Re: [Gnash-dev] [PATCH] Small rtmp_msg.cpp refactoring
On 08/12/09 02:22, stefasab wrote: On date Monday 2009-08-10 12:07:19 -0600, Rob Savoye gnashed: stefasab wrote: Rob, reading the source of cygnal I can't still figure out how the server side application is supposed to work, also I'm not still sure if there is already support for that. Yes, but I've been refactoring that part of Cygnal for the last few days in a branch. The new way supports server side scripting as a dynamically loaded plugin, so one can build a plugin out of the Gnash source tree for Cygnal. Along with that technique, Cygnal also supports remote server side scripting and can use a network connection to funnel data to and from the server script support. This also lets Cygnal talk to Cygnal, one working as the backend for the other. This is all reasonably complex, you might find IRC (#gnash @ freenode) is better for some of this than email. The cool part of the new code is that the data is now independent of the transport protocol, so it's possible to exchange data between HTTP and RTMP based connections. That is, how can I write for example a simple echo test server side application to interact with an echo client side application? That works now, with both RTMPT (HTTP) and RTMP. I took the Red5 echo test and converted it to Ming syntax. This sends a variety of complex AMF data types, including nested objects, Custom Classes, etc... - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Flash cookies and GNU Gnash
John Gilmore wrote: It's not the default, unfortunately. Perhaps when we tried making those cookies non-writeable, too many web sites stop working. Since Flash cookies (.sol files) are heavily used by many swf files for storing configuration settings, like game scores. I've looked at hundreds, and at least never saw anything privacy invading yet... But the ability to store executable code should scare people... I'm not so worried about SharedObjects as I am LocalConnection when it comes to security. option to Delete all Shared Object files (flash cookies) on exit, or Only keep Shared Object files in temporary RAM would protect users, while not breaking the web sites that attempt to sneak up on them. We do print out what is written to the .sol file, which is better than nothing. We could do a lot more like integrating support in the GUI to display, edit, etc... .sol files as used by the playing swf file. Patches gladly excepted. :-) We mention flash cookies once in our documentation, but we should use that term throughout the doc and the program, now that it's become the publicly accepted term. It's cause our documentation hasn't barely been updated in over a year... Gnash also provides a soldumper command-line program that can decode what's inside those .SOL files. It works on .SOL files created by Adobe Flash as well as those created by Gnash. Example: One of these days I intend to add support to soldumper to do much more than it does now. Most of the other Gnash developers ignore the little utilities Gnash has. I'd like to add an option to recursively find all the .sol file in ones $HOME/.macromedia directory, and to dump all the data. It'd also be nice to add the ability to modify the .sol file without using flash. With the release coming up, I know I don't have time to add this support before 0.8.6, but would love a patch. :-) I'm shocked that I have this many. I'd thought gnash was protecting me from them. You unfortunately need them for proper functionality for many things, like media players. soldumper -l is supposed to list all the cookies (like the above), but it doesn't actually work in the current release. In 0.8.5 ? It should, but is working in trunk, so will make the next release. which should be out in Sept. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: no AVM2 for a while?
Matthias Kramm wrote: Anything specific you need in as3compile for more extensive test cases? I haven't tried to keep up to date on as3compile this summer, as once I created a new testsuite with Haxe in May, we kindof got stuck with it. One issue is namely support for older swf versions. Haxe supports at least back to v6. The advantage to this is that for an API testsuite, which is basically what we have for ActionScript, can be used to cover all swf versions from v6 to v10. So to use as3compile, we'd be stuck maintaining (as now anyway), two ActionScript testsuites for the same classes, one for v9/10, and one for older versions. This isn't very efficient, as maintaining good test cases takes a lot of time. I did start on a Haxe to as3compile conversion script once, they're similar enough. The other issue that you probably can address is we use the type functions in Haxe pretty heavily in the test cases, as for API testing, the return data type, parameters, etc... is all very important. Maybe as3compile or obscure ActionScript commands can do the same thing, I don't know. We've found more than a few bugs in Haxe, and also found workarounds for what was needed. I'm not super happy with Haxe, but the two mentioned issues have us stuck there for now. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
[Gnash-dev] code freeze coming...
Now that the summer intern project is over, and the various distributions have their own code freezes in a month, it's that time real soon now... I'd like to know if anyone has any major work that is expected not to be done before early next week that was supposed to be in the release. Ideally there is isn't any. :-) It'd be nice to cleanup any random core dumping of Gnash or testsuite oddities before the freeze instead of after for a change... I would like to keep this release focused on just making a stable, player with the last 5 months of work included, instead of trying to pack as much into it as possible. I'll be honest, I'll probably be making Cygnal hacks till the very end, but these will only be to files in the cygnal directory, which won't effect anything else. This is mostly cause I'm just finishing up a big pile of refactoring changes, which I assume could take another week to fix all potential problems testing will find. I'm volunteering at a local music festival this weekend (http://www.NedFest.com), so Thurs-Sun I'm totally away from all computers, so won't be able to get this done by the freeze. :-) As I'm the one actually doing the release, I figure it's my own problem... - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] make check failing -- dejagnu problem?
On 08/30/09 12:21, strk wrote: We have an alternative used by other testsuites (testsuite/check.h) but libmedia.all seems not using it (see top of that file to note how it handles that situation). I am trying to depreciate the usage of the check.h and check.as header files. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Gnash cross compiler
On 09/01/09 08:57, JJON wrote: I am a student form Jinan University in China,I want to cross compile gnash for ARM9 architecture 's embedded Linux.But i don't know how to start it .I found some information about it but it is fragmentary. * *I've a gcc based cross compiler available in ubuntu. My Gnash is 0.8.5 released1)Can somebody tell me how to do this, especially modifying the configure script for the new target? I currently do ARM builds for Android, although I've built for other ARM targets in the past. This page will likely be of some help: http://wiki.gnashdev.org/Building_for_Android as will the other links here: http://wiki.gnashdev.org/Gnash#Building. You don't have to modify the configure script, you only need to use the proper options when configuring. The important thing to remember is to use the --top-level= option when configuring. Usually the only headache is building all the dependent packages like ffmpeg, boost, etc... that Gnash requires. Help on that is at: http://wiki.gnashdev.org/Building_dependant_packages 2)Can somebody tell me which module or function can be cutted, i just want gnash can play .swf file in a embedded Linux Try --enable-media=none --enable-gui=sdl. You can't really leave much out of Gnash. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] problem compiling 0.8.5 and trunk
On 09/02/09 08:58, dan...@rbzrobotdesign.com wrote: I'm working with gnash in order to have it working under freescale powerpc platform. I have been able to compile and use version 0.8.3 against FB using AGG. Later I started playing with version 0.8.5 and trunk. My system is ubuntu 9.04 and gcc is 4.2.4. Should work fine, I've run Gnash on multiple 32 and 64 bit PPC platforms. I have to manually set the SDL_CFLAGS in libbase/Makefile in order to compile, the system doesn't set this variable with the SDL header path. Later I in the compiling process I get this: There are the --enable-sdl-incl= and --enable-sdl-lib= to adjust the path when configuring. Most all the packages Gnash depends on have similar options. When cross-compiling, use the --top-level directory to point to your sysroot where all the other files are located. processor.cpp: In function 'int main(int, char**)': processor.cpp:354: error: 'handler' was not declared in this scope make[2]: *** [processor.o] Error 1 Hum, I don't remember that problem in 0.8.5, but it shouldn't be in trunk. I'm very close to a huge checkin to trunk effecting all the RTMP code, the Handler class has been completely refactored, and Cygnal's network engine has also been totally refactored... RTMP streaming is now starting to work... :-) ./configure --enable-gui=fb --enable-renderer=agg --enable-media=none --disable-shared --disable-menus --disable-plugins - With trunk version the compilation stops when compiling files related to the microphone. The configure call is the same. GST I've discovered doesn't cross compile now, I plan to fix this as I start the release testing process for 0.8.6. (sometime in the next few days) At the same time I'll fix the media=none (and ffmpeg) build problems. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] libltdlc not compiled anymore
On 09/07/09 07:30, Andrea Palmatè wrote: i've seen that my Makefile.am of libltdlc is empty. can someone confirm this? libltdl* is built in libbase, not the libltdl directory created by libtoolize. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] setting breakpoints in a supporting library
On 09/07/09 10:21, dww wrote: I am trying to debug gtk-gnash to become more familiar with the code and gain experience with gdb debugging. I was trying to set breakpoints in IOChannel.cpp I first did a symbol-file /usr/local/lib/gnash/libgnashbase-trunk.so then I set breakpoint in IOChannel.cpp Loading the symbol table isn't going to do it. To test an installed gnash, run GDB on the gtk-gnash (or kde4-gnash, etc...) executable. Use the GDB command set args to specify the command line for Gnash, including the name of the swf file you are running. Then it'll work. I use .gdbinit files to set the default commands for Gnash when debugging, including doing a tb main followed by run, which loads the symbol table making setting breakpoints easier. Otherwise just answer y when GDB prompts you about setting a breakpoint in a dynamically loaded library. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] setting breakpoints in a supporting library
On 09/07/09 12:41, dww wrote: at this point the symbols for libgnashbase-trunk.so were not loaded and when I set a pending breakpoint with the br command it still does not get hit. Did you compile with -g ? Otherwise no debug symbols. :-) - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] setting breakpoints in a supporting library
On 09/07/09 12:46, dww wrote: I followed the reference manual make CFLAGS=-g CXXFLAGS=-g correct? Yep, that should do it. Did you install the debug version ? I can try again to be sure nm --demangle will print lots of symbols if compiled with -g. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] cairo instead of agg for fb?
On 09/07/09 14:36, Leon Woestenberg wrote: Currently only agg is supported. Cairo / pixman recently got good ARM NEON optimizations that are worth exploring. When configuring Gnash, use --enable-renderer=cairo. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Building Gnash for the TDS/Trimble Nomad
On 09/09/09 09:30, Udo Giacomozzi wrote: This mail is mainly for Rob (who will be cross-compiling Gnash for the Trimble Nomad too), but might be useful for anybody who wants to build Gnash for that device. I'll try to expand this, but here's a start on my notes for Nomad: http://wiki.gnashdev.org/Building_for_Nomad. Last time I tried this, it worked well if you do NOT include any media handler, but that is what I'm trying to do right now (to have audio support). Good luck. From looking through the Nomad SDK, Gstreamer appears to be only partially installed. You might have better luck with ffmpeg directly. I did fix Gnash's configure to better handle a broken and partially installed Gstreamer, so it'll now build fine. I just had to disable some features, like auto-probing for input devices like a webcam. (you can set this in your .gnashrc file anyway) In the meantime there is a newer one, which I'm testing right now: http://developer.sdgsystems.com/sdk/sdk-addon-qtex4.4.3-release1.tgz I used the newer one, with my changes. The Nomad comes with an (old) GST pre-installed, but some other libs are missing. I downloaded and built them separately (for running Gnash I put the .so files in the same directory, for simplicity): Which I cleaned up an reorganized to work the standard way. The improved sysroot that successfully builds Gnash is at: http://www.gnashdev.org/tools/sdg/arm-angstrom-linux-gnueabi.tar.bz2 I invoked configure with these options: ../trunk/configure --with-top-level=/opt/sdg/angstrom2007.12/sdk/arm-angstrom-linux-gnueabi --host=arm-angstrom-linux-gnueabi --build=i686-pc-linux-gnu --enable-gui=fb --enable-extensions=fileio. Currently some manual patches are necessary to help building Gnash (current trunk). Bug reports are files for this. All fixed in trunk as of revno # 11502. First, the following define must be removed from gnashconfig.h because the GST version used in the SDK is 0.10.14 and Gnash detects 0.10.15: //#define HAVE_GSTREAMER_0_10_GST_PBUTILS_INSTALL_PLUGINS_H 1 Also, g_strcmp0() is not available and this line must be added to gnashconfig.h: #define g_strcmp0 strcmp I changed g_strcmp0 to strcmp where I found it. The other issue is pbutils headers are installed, but not the library support. This is now fixed in trunk. The other problem was the Gstreamer headers wanted libxml2 headers, which we don't look for. When building native, we use pkg-config to get the flags, which include libxml2. When cross configuring with no pkg-config, we have to find all the pieces of Gstreamer ourselves. So I had to add libxml to the sanity check of what is found. Could be that this function is not being used anymore, I didn't check this. Just a glib dependency missed cause of cross-configuring. Earlier the build filed for the utilities directory, so I modify that Makefile to completely skip building it. Also fixed. Without --enable-media=gst this produced a working Gnash build for the Nomad, altough this was some months ago. My newer build is at: http://www.gnashdev.org/tools/sdg/gnash-nomad.tar.bz2. This is trunk as of revno #11502. I don't have any Nomad hardware, so I can't test it. (add begging for cool hardware here... :-) ) - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] using gnash to play BBC iPlayer content
On 09/11/09 10:32, Matthew Spencer wrote: I currently have a pet project to get the BBC's iPlayer working with gnash. I am using WebKit to display the BigScreen interface, with the gnash plugin supplying the flash support. I have a few questions for the list on how best to progress with the work, seeing as I have a bit of time available to try and get this working. It'll be more than a little bit of time... The BBC iPlayer requires RTMP support. I have RTMP support in an experimental branch, but it's not going to be in this next release due to not bring quite ready for prime time... I'll gladly work with anyone who would like to take on this task from my branch and beat it into shape. Unfortunately the few people that have considered plugging into this effort realized there is a steep learning curve with both RTMP and Gnash. If you're willing to put weeks or longer into this, let me know, I'd love to hand this off to somebody while I continue hacking on Cygnal. (the server side of RTMP) - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] using gnash to play BBC iPlayer content
On 09/11/09 14:21, Matthew Spencer wrote: I do have a pressing need to get iPlayer working, so access to your branch would be advantageous. I understand that there will be a steep learning curve. I will be be upfront about the time I have available though, family The timing doesn't matter to me, an hour or so may be fine. If there are any design documents you can point me at, or pointers to your branch, that would be great.. What docs I've written are all on http://www.gnashdev.org. The OSFlash site is also somewhat useful, but I've discovered some of their info is completely bogus. I've also been trying to add many more comments into the code. Any additional notes of corrections on our wiki would be greatly appreciated. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] iPlayer using gnash
On 09/11/09 08:55, Matthew Spencer wrote: Firstly, I am getting a huge number of 'UNIMPLEMENTED' messages in the log for 'LocalConnection.send' of the form: 13979:1] 14:46:44: DEBUG: * The send function is called * 13979:1] 14:46:44: DEBUG: STOP! No memory allocated!! 13979:1] 14:46:44: UNIMPLEMENTED: LocalConnection.send unimplemented [string:emp], [string:traceIt], [string:{72} 14:46:44 [system_event]], [string:Confirm IplayerLiveStats sent] LocalConnection::send() was mostly implemented by one of the student interns that worked on Gnash this summer. The part that needed to be finished was having the notifications work so a movie being player can get an event when data is written. This looks to me to be a debug interface in the swf file that is not currently connected to anything, my feeling is that this is safe to ignore for the moment? Very likely, yes, as the real streaming doesn't have anything to do with LocalConnection. The second question. YouTube streaming seems to be working fine (although not through the XL interface at youtube.com/xl). Looking at the logs again, it seems that youtube is using http rather than rtmp streaming. I notice Yes, YouTube just uses HTTP, although at one time we noticed they were testing RTMP support for a pay-for-view service. RTMP is considered a form of DRM for streaming. (or was... :-) ) there is a lot of rtmp code in place in gnash, but an rtmp packet is never sent. Looking at the code in NetConnection_as.cpp, I also note that there is only an HTTPRemotingHandler created (line 846), if I want to get RTMP working, do I need to implement an RTMPRemotingHandler interface? I can send you my version of NetConnection_as.cpp, which has rewritten RTMPT support, and most of the client side of RTMP. Al the low level code is done, tested, and works. I just hadn't had time to glue it all into trunk. Remoting isn't really used for just RTMP based streaming like the BBC, that's more for video conferencing types of things. I recently checked in most of my latest RTMP improvements a few days ago, including an improved client side API, so NetConnection_as.cpp can now be simplified somewhat. At one time I had utilities/rtmpget working for the BBC. All you really need is the initial handshake and NetConnection packets, (look in libnet/rtmp_client.*), and sending the NetStream::play() INVOKE, also already supported by the client side API. All the RTMP code is in libamf and libnet, and works fine. Test cases are in testsuite/libamf.all, libnet.all, and network.all. The Network.all tests require a working RTMP server, Cygnal, Red5, or FMS all work fine. I was really hoping to get all this done by the release, but working with the student interns took more of my time than I had hoped it would. I may be going on vacation after the release, so for better response, hit me with questions now, :-) I'm working on the release, and hope to push it out next week. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] firefox warning to update flash player
On 09/14/09 16:44, John Riesen wrote: Firefox has started warning about out of date plugins. Using Firefox 3.5.3 + gnash-plugin-0.8.5-3 (on Fedora 11), it now warns that: Change the version number Gnash uses to tell Firefox it's the flash plugin. In your $HOME/.gnashrc file you can set this to match Adobe: set flashVersionString LNX 10,0,12,36 This is also setable via the Edit-Preferences menu. My gut feeling is this will now force people to constantely update their version string for Gnash, which will become a constant hassle. It's pretty obvious that the Mozilla Foundation doesn't really care about anyone but their corporate funding partners. The funny part is they did this because of all the security problems with Adobe's plugin. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] firefox warning to update flash player
On 09/15/09 05:30, dolphinling wrote: I think the more likely scenario is that none of the people who developed the update feature use gnash, so it just didn't occur to them. Figures... There's already a bug filed: https://bugzilla.mozilla.org/show_bug.cgi?id=516166 Ah, thanks. We could probably help them figure out a way to tell the difference if interested. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] iPlayer using gnash
On 09/15/09 02:17, Matthew Spencer wrote: I have now integrated the code, but it seems to be acting as it was with the old version of NetConnection_as.cpp. Could I ask a few more questions: You'd need to uncomment the lines in asobj/flash/net/net.am to build it there, and not build the version in asobj, which is the old one. 1, what is the best way to see what is happening in the actionscript? My understanding is that the code will normally fail silently if there are unimplemented functions, is it possible for me to track these (bearing in mind that I have not written the iPlayer code and have no visibility of how they have implemented their rtmp streaming client side). I use the Gnash debug log for seeing what is happening, combined with either wireshark, tcpdump, or ngrep running in another window. I spend *huge* amounts of time analyzing the actual network traffic. The other thing I do alot of is write specialized testcases (network.all), and combined with the above is how I've done most of my work. For more info, track down the video of my RTMP reverse engineering talk from FOSDEM this year. There are also some good docs on the gnashdev.org wiki, complete with beautiful color coded decodings of RTMP based video streaming. 2, Which documents are you using for your actionscript reference? I notice that you have a property NetConnection.isConnected, but according to the flash 9 specs ( isConnected() changed somewhat between swf v8 v9/10. 3, What is the process for pushing patches upstream? Do I send them to the list? I would be happy to send the initial work I have done to get the code compiling somewhere for review. As noted, the patch tracker on savannah is good, email to this list works also. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] iPlayer using gnash
On 09/15/09 09:24, Benjamin Wolsey wrote: Almost all properties in AS2 are either implemented, or logged as unimplemented. If you find one that is silently unimplemented, it's a bug. I don't think there are any missing NetConnection functions. No missing methods or properties that I'm aware of, just missing functionality. The trunk version is pretty limited, and only works with OpenStreetMap. The version I sent you is much more complete. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] iPlayer using gnash
On 09/15/09 12:01, Matthew Spencer wrote: Quick status update, I am at the point where the initial handshake seems to be working now. I am not getting a vaild response to the 'connect' packet sent. Then the packet you sent isn't 100% correct. This usually happens if you miss the 128 byte boundary for RTMP chunks, in this case it should be a 0xc3 exactly 128 bytes after the 12 byte header. the NetConnection packet is usually over 128 bytes, so it gets split into two pieces. Try the RTMPClient::connectToServer() method, which should be doing this for you. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] iPlayer using gnash
On 09/17/09 04:44, Matthew Spencer wrote: (back from being offline all week) The point here is that the NetStream is only created after a successful connection to the server. My feeling is that this should be signalled after the Invoke::connect response is returned from the server. The current code only attempts to make a connection when NetConnection::call is called, which will never happen in the code sample above. I therefore plan to issue the handshake and Invoke::connect when NetConnection::connect is called The current file I sent you only supported remoting, not streaming. I've been working on streaming support, but only on the server side. Secondly, the current rtmp_client code has a hardcoded value for swfUrl for the version of RTMPClient::encodeConnect that is used when generating the Invoke::connect call. Through investigation, I can prove that the iPlayer servers are sensitive to this value being correct, so I plan to add a new version of the call that allows me to define a proper version for swfUrl. Yep, that was a hack, sorry. You can extract the URL you need from the Gnash VM. Any comments on what I have said? You might also find http://tlb.org/rtmpout.html useful, although it's for a live RTMP stream, and not a disk based one. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Flash HD (H.264) video decoding acceleration
On 09/22/09 08:38, Benjamin Wolsey wrote: The size of the patch almost certainly means that copyright assignment is necessary before it can be applied, and only Rob Savoye can sort that out. I assume this would be a corporate copyright assignment or a personal one ? I can send you the appropriate paperwork. On occasion I've had corporations refuse to assign copyright to the FSF (we're a GNU project), which has forced me to remove code I wish I didn't have to. I can think of a few other projects (OLPC's new hardware is Via based) that would benefit from this patch, so hopefully there is no problem. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] MIPS 400?
On 09/25/09 17:19, Rob Moore wrote: I have a Linux MIPS Alpha 400 type mini netbook thing. Can I use Gnash in order to view streaming video? It's not like a normal machine and can't recognise normal linux software that I download. There is a website about these kinds of machines but nothing on there about flash for streaming video. http://www.littlelinuxlaptop.com/ I have one of these Netbooks too, the CnMbook. I have succeeded in getting a root shell on mine, but haven't had the time to pursue getting Gnash running on it. It won't be an easy port. The crappy flash player they ship only supports swf v5. Personally, the machine is so slow that doing 100% software rendering for YouTube is likely to be painfully slow. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Ubuntu compilation
On 09/29/09 16:51, Andrea Palmatè wrote: after latest changes gnash cannot be compiled on Ubuntu (native not cross-compiled) i receive this error (in italian..) make[2]: *** Nessuna regola per creare l'obiettivo ../libbase/libgnashbase.la, necessario a libgnashamf.la. I can unfortunately finally reproduce this, and it's completely perverted what's going on. I think it's an Automake bug in how it handles dependencies. I manage to fix this by rearranging the Makefile some, without any other changes. Weird... I just checked in my fix (revno 11528), hopefully it'll work for you too. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] [Enhancement] Move start in paused mode to a plugin (to enable white-listing and easier configuration)
Bram Neijt wrote: Currently, Flashblock does not properly work with Gnash. On the other hand, Gnash does not support white-listing, blacklisting and removal of flash parts in a page. Um, Gnash does support whitelists and blacklists, and has for many years. See http://www.gnu.org/software/gnash/manual/gnashuser.html#gnashrc. I think it might help Gnash if the pause option was removed and instead a development team/person was looked for to do browser plugins or Flashblock support instead. Gnash supports more browsers than Firefox, so this option is still required. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Mips port...
Ramesh Chandra wrote: Is this the usual behaviour of gnash in mips platform or is there any problem in my cross-compilation, is my target very slow for the desired flash file? any optimization is needed? 400Mhz is too slow to use software rendering for streaming video. I've done several MIPS Longsoon ports (recompiling mostly), and they all worked fine with software decoding, but had a higher clock rate. Simple stuff should work fine at 400Mhz though. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Mips port...
Ramesh Chandra wrote: My mips platform is little endian. What clock rate you are exactly mentioning? because we have a 700Mhz Mips cpu of the same series, yet to try gnash in it. Do you think I'm not home to double check, but recall my two MIPS/Loongson machines ran at either 700 or 800Mhz. At that speed I could do software rendering of YouTube adequately. Simpler swf files played just fine. Typically ffmpeg has better performance than gstreamer. Certain types of animations in Gnash may perform better or worse depending on our implementation. We spend more time focusing on compatibility than performance unfortunately. Once we get better swf v9/10 support working, we'll be able to consider going after performance issues more than we have been. Gnash is seriously under resourced these days, and with what little funding we manage to get, we focus on the compatibility issue, as it's important. More funding would mean more time to work on performance issues, which is a time consuming task. Patches would also be welcome. And also, the same target has been ported with macromedia flash running windows Ce. The same .swf files runs fine up there with out any performance degradation.(not videos, only simple animated .swf files)..This proves that the target can perform well on the same .swf files with 400 Mhz too... Some thing needs to be optimized? Please correct me if i am wrong. I've got several 400Mhz platforms (Geode GX Geode LX based), and had no problems with simpler swf files either. I've even been able to do YouTube at around 12fps with software rendering, but again, some files played great, and some didn't, and I'm not sure what the differences were. Course in those cases I had optimized both GCC and GLIBC, which I guess was cheating. :-) As far as I can remember, a 400Mhz MIPS is slower than a 400Mhz Geode. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] is it possible to port Gnash to Symbian platform?
chunxi_c...@hotmail.com wrote: I want to is it possible to port Gnash to Symbian platform. how difficulty it will be? Possible ? yes. I've thought about it, as I have a Nokia E71x phone. As Symbian is pretty different, it'd potentially be a lot of work. At least S60 uses GCC. I haven't really investigated the details, although I believe the S60 SDK is freely available. I'd have to dig into the SDK to really make a better guess. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/12/09 21:45, Bernie Innocenti wrote: However, symbols were present for other binaries and for other libraries. Do you mean the other Gnash utilities had debug symbols ? The gnash executable is just a shell script, that launches the appropriate binary, which for the XO is gtk-gnash. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Porting Flash to Embedded Device running with Linux kernel 2.4 Open Source
On 10/12/09 06:42, Vinayaka UP wrote: Our Embedded product, is having Freescale Coldfire Series MCF5271 controller (Configured with 120 MHZ) and Linux kernel 2.4 Open Source.This device also have features like input via touch screen mouse, 128 MB of RAM and display with resolution 800x600. That should work fine. Let us know any harware specifications needs to be taken care for porting flash to embedded device ? Gnash runs on most any hardware configuration if you use GCC. You'll want to read the gnashdev.org for notes on cross compiling and other issues. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: YouTube using SWF10 today
On 10/13/09 13:51, Marcin Cieslak wrote: Dnia 08.10.2009 John Gilmoreg...@toad.com napisał/a: I went to this page today, and got a black gnash box: http://www.youtube.com/watch?v=G4S9tV8ZLcE As a workaround, this works: http://www.youtube.com/v/G4S9tV8ZLcE Interesting, sounds like a parsing problem of the parameters. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: YouTube using SWF10 today
On 10/13/09 13:51, Marcin Cieslak wrote: Dnia 08.10.2009 John Gilmoreg...@toad.com napisał/a: I went to this page today, and got a black gnash box: http://www.youtube.com/watch?v=G4S9tV8ZLcE As a workaround, this works: http://www.youtube.com/v/G4S9tV8ZLcE Now I see what the difference is. This other version brings up the same video in the swf v8 version of the media player. Thanks for the workaround, it should be useful till swf v10 works in Gnash. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/12/09 22:44, Bernie Innocenti wrote: Yeah, this is the binary that was missing symbols. But, does it work for you? What version of oprofile? On what distro? With what compile flags? On Fedora 11, I get debug symbols in gtk-gnash using -g -O2, which is the default. I haven't been using oprofile obviously, I'll get it setup on my Fedora 11 machine and see what's up. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/12/09 18:30, Michael Stone wrote: I just wanted to mention to you that Bernie Innocenti and I spent a few minutes together last week trying to learn something about why gnash stutters on assorted content on our Rawhide and Debian Sid machines. The stuttering (if you mean for video) is a known problem, there are some incomplete patches in the gnash-dev archives that work for some people. We approached the problem via oprofile and learned two things: 1. gnash-0.8.6-1 spends tons of time (on our systems and test cases) calling memset(). We're kind of curious as to why but... Called from where ? I don't have oprofile setup on my other Fedora 11 machine. Memset() is often used to set a newly allocated buffer to zeros to make debugging easier, but can also be a performance hog. 2. we can't get any data on what gnash is doing the rest of the time because we don't have access to appropriate debugging symbols. When I build Gnash from source, I get debug symbols. So I can only guess it's an issue with how the packages are built. Some experimental RPM packages I built, includingg the debug info ones are at: http://www.getgnash.org/packages/snapshots/fedora/. These I built using my own gnash.spec file (which is checked into Gnash trunk). This is the first time I've heard any issues about the lack of debug symbols in the debuginfo packages. Me, I don't install packages except for my own for testing, which I build myself. anything about it. (Try, e.g., gnash debug info, gnash oprofile, gnash profiling; the latter query is the only one with real results on the first Try on which machine ? An XO 1.5, standard Fedora 11, or a Debian machine ? I made a shot at compiling Gnash on an XO 1.5 machine, but get too many dependency errors when trying to install the *-devel packages Gnash needs. I assume I can build on my other Fedora 11 machine and just copy the executables over like I used to do ? Is the Gstreamer packages standard now, or are the - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/13/09 23:39, Kevin Kofler wrote: On Wednesday 14 October 2009, Rob Savoye wrote: Is the Gstreamer packages standard now, We build Gnash against GStreamer in Fedora, for the usual patent reasons. (GStreamer's modular architecture allows adding the patent-encumbered codecs from third-party repositories. FFmpeg's monolithic structure makes it a no go for Fedora.) Sure, but even Gstreamer still needs the ffmpeg plugin to work with YouTube. On the XO 1.0, the version of Gstreamer was custom, which caused me endless problems with Gnash. So what I was wondering is if on the XO 1.5 is it's using the standard Gstreamer in Fedora 11, or the older custom version. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/12/09 18:30, Michael Stone wrote: Lastly, if there's a simple solution out there or reason why our approach was doomed from the start, then I'd also appreciate it if you could link to it a little bit more prominently from your wiki since Google doesn't seem to know anything about it. (Try, e.g., gnash debug info, gnash oprofile, gnash profiling; the latter query is the only one with real results on the first page and they say nothing at all about oprofile). Ok, I got my XO 1.5 updated, built fresh rpms, and installed them. You can grab my snapshot from http://www.getgnash/packages/fedora. Built 11560 should work, it's what I installed. As far as I can tell, the debuginfo package I built has all the symbols. Now I'm trying to get oprofile running, but all the docs I find are out of date, and only refer to the older Geode based machines. So if somebody can give me pointers to getting oprofile up and running on the newer VIA based machine, I'd appreciate it. Then I can check the performance issues you mentioned, and add some info on our wiki. One thing I did find is all the animations I ran worked fine, even the bug nasty ones where scaling hurts performance. What didn't work well was was youtube. It seemed even worse than the older G1G1 machines for some reason. Embedded video tests worked good, so I assume the network overhead is what is killing performance for my quick tests. Overall, the XO 1.5 is sure much faster! - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/14/09 14:40, Bernie Innocenti wrote: It's amazingly simple these days: just install the oprofile-gui rpm and then run sudo oprof_start. Of course you also need to install the kernel-debuginfo and gnash-debuginfo. You probably also want glib-debuginfo and gtk-debuginfo. Can I do this my own builds of Gnash or do I need to install packages ? I guess I could install the rpms I just built, it's just more hassle. On my Fedora 11 machine, the oprofile_gui came right up, but on the XO I keep getting too many need to specify events errors. Memset() is often used to set a newly allocated buffer to zeros to make debugging easier, but can also be a performance hog. How do we turn off this behavior? By defining NDEBUG? Unfortunately not, but I should consider adding it in the critical places. Once I get oprofile going, I'll see about doing that. I think I got the debug symbols all right even from the Fedora rpm. The problem was just with oprofile... as if gnash did something that would break oprofile. Something like multi-threading, perhaps? Gnash is a multi-threaded application, does that cause oprofile problems ? I'm more used to gprof, unfortunately. I think they're standard. Most of the ugly customizations that OLPC did are finally gone. Whew. :-) I did get my builds working with the default Gstreamer on the XO. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Technical thing
On 10/14/09 04:40, John Snowden wrote: something so you end up with nothing- wouldn't it be better if online video was anything like png or jpg or giff? This would solve all the problems, I wouldn't care more. Talk to the content producers to use free codecs. Gnash only handles the formats that the producers of videos use, so you really need to talk to them. Maybe its time to shift your efforts to open video formats and promote it forvideo? Maybe anti-adobe/ google /mpegLA or free the The HTML5 video tag is only for web browsers, Gnash is a swf file player. We already support free codecs like Theora, Vorbis, etc... so what's your point ? :-) - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] You Tube troubles
On 10/15/09 18:41, dww wrote: Lately I have had trouble running clips from YouTube with the trunk I went back to 0.8.6 and 0.8.5 and had the same trouble ( blacked out viewing area). Has something changed on the YouTube site or is somthing bad with my build? I saw no build errors. Has anyone else experienced a Very recently YouTube started using a swf 10 based media player, which won't work in Gnash at this time. I wouldn't expect a fast fix. Many videos still work if they use the older media player, but none of the latest ones do at all. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/14/09 14:40, Bernie Innocenti wrote: In oprof_start, enable call graphs and set the path to your elf vmlinux image (/usr/lib/debug/lib/modules/2.6.31.1-56.fc12.x86_64/vmlinux in my case). Ok after a few frustrating days, I've gotten oprofile to work. Bizarrely, it crashed another system I was working on. Here's a log I made from a YouTube video with callgraphs and symbol info. Now anyone motivated should be able to get started on tracking down performance issues. http://www.gnashdev.org/testcases/oprof_gnash_all.log I'll add some notes on getting oprofile working, like make sure you have sufficient disk space! And don't forget to turn the daemon off, or you need even more disk space... Memset() is often used to set a newly allocated buffer to zeros to make debugging easier, but can also be a performance hog. How do we turn off this behavior? By defining NDEBUG? Interesting enough, none of the memset() was in any of the debugging code I was originally thinking of. :-) I will start adding NDEBUG though in the appropriate places. I think I got the debug symbols all right even from the Fedora rpm. The problem was just with oprofile... as if gnash did something that would break oprofile. Something like multi-threading, perhaps? I seemed to have no problem with multi-threading and oprofile. I've updated the gnash.spec file in Gnash trunk, it builds rpms just fine now, including a working debuginfo package. (which appeared to always be ok anyway). I've also been hacking on the Gnash Debian packaging files, and after one more test (adjusting dependencies) I'll check them in. I can now building the gnash-dbg package. The rpm and deb snapshots I've uploaded to http://www.getgnash.org under fedora and karmic snapshots. (both 32bit) My oprofile runs were built from trunk using -g -O2, which worked just fine. I've still failed to get oprofile working on the XO 1.5, I get many errors about not specifying events, and other problems. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] How to speed up gnash performance
On 10/15/09 08:22, Weidong Li wrote: slower than flashlite. However, gnash has a much better user interface than flashlite. It is possible that I did not do everything right with gnash so that it appeared to be slower. Were your tests with swf animations or streaming video ? In general, does anyone have any good idea to make gnash do the best it can? As a start, I managed to beat oprofile into a streaming video performance test. You can see that log here: http://www.gnashdev.org/testcases/oprof_gnash_all.log. Oprofile seems to be a reasonably good tool for performance analysis, it's easier than using gprof. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] oprofile difficulties, easy access to debuginfo, and other thoughts
On 10/16/09 14:33, Bernie Innocenti wrote: motivated should be able to get started on tracking down performance issues. http://www.gnashdev.org/testcases/oprof_gnash_all.log Very interesting. This looks by far the hottest spot: gnash::PropertyList::setReachable() That adds it to the garbage collector. http://gnashdev.org/doc/html/classgnash_1_1PropertyList.html#d7bae96d62a404a9656db68cc7aa408e. I just updated the doxygen pages on Gnash from trunk, the uptodate version is at: http://gnashdev.org/doc/html/index.html I'm not familiar with Gnash's codebase, but if PropertyList is really what its name suggests, it is a typical performance bottleneck in many applications. All ActionScript classes have properties, so you assume correctly. Some property maps implementations I've seen try to manage typed values using RTTI or fancy things such as boost::variant. In my experience, this is super-inefficient compared to storing plain strings and doing simple marshaling and unmarshaling as needed. We're using Boost already, but maybe this time we should do a little experimenting and see what's the fastest. Here's another performance tip: most (sane) PropertyMaps tend to allocate copies of the keys and values that are being stored in it. typically with std::string. All these micro allocations can hurt performance a lot. I suggest pooling allocations with a custom allocator or even a specialized string class. Even better, allocating strings from Enabling jemalloc helps, as it's more tuned towards lots of the small allocations both C++ and ActionScript use heavily. I'm rebuilding the srpm because I need x86_64 for my system. Are there any notable differences between this spec and the official one in rawhide? No idea, I haven't looked at the one in Rawhide, nor Fedora 11. I've used my own for ages as I like to configure things differently than a distribution would, often enabling experimental features like avm2, or rtmp. Tip: if you use Koji to build the Fedora packages (as opposed to building packages locally) you'd get builds for all architectures for free: I'm often having to build branches were everything isn't all checked in for testing, my source directory is mounted to all the machines in my office. So I don't think Koji would work as well that way. Plus I also do *BSD, and other distributions, so being able to build my own packages is useful, although sometimes the maintainance is a pain in the neck. I think Chris Ball and/or Daniel Drake may be able to tell us why this is happening. I assume the results would be mostly the same, but it'd be good to know if there are any differences. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Can it be done?
On 10/25/09 11:00, Scott Good wrote: Can Gnash be compiled to run on an ARM Based system such as the HTC dream and googles Android OS? Yes, Gnash has run on several ARM based systems including Android. More details at: http://www.gnashdev.org/?q=node/68 -rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] static cross compiling gnash for arm
On 10/29/09 06:47, wrote: i've successly cross compiled gnash-0.8.6 for arm,and i've got the sdl-gnash binary after the make option.but for some reason i want to get a static version of the executable binary.so i added --disable-shared --enable-static to the configure option.unfortunately,it failed with the mistake showing below: Fixed as of revno #11592 in trunk this morning. You found a bug! :-) - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Fw: Help - Adobe Flash Lite- Vx-Works - MPC880-Porting to Embedded Device.
On Mon, Nov 09, 2009 at 04:18:51PM +0530, Indrajeet Gajanan wrote: http://wiki.gnashdev.org/Building_dependant_packages#Dependant_Libraries. Let us know what are files are to be needed to port Gnash to an Embedded Device. You need a bunch of other libraries Gnash depends on. The list is at: http://wiki.gnashdev.org/Building_dependant_packages#Dependant_Libraries. If you are cross compiling, there are also tips on the wiki about how to get some of the libraries required to cross compile. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] static cross compile gnash
On 11/10/09 05:40, wrote: Dynamic section at offset 0xe36e1c contains 30 entries: TagType Name/Value 0x0001 (NEEDED) Shared library: [libpthread.so.0] 0x0001 (NEEDED) Shared library: [librt.so.1] 0x0001 (NEEDED) Shared library: [libdl.so.2] 0x0001 (NEEDED) Shared library: [libstdc++.so.6] 0x0001 (NEEDED) Shared library: [libm.so.6] 0x0001 (NEEDED) Shared library: [libc.so.6] 0x0001 (NEEDED) Shared library: [libgcc_s.so.1] 0x000c (INIT) 0x2175fc ... Now that I compiled in a static way,why does it still need these shared libraries?? These shared libraries are in the toolchain directories.And all these shared libraries have static ones correspondingly in the toolchain.Why doesn't the toolchain built these libraries in a static way? Configure with --enable-static to make Gnash built static libraries, but these other libraries are system ones. Statically building Gnash just makes the Gnash libs be compiled in. If any of these libraries is static archive, then they'll be compiled in too, but in this case they're dynamic ones. I think there is an option to libtool to only statically link everything, I can see if it works. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] How we can cofigure gnash for Freescale MCF5271 controller running on Linux kernel 2.4 Open Source.
On 11/19/09 02:06, Indrajeet Gajanan wrote: 1- I successfully configured the Gnash source code on Ubuntu and I am able to run the SWF file on my system. I want to know how we can configure the Gnash source code to Freescale Coldfire Series MCF5271 controller running on Linux kernel 2.4. What are options need to Enable/Disable while configuring Gnash so that i can run my SWF file on my Embedded device. Depending on whether your device runs X11 or not, just recompile Gnash for your target. You don't have to do anything special. 3- What additional packages need to be used to port the Gnash to an Embedded device. The same as what you used on Ubuntu. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] touch screen X Y axis inverted ?!?
On 11/19/09 05:34, moquette wrote: I compiled gnash-fb 0.8.5 from the openembedded.org cross compilation tool. The current release is 0.8.6, btw. It seems that the X and Y axis are inverted even for the mouse resolution. I'd try a newer release of Gnash, or better yet trunk, to see if the problem still exists. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] RTMP status
strk wrote: Does anyone have an idea on the status of the RTMP work ? Is the shared branch still being maintained (I don't see it in sync with trunk for the libcore/asobj part at least) ? Ignore the shared RTMP branch, I've been working in a different local branch since most people don't care about RTMP. I could make that the shared RTMP branch if anyone actually wanted to use it. Does rtmpget succeed at getting streams from servers ? Not anymore. My plan is to fix it up before the next release. The client side API has been changing, so rtmpget has bit-rotted. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Problem building Gnash 0.8.6 - Cygnal
On 12/02/09 15:37, Kristian Erikson wrote: We're having some problems with Building Gnash 0.8.6 using our build script (SPEC file and rpmbuild) and I was wondering if anyone has come across this one before. This is fixed in a branch, but I doubt you want to build Cygnal, so dropping --enable-cygnal should work. There is also an rpm packaging file in packaging/redhat, which is the one I use. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] How we can cofigure gnash for Freescale MCF5271 controller running on Linux kernel 2.4 Open Source.
On 12/09/09 05:41, Indrajeet Gajanan wrote: My questions are 1- How we can run the Gnash player on our device. After cross-compiling, load the executable on your target. For a UI, just have standalone Gnash launched when you boot or login. 2- Where I have to load swf file and which API class is used to read the swf file. 3- How gnash gets command from swf(our UI)i.e.how gnash capture buttons in swf file. Knowing nothing about your target makes it difficult to guess. You can either specify the file on the command line, or send it via stdin. gnash FileName.swf to run FileName.swf file, but how to run the same FileName.swf file on target. There are a variety of ways to do this. Enable auto-login, and just have that user's config files (.login) start it up. Or maybe add it to /etc/rc.local. This has nothing to do with Gnash really. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] How we can cofigure gnash for Freescale MCF5271 controller running on Linux kernel 2.4 Open Source.
Indrajeet Gajanan wrote: m68k-elf-tools-20030314.sh tool chain.Which command i have to use to generate executable file for target. 2- I try to configure the gnash by using ./configure --build=m68k-elf --target=m68k-elf --prefix=/usr/local/bin --enable-renderer=AGG --enable-media=none --enable-gui=FB --disable-plugin command whther it is the right way to congfigure the gnash for target. -Indrajeet G As Gnash isn't a cross compiler itself, don't use --target=m69k-elf, use --host=m68k-elf instead. You can also (if you need it) use --top-level to point to your sysroot. (where all the headers and libraries are installed) - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Memory use
On 12/14/09 15:34, strk wrote: Can anyone think of other strategies to decide *when* to run the GC ? Back when I worked on LISP systems (25 years ago), the GC would run in a separate thread based on a timer, as opposed to any other action. Each object was tagged as to it's state, and the GC would just remove anything with the appropriate state. Some customers of mine back then would often play games with the time slice for the GC to get the optimal memory uses. (back then memory was still expensive) This might not help in this case, unless the objects are tagged as unused before the action is complete. I'm not exactly suggesting we do it this way, I haven't really thought through how it would work in Gnash, but you did ask for other ideas... :-) Just as a note, there is some other memory profiling built into Gnash that uses mallinfo() to get memory stats from the kernel. There are also some utility functions in gmemory.h like startCheckpoint(), endCheckpoint(), to get finer grained info of where memory was going. I also added support to jemalloc to support mallinfo() as well. Out of curiosity, did you run your tests with or without jemalloc ? It's better tweaked for all the small allocations of memory that an OOPS language uses. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] how to run standalone or make a wrapper?
On 12/15/09 08:50, Dmitry Shalnoff wrote: I'm designer/ActionScript developer and I'm trying to make interactive GUI/game for OpenMoko by using AS. gnash perfectly installs on the device (and even show more-less nice performance), but I stuck in a Just as a question, did you install Gnash from an ancient ipk package (the current release is 0.9.6), or built it yourself from source ? Anyway now I think that will be better to use pythin for that (becouse python preistalled on OM and not necessary to compile anything). But, in any case, I never have a deal with python too. And my experiments finished also unsuccessfully. Gnash also has support to function as a python or GTK widget (donated to us by Sugar Labs), so that's an option too. Course that just lets you render swf files in a window. I know, that gnash was created as a GUI platform for embeded system. So, maybe somebody just show me a little trick for running my swf in standalone windowed/windowless/fullscreen mode? some python wrapper or something like that? Try gnash --fullscreen with the standalone player. :-) - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Configure the browser plugin to not print messages to stdout/stderr?
On 12/17/09 04:15, Petter Reinholdtsen wrote: This is 0.8.4-2 in Debian/Lenny. That's two releases behind, 0.8.6 being the most current. You might want to grab one of my Debian snapshots from getgnash.org and see if it's still a problem. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] how to run standalone or make a wrapper?
On 12/17/09 09:22, Dmitry Shalnoff wrote: that was prepared package form here http://www.getgnash.org/packages/releases/openmoko/ and there is 0.8.1 version only 0.8.1 is seriously ancient. I thought OpenMoko was a dead project ? I would like to try build gnash on OM platform by myself on weekend, but I never did it before ... suppouse that won't be easy for newbie. Hard to say. This page: http://wiki.gnashdev.org/Building_on_OpenMoko is also out of date. Their may be a newer bbfile in OpenEmbedded, if so, that'll make it easy. When I was doing OpenMoko builds, I ignore the bbfile nonsense, and just did traditional cross-compiling, which Gnash has good support for. but I can't found any links or other examples. May be I just don't understand something trivial, but, how can I use it?. If I right understand, I need some extra GTK library for that, which I can't found. The Python/GTK widget is being used for Sugar, so I'd guess there are example in that. It's a new feature, included in 0.8.6 for the first time. that was first idea :) doesn't work (at least on 0.8.1 version) Right, more reasons to use a newer Gnash, - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: Configure the browser plugin to not print messages to stdout/stderr?
On 12/20/09 09:16, Petter Reinholdtsen wrote: packages working. Luckily there are packages for Debian Lenny. I tested them, and they seem to fail completely. Found this message in ~/.xsession-errors: The Debian lenny x86 packages are kindof out of date, I need to fix my Debian machine. I do have some recent snapshots for GNewSense, which is basically Debian Lenny. The ubuntu packages should work as well. /usr/bin/gtk-gnash: error while loading shared libraries: libgnashnet.so.0: cannot open shared object file: No such file or directory LDD show that several libraries are not found: % ldd /usr/bin/gtk-gnash|grep not libgnashnet.so.0 = not found libgnashsound-trunk-20090313.so = not found libgnashnet.so.0 = not found libgnashsound-trunk-20090313.so = not found Let me check, this sounds like a bug in my packaging control files, as builds from trunk work just fine. I don't use the stock Debian control files, I use my own so they stay more uptodate. When installing the packages, I ran into problem because these versions of klash and gnash can't be installed in parallell (which work with the Debian packages). This make it impossible to install the konqueror and firefox plugin at the same time. Klash is built for kde4, btw. When you say installed in parallel, you mean as an upgrade for an existing installed gnash package ? Anyway, there is hope of having 0.8.4 fixed in Lenny, so we might continue with that for now. Sticking with 0.8.4 is completely silly, as we've made huge changes since then. For example, with 0.8.4, most video sites won't work at all anymore. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: Configure the browser plugin to not print messages to stdout/stderr?
On 12/20/09 10:34, Petter Reinholdtsen wrote: I mean by trying to install mozilla-plugin-gnash and konqueror-plugin-gnash together. I'll check, I'm working on the packaging files for Debian now. Part of my problem was the machine I was building on had odd versions of things (as I upgrade every 6 months, install of a fresh install). So I'm cleaning out the old crufty stuff and tweaking files on my karmic machine. Maybe I can get my Debian build slave to boot. Well, given that we lack the man-power to maintain our own set of gnash packages (to ensure timely bug and security fixes for the duration of the Debian Edu/Lenny lifetime), it is the best option. At one point the Debian packages for Gnash were up to date, I don't know why they aren;'t now. Ubuntu stays uptodate with our current versions, so it can't be that hard. I also barely have the time to maintain packaging files, but do so because we fix so many bugs between releases, it's often the only way (other than compiling from source) for users to get version that work for them. I'd gladly work with whomever is or can be a package maintainer to get the current release of 0.8.6 into Lenny. The alternative is to drop Gnash completely from the default build and ask schools to install the Adobe Flash packages if they need Flash. I thought that's what you were already doing. Several of our test sites as listed on URL: http://wiki.debian.org/DebianEdu/FlashInDebianEdu do not work as they should with gnash, and until they do it is hard for schools to I tried a handful of those links at random with Gnash trunk, and they appear to mostly all work. Maybe I'm missing something. Basing what works or not on an ancient, seriously out of date release doesn't accomplish anything. use only free software. I really hope Gnash reach that point soon, preferably in time for Debian/Squeeze. :) Well, we're seriously under-resourced these days, down to a small handful of developers hacking part-time, and a mostly broken build farm. :-( Everbody wants Gnash to be completely compatible, but almost nobody seems to want to help us get there... - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: Configure the browser plugin to not print messages to stdout/stderr?
On 12/20/09 11:26, Petter Reinholdtsen wrote: I suspect you are confused by the difference with the stable release in debian (Lenny), and what is available in other versions of debian. Gnash 0.8.6 is available in Debian, in the unstable distribution and is also currently in the testing distribution which represent what hopefully will be included in the next stable release (Squeeze). Ah yes, stable to the point of unusability on the Desktop. :-( The version in Lenny will probably not be changed, while important bugs might be fixed using backported patches or home made bug fixes. Which is nearly impossible for Gnash, often bugs get fixed by heavy code refactoring... Not sure if that is possible. But we could do this in Debian Edu, which is based on Debian Lenny with a few updated packages to get very important fixes in. Are you interested in maintaining gnash in Debian Edu/Lenny for a few years (until Debian Lenny no longer have security support), fixing bugs and security issues? Aren't we already doing that ? I just checked in a fix (#11723), which primarily fixes all the problems with the deb packages. I cleaned up all the dependencies, till they all can be installed in parallel, mozilla, klash, and konqueror too. I built new x86-karmic packages, now building the other platforms. This is what we did for Debian Edu/Etch, as gnash was not really fit for general use. Currently we are working on the Debian Edu/Lenny release, and here we have Gnash 0.8.4 which seem to be working for several sites, and thus might be better than providing no Flash implementation in the default installation and only document how schools can install Adobe Flash. It seems better to use a newer release. stable in this case means not keeping up to date with the web. Ubuntu keeps up to date with Gnash, and it hasn't ever been a problem. And definetely more sites work now than did for 0.8.4. The problematic sites I have seen and tend to visit to check the curren status is URL: http://www.steinalder.no/flashsite.htm , where Gnash earlier failed to wrap the text, leading to only the first words of the history from the stone age to show up in the browser and As I can't read Norwegian, I can't tell which is the history link. But everything I clicked on seemed to work. There was a lot of work done to Text fields, text formatting, etc this summer, so maybe it's all fixed. URL: http://met.hia.no/ which only report an error instead of showing the weather status, and finally URL: http://www.vg.no/ which have an insane amount of flash ads which really should not bog the machine down as much as it does. Hard to tell, I use AdBlock. :-) The photo galleries and the other stuff I randomly tried all worked fine. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
[Gnash-dev] Re: How can I help? [Was Re: Configure the browser plugin to not print messages to stdout/stderr?]
On 12/20/09 12:33, Richard Wilbur wrote: I can only offer part-time help as well, but I am interested in helping. 1. What do you folks need the most help on? Keeping our kde4 support working as you do now is a big help. There are always other little things than never get done, like rescanning all the strings for the translations, documentation improvements, bug fixing, more testcases, etc... 2. What's wrong with the build farm? A mix of hardware problems, and system administration. The hardware is all really old PCs, and they're all showing their age. It'd probably be better to replace the hardware than it is to fix it. I use primarily real PCs instead of VM images, as in my experience, VM images can be unstable, especially as you work with other OS's than linux. But it's mostly sysadmin tasks. Keeping all the machines up to date is important, but most are already one release behind for the BSDs. The other problem got to be buildbot was overly sensitive to minor problems, so it got to be where most of the time, it thought the builds had failed. Fixing all the little subtle bugs found by distcheck, make check, etc... turns out to be a time consuming task. It's almost more efficient for me to do builds by hand, as it's pretty well automated in Gnash. I got tired of how just keeping it all working could suck up much of my day when I'd rather be coding. Lately I've been getting access to other's build machines, which helps alot. That's what I do for my Debian-mipsel builds, and now GNewSense-amd64. I'd love to get access to more machines, like native ARM. Using other people's machines helps, as they do the maintainance. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] C++ IDE
On 01/01/10 03:56, Eric Pascal wrote: I'm new to this project, could you tell me which C++ IDE you are using to develop Gnash ? Well, I use emacs, but I think most of the other developers use vim. :-) This summer some of the student interns used something I had never heard of called Geary, and of course Eclipse has C++ support. Gnash uses Makefiles, which often don't work well under an IDE. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
[Gnash-dev] code freeze coming...
Just so there is plenty of warning, I was hoping to start a code freeze for the next release sometime in mid Jan. That should give me a few weeks to do testing, and get 0.8.7 out before both the Fedora and Ubuntu code freezes, which are both in Mid Feb. I just migrated a big pile of RTMP changes from a branch to trunk, does anyone else have a big checkin to make anytime soon ? Sorry for the other big checkin, but now it's 2010. :-) Some pre-release info is already up at http://wiki.gnashdev.org/Release_0.8.7. One thing to discuss is whether avm2 should be enabled by default. I configure with --enable-avm2 all the time, and haven't seen any stability problems. Has anyone else's experiences been different to the point that we should keep avm2 off by default ? What's the status of the GC hacking as far as any configurable behaviors that should be enabled or disabled by default for the next release ? - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] RTMP
On 12/29/09 13:21, Vishnu Viswanath wrote: To implement RTMP should we start studying how the other protocols are implemented ?? I just migrated my RTMP branch to trunk, that should help some. This has the changes to the client API like I had mentioned on IRC. I'll try to make some updates to the doc on how it currently works. Basically the rtmp.* files are the base class, with all shared functionality between the client and server side. Then there is rtmp_client.*, which does all client side of course. RTMPClient::connectToServer() now does all the grunt work of establishing an RTMP connection. both rtmp.* and rtmp_client.* are for encoding/decoding RTMP packets. The rtmp_msg.* is the overall class that holds multiple RTMP packets with the transport packet header. You can look at the Peering code in cygnal.cpp, which uses this to connect to other RTMP servers. I just merged this feature in from my branch, it'll let Cygnal distribute the load over a cluster by itself, and will let me do some other useful things by acting as a bridge between networks. So now Cygnal functions as both an RTMP client and server. I fixed those bugs with building the testcases in testsuite/network.all that you had a problem with. I fixed the other RTMP test cases while I was at it, everything tests valgrind clean. So you'll want to update, this is all checked in as of revno #11741. My suggestion is to fix the currently broken rtmpget to use the newer client side API to establish the RTMP connection, start the stream playing, and then write each packet to disk. This will be much simpler than jumping deep into Gnash right away. We can then fix any bugs in any of the lower level code you find without having them obscured by Gnash's whole code base. There are docs on RTMP on our wiki, the osflash website, and various other places if you search around the net. To stay sane, you'll want to learn about tcpdump, ngrep, and wireshark. Warning, wireshark has some bugs in it's RTMP decoder, so it can fool you if you're not careful. I primarily use ngrep and tcpdump. Anyway, this is so you can watch the network traffic, which is the only way to see the raw packets. That way you can see any decoding/encoding bugs by comparing the hex with what should be there. For lots of hex decoding examples, look in the libnet.all/test_rtmp, as it uses binary captures from the other player as test case data for the RTMP decoding API. Briefly, you start by doing the handshake, whose final packet after going back and forth is an AMF encoded NetConnection() object. This is what establishes the connection. To start a stream you send a NetStream() object, followed by a 'remoting call' of the NetStream::play() method. Then the server sends the stream, with an RTMP header every 'chunksize'. The RTMP protocol allows for multiple channels on a single connection, so then those all have to be split apart into each channel. For streaming though, it's pretty easy and only a few channels are used. For example, the audio and video are on separate channels. 02 is always the system channel. The server can also issue a 'new chunksize' command, which it usually does in the beginning, to change the packet size from the default 128bytes, often to 4096 or 1448. Bigger chunks means less processing, so better performance. There are functions for all of this in the client side API. As you figure things out, I'd appreciate expanding the documentation in the manual and our wiki. I've obviously been focusing more on code than docs. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] NPAPI plugin with KDE4 gui
On 12/31/09 12:06, John Wimer wrote: I'm no expert with the auto* suite, so I may have overlooked some things. But if no one sees a problem with this, I'll go ahead and check it in next week. I checked it in for you, as I was doing testing of some other changes anyway. One thing I'm having problems with is, separate from this patch, is the kde4-gnash segfaults in a constructor someplace on Fedora, but works fine on Ubuntu. There must be a subtle difference between the two systems kde4 installations. On Ferdora, the kde3 gui works just fine. I think the sdl and fltk GUIs also support XEmbed, so should they be supported in this manner also? I don't think either SDL or FLTK support XEmbed, but maybe things have changed since I last checked. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] [PATCH] Fix build with libtool2
Gwenole Beauchesne wrote: Current trunk doesn't build on my Ubuntu 9.04 system with libtool 2.2.6a. Well, you don't strictly notice this problem with the original trunk but manual inspection of libbase/Makefile.am reveals that some variables like noinst_HEADERS are only available in the libltdl1 path. Is this expected? This also means you can't generate a correct make dist with libtool2. This is very strange, as I build Gnash with both libtool 1.5 or libtool 2.x on Ubuntu/Debian/gNewSense all the time. I'll have to bring up a 9.04 VM image and try it there, it works fine on 9.10. BTW, it's no longer possible to run gnash within the build-dir on trunk either. gtk-gnash complains that: error while loading shared libraries: libgnashbase.so.0: cannot open shared object file: No such file or directory Sounds like more libtool issues, I also do this hourly... I'm driving home today after my ice climbing festival, I'll dig into this in more detail tomorrow. -rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Dear Gnash-Dev, If you can't find in google, try JUSTDIAL.COM
Andrea Palmatè wrote: Damn... SPAM everywhere.. :D Tell me about it. This account is now blocked. Considering whomeever had to get on the list to post this spam, I'm always amazed at the lengths spammers go to... They seem to get our wiki every few weeks as well. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] [PATCH] Fix build with libtool2
On 01/13/10 03:59, Gwenole Beauchesne wrote: I believed I fixed the libtool problem as of revno #11749. I dropped the support for the installable libltdl, as we only build the convenience library (static) for both libtool 1.5 and 2.x. My Cygwin build is still going, but it got past that point now. I tested this with both libtool 1.5 and .libtool 2.x, so hopefully it'll now work out of the box for you too. I'll be checking in a few other Cygwin Mingw32 patches too. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] [PATCH] Fix build with libtool2
On 01/15/10 02:36, Andrea Palmatè wrote: libtool: link: cannot find the library `../libltdl/libltdlc.la' or unhandled argument `../libltdl/libltdlc.la' It is searching the lib always in libltdl directory I believe I just fixed this as of revno #11788. Can you try again ? At least for me, trunk builds with libtool 1.5 or 2.x. Cygwin and mingw32 built as well without this problem. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Testsuite libtool errors
On 01/22/10 07:00, Benjamin Wolsey wrote: I have some libtool problems now. I did a fresh build; Gnash itself is fine, but the testsuite (misc-ming.all, libcore.all, libamf.all etc) fails to compile with errors like: I need to add these two libraries to GNASH_LIBS in the teststuite now. Not sure why it worked for me. Hopefully now the other libtool problem is cleared up, - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Anyone having build troubles on 'make check' ?
Anyone else with 8.10 ? You might trying staying more up to date. :-) Since Hardy is an LTS release, I installed Hardy as a VM image, updated it, built current versions of the testing dependencies, reproduced, and fixed this bug. So as of revno #11815, it should work fine, at least it does for me. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] [PATCH] Fix build with libtool2
On 01/28/10 08:03, and...@amigasoft.net wrote: On *unix: ltmain.sh (GNU libtool) 2.2.6b On Cygwin: libtool (GNU libtool 1.3140 2009-12-30) 2.2.7a ltdlver=`${LIBTOOLIZE:-libtoolize} --version | head -1 | cut -d ' ' -f 4` ltdlmajor=`echo $ltdlver | cut -d '.' -f 1` are wrong since they return ltdlmajor as version 1 since ltdlver is 1.3110 forcing ltdltmajor=2 cure the problem.. :) Aha! This has been driving me nuts as I've been trying Cygwin builds too. Glad you tracked it down. I just checked in a proper fix for this bug. Now autogen.sh strips out anything between (.*), so the version number is in the same place on *unix or cygwin. Fix checked in as of revno #11851. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Gnash replacement to Adobe flash Lite
Indrajeet Gajanan wrote: I want to know whether Gnash player can be used for these devices. Whether someone has tried to port the Gnash player to any Embedded devices as replacement to Adobe Flash Lite. Gnash star6ed it's life running on embedded devices, and we've done ports to several devices like the Sharp Zaurus, OpenMoko, Android, etc... We want port the Gnash player to our embedded device can you just tell me- You seem to ask this monthly, are our answers getting b;locked somehow ? What are the system/hardware consideration required to port the gnash player to an embedded devices. It depends on what you want to do. I've run Gnash on as low as 200Mhz processors, but it could only handle simple animations. around 400Mhz you can start to handle video (if you have to do software rendering), but it won't be great. Add hardware acceleration support, then video works much better. Without trying to do YouTube, most animations, user interfaces, etc... work fine on anything past 400Mhz. There is lots of room for optimizing Gnash, we haven't had the time or funding to do it. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Nellymoser
On 02/03/10 05:43, Benjamin Wolsey wrote: Are there any objections? Or can anyone demonstrate that the code works? Tgc added that code, but I don't think it's even been used, so removing I doubt will cause any problems. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
[Gnash-dev] release 0.8.7 branch made
I tagged and made the release branch today. It's tagged 'release_0_8_7_start', but what you'd really want to do is just do a bzr checkout of the 'release_0_8_7' branch from savannah. I have a few release specific tweaks to make, does anyone else have something that is release specific ? The goal is to push the release out by Monday... At this point I think most of the code changes and bug fixing are done. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] release 0.8.7 branch made
On 02/10/10 14:25, Robinson Tryon wrote: If there are any other bazaar commands people ought to know, we can document them on that page. How to use the -r revision option as a way of getting a revision that isn't a release. Also maybe a quicky on bzr st and bzr diff to see any changes people may have made to the sources locally. Sometimes people do this, then forget, and have weird bug reports... Back to release hacking... - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] LocalConnection
On 02/11/10 12:59, Benjamin Wolsey wrote: This problem should be fixed too now with a signal handler (revno 11936). We might consider adding this to the release branch too, as it is better to make sure things are cleaned up. Is this fixing an old bug, or one you only noticed with your changes for LocalConnection ? - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] LocalConnection
On 02/11/10 14:31, Benjamin Wolsey wrote: So there's no doubt we should be doing this, but the flip side is that it creates more opportunities to hang the browser, so maybe it needs more than a couple of days' testing before going into a release. I agree it's a good idea to handle this correct, I just have to decide whether the release should wait for it or not. Just before your email arrived, I had just uploaded source tarballs, debs, and rpms, and was about to ask for pre-release testing. :-) They're all at http:///www.getgnash/packages/pre-release-testing/. Actually regardless of this issue, I'd love to see some testing of the release candidates. While rebuilding the packages if I applied this patch is easy, it's the time for properly testing it I wonder about. Ideally the release goes out Monday at the latest, it's pretty much ready unless somebody finds a critical bug. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] pre-release 0.8.7 libxml header files not found
Peter Vermaas wrote: I'm trying to compile the 0.8.7. pre-release from http://www.getgnash.org/packages/pre-release-testing/ for the arm on openembedded and I'm getting the following errors: Did you use the old bbfile, or the new one ? My last OE build worked fine. So it looks like the libxml include files cannot be found. Well, Gnash doesn't use libxml2 at all, we have our own XML parser. Usually the flags from glib or gst from pkg-config have libzml2. Course with OE, you're not using pkg-config, so you don't get those flags. Sigh, I will have to add the config test back in I guess, I only removed it a few days ago. Is it still possible to add this for 0.8.7? I can add it to the bb file for Open Embedded, Ill have to think about adding this back in for the release. If I add the other patch for the signal handler, then I guess I can fix this too. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] pre-release 0.8.7 libxml header files not found
Benjamin Wolsey wrote: I am rather ignorant of autoconf and packaging, but it seems completely crazy to me that we should ever add config tests for dependencies of dependencies. Is there really not a better way of doing it? Not really... It's mostly just the header files included by the other projects, ie.. gstreamer needs libxml2 headers. If the platform doesn't have pkg-config, then we have to find all the cflags ourselves. Ugly, but it makes Gnash more portable. More interesting was the config test was broken anyway. :-) I fixed it the proper way, it works on Ubuntu, I still need to test it on OE. At least this does add the configure flag for the headers that lets one specify the path manually. I don't think I would add the patch for the signal handler. The risks appear to outweigh the benefits: it works as it is now and technically there's no problem with it, even though it's not pretty. Ok, I'll leave it out of the release. Time to build new tarballs... - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] pre-release 0.8.7 libxml header files not found
Peter Vermaas wrote: So it looks like the libxml include files cannot be found. Version 0.8.6 had a configure option I used to set the path --with-libxml2-incl=. I'm missing this option in 0.8.7. Is it still possible to add this for 0.8.7? I made new rc2 release files, with this fixed yesterday. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
[Gnash-dev] Gnash 0.8.7 Released
The 0.8.7 release Gnash is now out. It's available from ftp.gnu.org, and getgnash.org, and soon in your distribution of choice. More information about how to get Gnash is at these links: http://wiki.gnashdev.org/Gnash#Obtaining_Source_Code http://wiki.gnashdev.org/Gnash#Obtaining_Pre-built_Packages Improvements since the 0.8.6 release are: * Automatic and spontaneous screenshots support in all GUIs * Significant memory savings in parsing large XML trees and in some function calls * Enhancements in video streaming * Non blocking load of bitmaps, movies, data * Refactoring to eliminate most static data and get closer to re-entrant VM * Cygnal now supports multiple network connections, handling multiple video streams * Cygnal now supports plugins for server side scripting in C/C++ * Improved packaging support for deb and rpm See the NEWS file for more improvements. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Install instructions in the wiki fail for Ubuntu Jaunty
On 02/16/10 21:41, John Gilmore wrote: which says the current release is 0.8.5. Oops! Oops, now fixed. W: GPG error: http://www.getgnash.org ubuntu Release: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY 50B9A110A0B6D3FE I haven't figured out how to make the new repository somehow put up it's GPG key. I'm new to the whole repository administration thing, but it seemed a good idea to make the most recent version of Gnash available to people so they don't have to build from source. It was a nice try to specify ubuntu rather than jaunty, but unfortunately our built packages seem to be specific to Ubuntu Karmic rather than generic to multiple Ubuntu releases. Right, because I only have Karmic build machines for Ubuntu. It would be nice to support at least one distro back. I did make sure Gnash compiled on Hardy, an LTS release. I sortof missed Jaunty. I'd need to build up either VM images, install old versions on old hardware, or get access on other build slaves. I gave up on buildbot, so these days I do it partially by hand and just login to the build slaves. Most of the work is built into Gnash's Makefiles, make deb does all the work. If anyone else wants to build debs, I can give them upload access to the repository's incoming directory. I'll see about adding support for multiple versions of multiple distributions in the repository. Help from anyone that's done this before would also be nice. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Haiku support
On 02/17/10 07:53, Adrian Panasiuk wrote: In context of http://bzr.savannah.gnu.org/r/gnash/haiku/, here are some more changes needed for Haiku support: No shared memory support at all ? I just checked this into the branch, with a few changes. Thanks for catching a bunch of obscure typos. The only thing I left out are the freetype2 changes, I just built freetype2 for Haiku so I can finish that up. I know Haiku isn't GPL'd, but Gnash is, so I hope you don't mind doing the usual GPL copyright assignment forms for all the new Haiku specific files this patch adds. I can send them to you in a private message. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Haiku support
On 02/17/10 07:53, Adrian Panasiuk wrote: In context of http://bzr.savannah.gnu.org/r/gnash/haiku/, here are some more changes needed for Haiku support: I got these merged into the branch slightly massaged, which then configured and compiled on a Haiku VM image. The only problem was when I ran gnash, it appears to be running, I get debug messages in the terminal, etc... but I never get a window opened with the rendered scene in it. I assume this is in the Haiku specific code, but the branch now builds without any patches. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] gnash cross compiler
On 02/18/10 07:39, ON JJ wrote: HI I am trying to cross compile gnash to arm liunx, but when i try to ./configure glib i get this message then,i try to ./configure gettext(0.17),configure is ok,but when i make gettext i get this message mbslen.c: In function ‘mbslen’: mbslen.c:37: error: storage size of ‘iter’ isn’t known mbslen.c:40: error: ‘mbstate_t’ undeclared (first use in this function) make[4]: *** [mbslen.o] Error 1 . how to solve this problem,i cann't find any information in gnash-dev Archives,please tell me how to solve this thanks Sounds like a gettext problem, not a Gnash one. Is your arm system ulibc based instead of glibc ? - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Re: Failing to build mozilla plugin for gnash-0.8.7
Rupert Swarbrick wrote: I suspect that it's a result of ~/.mozilla/plugins not being in your dynamic link library path. Try running with LD_LIBRARY_PATH=/home/user/.mozilla/plugins firefox Actually you don't have to do that for the plugin, but you do need to have /usr/local/lib/gnash in your LD_LIBRARY_PATH. make install-plugin From your source tree. It should stick the plugin in ~/.mozilla/plugins for you, I think. That said, you might need to install the rest of gnash to a more standardised location for this to work (for the same reason). The default is to install the plugin in ~/.mozilla/plugins, as that way you don't have to be root. If you want to change it, you can use --with-npapi-install=system (to go in /usr/lib/mozilla/plugins) if you do want to do a system wide install. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Wrong md5sums for bz2 versions of 0.8.7 ?
On Thu, Feb 18, 2010 at 04:27:51PM +, Ken Moffat wrote: Thanks for the new release, but the md5s for .bz2 and .bz2.sig seem to be wrong in the announcement at http://www.gnu.org/software/gnash/release-0.8.7.txt I had to re-upload the bz2 tarball, because somebody noticed the original was corrupted. I did that, fixing the tarball, but couldn't change the release announcement, as it had already gone out. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Install instructions in the wiki fail for Ubuntu Jaunty
On 02/24/10 14:36, Richard Wilbur wrote: How is your repository hosted? Do you put files into it using FTP? Launchpad automatically creates a key for a Personal Package Archive I'm not using launchpad. :-) I've been using a utility called reprepro that's been very useful. I'm also using dupload to upload the signed binary packages. I have used pbuilder under Ubuntu to build binary packages from source packages for a particular Ubuntu release. It does a decent job as it creates a chroot build environment with the build dependencies from the target Ubuntu release repositories. The chroot build environment takes a little hard disk space to keep around (not too much as they are minimal to start with and also stored compressed) but updates are much faster than creating a new one. I may try pbuilder in the future, as building in a chroot is a good idea. Right now I have package building working good, and just added support for building multiple configurations as packages (http://wiki.gnashdev.org/Packaging), which can all go in the repository. I wrote a cronjob that checks for files uploaded by dupload, and then it uses reprepro to add them to the repository. Right now the problem seems to be GPG key handling. I'm having the same problem with the yum repository, which I setup with createrepo. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
[Gnash-dev] Hardware acceleration support
I've been hacking on hardware video decoding support in an experimental branch ('hwaccel'), and just added support for Gnash to load the renderer at runtime. So that means one can change between Cairo,. OpenGL, and AGG when gnash is started. It's also configurable by a gnashrc setting. Currently I'm just putting all the backends in one big library, and also all the GUI glue code. Luckily there were zero name collisions on anything. The original idea was to make them dynamically loadable plugins, but for now the big render library works. (in the branch, that is). I'm also considering renaming backend to librender, and in the branch, also moving libvaapi under librender. If anyone decides to play with this, run 'gnash --render-mode (opengl|cairo|agg)'. The new option for HW video accelerators is '--hwaccel (none, vaapi, xv)'. So far I've tested this on an Nvidia (binary blob and libvdpau required), and Intel (965 required), and supposedly ATI works with libxvba. The last thing I need to fix before migrating this to trunk is currently if vaapi is enabled, but you don't have supported hardware, it doesn't handle this gracefully at all. (it segfaults, actually) it should also be possible to find the best defaults for a platform, for example testing for opengl/vaapi support, then dropping back to agg/vaapi, then agg/xv, then agg/none. Unfortunately Gnash's Xv support has a major scaling bug... So far I actually haven't seen much performance improvement in my testing, (mpeg4 n an FLV) but it looks like the bottleneck is in our renderers. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Benjamin Wolsey wrote: I'm also not sure it's currently really worth making them really dynamically loadable (you mean dlopening them, I guess?). The only advantages I can see is that they'd be switchable during runtime (which the design currently doesn't allow) and that new renderers could be added without recompiling. The current way saves having to make an ugly C API so they could be dlopen'd at the expense of making the Gnash footprint larger. It should still be possible to build with only one library like we usually do, +1 for librender. Not sure exactly what you mean by moving libvaapi. I would prefer that libvaapi code is kept together, and separate from other code, though whether it's under librender or at top level isn't so I was going to move libvaapi, directory and all, under librender. Just a reorg thing. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Benjamin Wolsey wrote: On the other hand, gui/Player.h in the hwaccel branch is generally better than in trunk because most tabs are replaced with 4 spaces. But it now has some new 8-space tabs, so the indentation is also messed up in places. My problem... On my laptop I wasn't converting all the tabs to spaces, just some unlike my main computer. I fixed my emacs config so this shouldn't be a problem now. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Benjamin Wolsey wrote: Am Montag, den 01.03.2010, 19:08 -0700 schrieb Rob Savoye: I also don't think it's a good idea to use -r / --render-mode to select a renderer. Why not add --renderer, even with a short -R option? We're running out of letters for options. :-) I can change it, I'm not hung up on using -r, it was just available and made sense at the time. I'm obviously still cleaning this up, which is why it's all still in a branch. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Gwenole Beauchesne wrote: Actually, this won't work for VAAPI at this time because you won't have the right VaapiGlobalContext initialized for OGL. I will look into his once the remaining bits are merged. I noticed too that OpenGL doesn't work anyway, I obviously broke something. All I get now is a blank grey box. The changes for switching renderers is separate from the VAAPI code, even though they're in the same branch. This works on GMA500 too. The next change in Gnash would require an unreleased driver, though. All I had was an old Intel i915 which doesn't work with this at all. The performance improvement is really effective for H.264 videos playback only. You probably should downgrade your system to an Atom. ;-) You have a point. :-) I have no Atom hardware to test on, so yeah, on a dual core it may make no difference. Course that means on a fast enough CPU, VAAPI doesn't buy anything, so we'd probably want to disable it. I guess I had just assumed that a hardware decoding would be faster than software decoding, even with a dual-core. The only lend end hardware I have is a Via C7, which I don't think VAAPI supports. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Gwenole Beauchesne wrote: I think you should also replace sources annotations like // indent-tabs-mode: t to // indent-tabs-mode: nil That was the problem, I started fixing them all. I cut pasted a standard block for emacs, and grabbed the wrong one. :-) It does 4 space indents correctly, but the next lines wind up with an 8 space tab. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Gwenole Beauchesne wrote: On Tue, 2 Mar 2010, Benjamin Wolsey wrote: It's better to keep libvaapi at the top-level because you would run into weird dependency problems otherwise. libmedia, libbase, librender all require libvaapi. Ah, I guess it should stay top level then, plus we should move the Vaapi file in libbase. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Gwenole Beauchesne wrote: I have a Phenom X3 (triple core) and it does help. For example, the 2012 trailer I pointed to you doesn't decode in real time with the SW decoder. So, this generally is a win for H.264 on all platforms. Which is what I'd think. Please point me to your samples so that I can check where the bottleneck is. I believe the clip was poorly encoded. http://www.gnashdev.org/cygnal/videos. I'm also setting up a test page with a bunch of videos of varying types, and to test against various media servers. I was using the JW Flowplayer, with the mp4 files loaded via URL. I noticed you have a different version of the jw flvplayer, which doesn't work the same in Gnash as the old version I'm using. :-( This is how YouTube doe it, I didn't try an MP4 in an FLV container. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Gwenole Beauchesne wrote: I wanted to mention that the latest Cairo prerelease has a working OpenGL backend. This means we could drop our OpenGL renderer and rely on Cairo for crisp OpenGL rendering instead. Course our Cairo renderer is quite slow for some reason, I assume it's in how Gnash uses Cairo more than Cairo itself. But one of the main ideas behind the Cairo backend was that it was going to support both framebuffers and OpenGL, thereby freeing us from maintaining multiple renders like we do now. Anyway, I just checked in your patch to this branch as of revno #11825. For some reason the git patches don't apply 100% correctly, so although I went through this by hand, I could have missed something. It does compile and run though. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
Gwenole Beauchesne wrote: It seems OGL doesn't work at all, no matter VA-API is built-in or not. Something is probably wrong with the DRI/GLX check. There is a new check in the latest version that checks to see if DRI or GLX extensions are enabled, but nothing is done with that result yet. The idea is that if somebody did have DRI and GLX enabled, we'd probably want to use it. ** (lt-gtk-gnash:14386): WARNING **: This systms lacks a hardware OpenGL driver! (lt-gtk-gnash:14386): GtkGLExt-CRITICAL **: gtk_widget_set_gl_capability: assertion `!GTK_WIDGET_REALIZED (widget)' failed [...] We use the GtkGL widget for OpenGL and GTK. It was easier to do it that way at the time many years ago. I'm not 100% sure it's still the best option, but not being a graphics hacker, it's hard to know. All I know is that I would like Gnash to be able to take as much advantage of hardware acceleration whenever possible, as on a netbook, it's about the only way to get decent performance. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev
Re: [Gnash-dev] Hardware acceleration support
On 03/02/10 12:05, John Gilmore wrote: Great! It's good to stir up some fun in gnash like this. I'm still afraid that gnash will soon be irrelevant to most people, until it can play the friggin' scripts that the most popular video sites surround their videos with. While I agree with you, it's also important to make what we have currently run well. Plus hacking on this is an order of magnitude less work than avm2... and sometimes it's nice to work on something with a turn around time of a few days. More importantly, it's donated code, and Gwenole has been responsive on making the changes to the patch that are needed. All I'm doing is making sure the patches are better integrated into Gnash so we can actually benefit from this work, instead of watching it bit-rot. The switchable renders is a long-desired feature, and unrelated to the hardware acceleration patch. It'll make packaging Gnash easier, and hopefully get more usage and bug fixing in the opengl and cairo renderers. librender is such a generic name; if it's gnash-specific, I'd call it libgnashrender or something like that. The library itself would would be libgnashrender.*, just the directory name would be librender. It's probably a better name than backend, which is what we currently use. - rob - ___ Gnash-dev mailing list Gnash-dev@gnu.org http://lists.gnu.org/mailman/listinfo/gnash-dev