <69f7d8470808190651l27dfa9abw3d0894752d48ab30@mail.gmail.com> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <48AAB905.5060605@ximian.com> <69f7d8470808190651l27dfa9abw3d0894752d48ab30@mail.gmail.com> Message-ID: <48AAF749.8060407@ximian.com> Hola, > Hey, > > On 8/19/08, Atsushi Eno wrote: >> You meant System.ServiceModel*. > > Of course, thanks :) > >> Most generally, we have two choices: > > Another option I'm thinking about, is to have SL2 *only* versions of > such assemblies in mcs. That is, the code still lives in olive, and is > developed there, and we manually backport code to mcs, where those > assemblies are built only for the net_2_1 profile. Ah, that sounds clever ;) In that way we could also use some sort of configuration option like --enable-preview to determine whether to build 3.0 assemblies or not. They can be skipped when it is not enabled. Atsushi Eno From jb at nurv.fr Tue Aug 19 14:00:30 2008 From: jb at nurv.fr (Jb Evain) Date: Tue, 19 Aug 2008 20:00:30 +0200 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <1219168700.4178.72.camel@linux-1p3a.site> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <48AAB905.5060605@ximian.com> <1219168700.4178.72.camel@linux-1p3a.site> Message-ID: <69f7d8470808191100j691e2512o6d85792be7ed06ec@mail.gmail.com> Hey, On 8/19/08, Miguel de Icaza wrote: > > > Silverlight 2.0b2 contains three assemblies that we do not provide yet: > > > > > > * System.ComponentModel > > > * System.ComponentModel.Web > > > * System.Runtime.Serialization > > > > You meant System.ServiceModel*. > > The assemblies shipping with Silverlight 2.0 are: > > System.Runtime.Serialization.dll > System.ServiceModel.Web.dll > System.ServiceModel.dll > > I do not see references to System.ComponentModel assemblies, at least > not with the runtime, maybe the SDK has them, but we can safely ignore > those. It was a typo in my mail, there's no System.ComponentModel assemblies in SL2. -- Jb Evain From mkostrzewa at novell.com Mon Aug 11 12:59:54 2008 From: mkostrzewa at novell.com (Michael Dominic K.) Date: Mon, 11 Aug 2008 18:59:54 +0200 Subject: [Moonlight-list] Moonlight Performance Suite Message-ID: <1218473994.28838.12.camel@nude.site> Hey, I was working last week on the performance suite for Moonlight and it's almost ready for deployment. The kind of output it generates: http://www.blondeisthenight.com/perf-report/ The main idea behind this tool is to make it easy for us to spot performance regressions and have a general understanding of how the code changes affect performance -- we currently have some performance-related bugs in the Bugzilla which have "I think it used to be fine in the past..." kind of comments. Not good. The general idea of benchmarking the perf is to measure the time it takes to render a given sample "as fast as possible" using the manual time source. In this example report above there are only 3 tests, but obviously we need much more of them. Some info about the tool: * Written in C# (plus small C tool to run the tests) * Uses sqlite simple relational database for storage * Separate runner/report generator tools * Uses drtlist (similiar to the test harness) for input and test configuration We obviously need to pair this tool with Monobuild (?) for building the revisions and spawing the benchmarking process. Comments welcome, -- Michael From kumpera at gmail.com Mon Aug 11 14:52:28 2008 From: kumpera at gmail.com (Rodrigo Kumpera) Date: Mon, 11 Aug 2008 15:52:28 -0300 Subject: [Moonlight-list] [Ximian-mono-list] Moonlight Performance Suite In-Reply-To: <1218473994.28838.12.camel@nude.site> References: <1218473994.28838.12.camel@nude.site> Message-ID: <8cca42d80808111152u1aa6b159w50bc46dbd5f76367@mail.gmail.com> On Mon, Aug 11, 2008 at 1:59 PM, Michael Dominic K. wrote: > Hey, > > We obviously need to pair this tool with Monobuild (?) for building the > revisions and spawing the benchmarking process. > Geoff and I will be working on monobuild next week as part of HackWeek, we sure will be looking into this integration. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ximian.com/pipermail/moonlight-list/attachments/20080811/0d313dbc/attachment-0001.html From kumpera at gmail.com Mon Aug 11 14:57:18 2008 From: kumpera at gmail.com (Rodrigo Kumpera) Date: Mon, 11 Aug 2008 15:57:18 -0300 Subject: [Moonlight-list] [Ximian-mono-list] Moonlight Performance Suite In-Reply-To: <1218473994.28838.12.camel@nude.site> References: <1218473994.28838.12.camel@nude.site> Message-ID: <8cca42d80808111157o54f04b16w5ec93f15e68469eb@mail.gmail.com> Michael, Did you take a look at cattrack for doing performance regression reporting? It's pretty complete and simple to use. My plan for mono is to try this tool first. http://jikesrvm.anu.edu.au/cattrack The source code you can find in the jikesrvm SVN. Cheers, Rodrigo On Mon, Aug 11, 2008 at 1:59 PM, Michael Dominic K. wrote: > Hey, > > I was working last week on the performance suite for Moonlight and it's > almost ready for deployment. The kind of output it generates: > > http://www.blondeisthenight.com/perf-report/ > > The main idea behind this tool is to make it easy for us to spot > performance regressions and have a general understanding of how the code > changes affect performance -- we currently have some performance-related > bugs in the Bugzilla which have "I think it used to be fine in the > past..." kind of comments. Not good. > > The general idea of benchmarking the perf is to measure the time it > takes to render a given sample "as fast as possible" using the manual > time source. > > In this example report above there are only 3 tests, but obviously we > need much more of them. > > Some info about the tool: > > * Written in C# (plus small C tool to run the tests) > * Uses sqlite simple relational database for storage > * Separate runner/report generator tools > * Uses drtlist (similiar to the test harness) for input and test > configuration > > We obviously need to pair this tool with Monobuild (?) for building the > revisions and spawing the benchmarking process. > > Comments welcome, > > -- > Michael > > _______________________________________________ > Ximian-mono-list mailing list > Ximian-mono-list at lists.ximian.com > http://lists.ximian.com/mailman/listinfo/ximian-mono-list > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ximian.com/pipermail/moonlight-list/attachments/20080811/4c8183b8/attachment.html From monodanmorg at yahoo.com Mon Aug 18 13:14:08 2008 From: monodanmorg at yahoo.com (Daniel Morgan) Date: Mon, 18 Aug 2008 10:14:08 -0700 (PDT) Subject: [Moonlight-list] Moonlight 0.8 - Where are they Release Notes In-Reply-To: <48A5E296.9090400@novell.com> Message-ID: <78577.31022.qm@web30804.mail.mud.yahoo.com> As Eno pointed out, where are the release notes for Moonlight 0.8? Is Moonlight good enough to watch video from http://www.nbcolympics.com/ on Linux now? --- On Fri, 8/15/08, Rusty Howell wrote: > From: Rusty Howell > Subject: [Mono-list] Moonlight 0.8 Released > To: moonlight-list at lists.ximian.com, "mono-list" , "Ximian Mono List" , "mono-devel" > Date: Friday, August 15, 2008, 4:09 PM > Hey folks, > > We just released the next version of Moonlight! > > You can download the Firfefox plugin from > http://go-mono.com/moonlight > or get the sources from > http://ftp.novell.com/pub/mono/sources/moon/ > > Install it and give it a try! If you come across a bug, let > us know > about it > ! > You can also search our current bug > > > list to see if your bug has already been found. > > Rusty Howell > Moonlight QA > _______________________________________________ > Mono-list maillist - Mono-list at lists.ximian.com > http://lists.ximian.com/mailman/listinfo/mono-list From jb at nurv.fr Tue Aug 19 14:08:19 2008 From: jb at nurv.fr (Jb Evain) Date: Tue, 19 Aug 2008 20:08:19 +0200 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <1219167914.4178.67.camel@linux-1p3a.site> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <1219167914.4178.67.camel@linux-1p3a.site> Message-ID: <69f7d8470808191108o1a926424w7ef7c892e78f2e2a@mail.gmail.com> Hey, On 8/19/08, Miguel de Icaza wrote: > Lets move the assemblies into mcs. Ok. That will require to move the folders: * System.Runtime.Serialization * System.IdentityModel * System.IdentityModel.Selectors * System.ServiceModel * System.ServiceModel.Web from olive to mcs. Right now, for the desktop case, they're built and installed in the 3.0 profile of olive, and installed in $(prefix)/mono/lib/3.0, people have to use -pkg:olive to reference them. Do we want to preserve that when being built in mcs, or should we simply build and install them in the 2.0 profile? -- Jb Evain From jb at nurv.fr Tue Aug 19 14:51:01 2008 From: jb at nurv.fr (Jb Evain) Date: Tue, 19 Aug 2008 20:51:01 +0200 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <1219171371.4178.91.camel@linux-1p3a.site> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <1219167914.4178.67.camel@linux-1p3a.site> <69f7d8470808191108o1a926424w7ef7c892e78f2e2a@mail.gmail.com> <1219171371.4178.91.camel@linux-1p3a.site> Message-ID: <69f7d8470808191151h683948b9n3abf62846117c835@mail.gmail.com> On 8/19/08, Miguel de Icaza wrote: > I guess it should go into the 2.0 directory, to be > a default reference, and we should also add some new .pc files -pkg:wcf > and -pkg:cardspace > > Does that make sense? > Makes sense to me, I'll work on moving them, integrating them in the build and adding support for them in the tuner tomorrow. -- Jb Evain From miguel at novell.com Tue Aug 19 14:42:51 2008 From: miguel at novell.com (Miguel de Icaza) Date: Tue, 19 Aug 2008 14:42:51 -0400 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <69f7d8470808191108o1a926424w7ef7c892e78f2e2a@mail.gmail.com> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <1219167914.4178.67.camel@linux-1p3a.site> <69f7d8470808191108o1a926424w7ef7c892e78f2e2a@mail.gmail.com> Message-ID: <1219171371.4178.91.camel@linux-1p3a.site> Hello, > On 8/19/08, Miguel de Icaza wrote: > > Lets move the assemblies into mcs. > > Ok. That will require to move the folders: > > * System.Runtime.Serialization > * System.IdentityModel > * System.IdentityModel.Selectors > * System.ServiceModel > * System.ServiceModel.Web > > from olive to mcs. > > Right now, for the desktop case, they're built and installed in the > 3.0 profile of olive, and installed in $(prefix)/mono/lib/3.0, people > have to use -pkg:olive to reference them. Do we want to preserve that > when being built in mcs, or should we simply build and install them in > the 2.0 profile? Mhm, good question. I guess it should go into the 2.0 directory, to be a default reference, and we should also add some new .pc files -pkg:wcf and -pkg:cardspace Does that make sense? From miguel at novell.com Tue Aug 19 15:51:36 2008 From: miguel at novell.com (Miguel de Icaza) Date: Tue, 19 Aug 2008 15:51:36 -0400 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <48AB1786.7050009@gmail.com> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <1219167914.4178.67.camel@linux-1p3a.site> <48AB1786.7050009@gmail.com> Message-ID: <1219175496.4178.114.camel@linux-1p3a.site> > We were discussing about a similar issue inside the UIA team. Can we > then take a similar approach for our assemblies in olive when our > release dates arrive? We will discuss that case later. From atsushi at ximian.com Tue Aug 19 21:13:42 2008 From: atsushi at ximian.com (Atsushi Eno) Date: Wed, 20 Aug 2008 10:13:42 +0900 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <1219171371.4178.91.camel@linux-1p3a.site> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <1219167914.4178.67.camel@linux-1p3a.site> <69f7d8470808191108o1a926424w7ef7c892e78f2e2a@mail.gmail.com> <1219171371.4178.91.camel@linux-1p3a.site> Message-ID: <48AB6FC6.7020902@ximian.com> Miguel de Icaza wrote: > Hello, > >> On 8/19/08, Miguel de Icaza wrote: >>> Lets move the assemblies into mcs. >> Ok. That will require to move the folders: >> >> * System.Runtime.Serialization >> * System.IdentityModel >> * System.IdentityModel.Selectors >> * System.ServiceModel >> * System.ServiceModel.Web >> >> from olive to mcs. >> >> Right now, for the desktop case, they're built and installed in the >> 3.0 profile of olive, and installed in $(prefix)/mono/lib/3.0, people >> have to use -pkg:olive to reference them. Do we want to preserve that >> when being built in mcs, or should we simply build and install them in >> the 2.0 profile? > > Mhm, good question. I guess it should go into the 2.0 directory, to be > a default reference, and we should also add some new .pc files -pkg:wcf > and -pkg:cardspace Do we really want such individual packages, especially even if we come up with WPF and WF (WF assemblies actually exists) ? Since they are not supported assemblies, we could rather live without them, or something like pkg:dotnet-preview would be better (just to replace olive, minus UIA stuff). At least -pkg:wcf and -pkg:cardspace sounds identical. Atsushi Eno From jaharper at novell.com Mon Aug 18 17:15:53 2008 From: jaharper at novell.com (Jackson Harper) Date: Mon, 18 Aug 2008 17:15:53 -0400 Subject: [Moonlight-list] libmoon.so: undefined symbol: XML_ParserCreateNS In-Reply-To: <9c037bdd0808181337v41e06014u31db6c5f8331b48@mail.gmail.com> References: <9c037bdd0808181337v41e06014u31db6c5f8331b48@mail.gmail.com> Message-ID: <1219094153.11900.3.camel@linux-isls.site> Do you have libexpat installed? Jackson On Mon, 2008-08-18 at 22:37 +0200, Marcello Magaldi wrote: > Hi, > I'm using gentoo and I have installed moonlight 0.8 using the ebuild > provided in this bug report : > http://bugs.gentoo.org/show_bug.cgi?id=234189 , with that ebuild, > after installed the novell-moonlight.xpi file in firefox (I'm using ff > 2.0.0.16) if I open a silverlight page my firefox crashes with the > error in subject : > PluginXamlLoader::TryLoad, filename: > /home/magowiz/.mozilla/firefox/ig4xtyjr.default/Cache/3A660E95d01, str: (null) > /usr/lib/mozilla-firefox/firefox-bin: symbol lookup error: > /home/magowiz/.mozilla/firefox/ig4xtyjr.default/extensions/moonlight at novell.com/plugins/moonlight/libmoon.so: > > undefined symbol: XML_ParserCreateNS > firefox-bin exited with non-zero status (127) > > This error comes out on EVERY silverlight enabled page > on my system I have mono 1.9.1 , which could be the issue? > > Best Regards > > > Marcello Magaldi > > > _______________________________________________ > Moonlight-list mailing list > Moonlight-list at lists.ximian.com > http://lists.ximian.com/mailman/listinfo/moonlight-list From magowiz at gmail.com Wed Aug 20 08:41:29 2008 From: magowiz at gmail.com (Marcello Magaldi) Date: Wed, 20 Aug 2008 14:41:29 +0200 Subject: [Moonlight-list] libmoon.so: undefined symbol: XML_ParserCreateNS In-Reply-To: <1219094153.11900.3.camel@linux-isls.site> References: <9c037bdd0808181337v41e06014u31db6c5f8331b48@mail.gmail.com> <1219094153.11900.3.camel@linux-isls.site> Message-ID: <48AC10F9.5060603@gmail.com> Jackson Harper ha scritto: > Do you have libexpat installed? > sure, on my system I have expat 2.0.1 . Anyway it seems that somehow on building moonlight doesn't link against expat , since ldd /usr/lib/libmoon.so doesn't report any linkage to expat , the ebuild maintainer is working on it... If you want to know everything about it you can follow this bug report : http://bugs.gentoo.org/show_bug.cgi?id=234189 > Jackson > > > On Mon, 2008-08-18 at 22:37 +0200, Marcello Magaldi wrote: > >> Hi, >> I'm using gentoo and I have installed moonlight 0.8 using the ebuild >> provided in this bug report : >> http://bugs.gentoo.org/show_bug.cgi?id=234189 , with that ebuild, >> after installed the novell-moonlight.xpi file in firefox (I'm using ff >> 2.0.0.16) if I open a silverlight page my firefox crashes with the >> error in subject : >> PluginXamlLoader::TryLoad, filename: >> /home/magowiz/.mozilla/firefox/ig4xtyjr.default/Cache/3A660E95d01, str: (null) >> /usr/lib/mozilla-firefox/firefox-bin: symbol lookup error: >> /home/magowiz/.mozilla/firefox/ig4xtyjr.default/extensions/moonlight at novell.com/plugins/moonlight/libmoon.so: >> >> undefined symbol: XML_ParserCreateNS >> firefox-bin exited with non-zero status (127) >> >> This error comes out on EVERY silverlight enabled page >> on my system I have mono 1.9.1 , which could be the issue? >> >> Best Regards >> >> >> Marcello Magaldi >> >> >> _______________________________________________ >> Moonlight-list mailing list >> Moonlight-list at lists.ximian.com >> http://lists.ximian.com/mailman/listinfo/moonlight-list >> > > From miguel at novell.com Wed Aug 20 13:17:11 2008 From: miguel at novell.com (Miguel de Icaza) Date: Wed, 20 Aug 2008 13:17:11 -0400 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <48AB6FC6.7020902@ximian.com> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <1219167914.4178.67.camel@linux-1p3a.site> <69f7d8470808191108o1a926424w7ef7c892e78f2e2a@mail.gmail.com> <1219171371.4178.91.camel@linux-1p3a.site> <48AB6FC6.7020902@ximian.com> Message-ID: <1219252631.4178.149.camel@linux-1p3a.site> > At least -pkg:wcf and -pkg:cardspace sounds identical. > Ok, then lets just use wcf. From knocte at gmail.com Tue Aug 19 14:57:10 2008 From: knocte at gmail.com (=?ISO-8859-1?Q?=22Andr=E9s_G=2E_Aragoneses=22?=) Date: Tue, 19 Aug 2008 20:57:10 +0200 Subject: [Moonlight-list] WCF in Moonlight In-Reply-To: <1219167914.4178.67.camel@linux-1p3a.site> References: <69f7d8470808190347h1e7d471bq7154d987e3ff2cf3@mail.gmail.com> <1219167914.4178.67.camel@linux-1p3a.site> Message-ID: <48AB1786.7050009@gmail.com> Miguel de Icaza wrote: >> How to do 2) properly then? Should we simply move those assemblies to >> mcs, and develop there, or continue to develop in olive, and merge >> changes when we feel confident into mcs? I'd rather simply move them, >> but then we would have unstable assemblies in mcs. Should we apply the >> same policy we apply for Cecil, that is not exposing it to the >> compiler without using a pkg? > > Lets move the assemblies into mcs. There are plenty of pros and cons > to various other approaches (like stabilizing Olive and depending on it) > but it will be cumbersome and will add another dependency. We were discussing about a similar issue inside the UIA team. Can we then take a similar approach for our assemblies in olive when our release dates arrive? Andr?s -- From rhowell at novell.com Mon Aug 25 13:01:34 2008 From: rhowell at novell.com (Rusty Howell) Date: Mon, 25 Aug 2008 11:01:34 -0600 Subject: [Moonlight-list] moonlight performance box is up and running Message-ID: <48B2E56E.1010902@novell.com> Hey all, The Moonlight performance box is up and running. You can login to it via moon-perf.mono-project.com or 151.155.5.147 Rusty From oe014b3624_3 at blueyonder.co.uk Mon Aug 25 17:40:55 2008 From: oe014b3624_3 at blueyonder.co.uk (O. Emmerson) Date: Mon, 25 Aug 2008 22:40:55 +0100 Subject: [Moonlight-list] Video from Democratic Convention 2008 using Silverlight Message-ID: <48B326E7.2040609@blueyonder.co.uk> Hi, I went to http://gallery.demconvention.com/ to watch some of the videos and found they are using Silverlight. I know roughly what Silverlight/Moonlight is but have not followed it's development closely. Can I please get some feedback from people who are already using Moonlight if it is mature enough to play the videos I want to see before I go to the trouble of compiling it all. Thanks From miguel at ximian.com Mon Aug 25 20:25:32 2008 From: miguel at ximian.com (Miguel de Icaza) Date: Mon, 25 Aug 2008 20:25:32 -0400 Subject: [Moonlight-list] Video from Democratic Convention 2008 using Silverlight In-Reply-To: <48B326E7.2040609@blueyonder.co.uk> References: <48B326E7.2040609@blueyonder.co.uk> Message-ID: <1219710332.5649.44.camel@linux-1p3a.site> > I went to http://gallery.demconvention.com/ to watch some of the videos > and found they are using Silverlight. > > I know roughly what Silverlight/Moonlight is but have not followed it's > development closely. > > Can I please get some feedback from people who are already using > Moonlight if it is mature enough to play the videos I want to see before > I go to the trouble of compiling it all. It is currently using Silverlight 2.0, which we are currently working on and is at this point not ready for use. But additionally, they seem to use another plugin to download files which is not Silverlight and that we have no idea how it work. From mdk at mdk.am Tue Aug 26 05:47:09 2008 From: mdk at mdk.am (Michael Dominic K.) Date: Tue, 26 Aug 2008 11:47:09 +0200 Subject: [Moonlight-list] moonlight performance box is up and running In-Reply-To: <48B2E56E.1010902@novell.com> References: <48B2E56E.1010902@novell.com> Message-ID: <67ed228e0808260247u6623839es2dcb61555151eef7@mail.gmail.com> On Mon, Aug 25, 2008 at 7:01 PM, Rusty Howell wrote: > Hey all, > > The Moonlight performance box is up and running. You can login to it via > moon-perf.mono-project.com or 151.155.5.147 Stupid question, what to use as a login? ;) -- Michael.