Flex with a virtual frame buffer

I have to run an automated version of my flex app on a
headless server so I'm using a virtual frame buffer (Xvfb) running
on Solaris in an oc4j container.
I can get the following command to 'apparently' launch the
app:
String cmd = "/bin/sh DISPLAY=:5 /opt/sfw/bin/firefox "+
url+" -width 1600 -height 1200";
I say apparently because it does not cause any errors. But
when I send messages to the flex app via BlazeDS for it to do it's
thing...they never get there. I know the program works because it
is fine in my localhost development environment, but how do I get
the app to launch with a virtual frame buffer or more generally, on
a headless server?

This is a wild guess based on a small amount of research, but
you might want to check the 'headless-server' compiler option - see
http://livedocs.adobe.com/flex/201/html/wwhelp/wwhimpl/common/html/wwhelp.htm?context=Live Docs_Book_Parts&file=performance_118_41.html
and
http://community.adobe.com/help/search.html?q=headless-server=true&loc=en-us&hl=en-us&lbl= flex_product_adobelr&go=Search&self=1,
also
http://osflash.org/pipermail/osflash_osflash.org/2008-August/015858.html
hth...

Similar Messages

Xvfb Virtual Frame Buffer in a local zone?

Ok, so this is a bit of a science project....

Was wondering if anyone had gotten this working?

in the GLOBAL: the following works...

/usr/openwin/bin/Xvfb screen 1024x768x24


Doing the same in a LOCAL gives:

# /usr/openwin/bin/Xvfb screen 1024x768x24 
There is no XDISPLAY information for display 0. 
Check the OWconfig file(s) and try again. 

Fatal server error: 
The server will now exit because of configuration problems. 
# able to restore text mode on vfb# 
# 

So lets get a little fancy to get rid of the "XDISPLAY" error... edit /usr/openwin/server/etc/OWconfig and add:

# Virtual X Display for virtuals
class="XDISPLAY" name="5" 
coreKeyboard="NKBD" corePointer="NMOUSE" 
dev1="vfb"; 

Ok, now lets try the Xfb command in the LOCAL again...

# 
# /usr/openwin/bin/Xvfb :5 screen 1024x768x24 

Fatal server error: 
The server can't access IO registers. 

#

trussing this event shows lots of stuff, but no smoking gun (from what I can tell...)


Anyone have any experience in this area?

many thanks,

-- MikeE

Oops... forgot to mention...

This seems to work just fine for SPARC, its on x86 (SUN v40z) that there seems to be some sort of issue?

thanks, 

-- MikeE
(truncated truss output for those interested...)
sigaction(SIGXCPU, 0x08047CE0, 0x00000000) = 0 
sigaction(SIGXFSZ, 0x08047CE0, 0x00000000) = 0 
sigaction(SIGTHAW, 0x08047CE0, 0x00000000) = 0 
sysi86(SI86V86, 0x00000004, 0xFFFFFFFF, 0x08415378) Err#1 EPERM [sys_config]
fstat64(2, 0x08046FE0) = 0 

Fatal server error: 
write(2, "\n F a t a l s e r v e".., 21) = 21 
The server can't access IO registers. 
write(2, " T h e s e r v e r c".., 38) = 38

Fail to download X virtual frame buffer (xvfb)

I went to the page http://otn.oracle.com/products/ias/ias_utilities.html#xvfb in hope to download the oracle prebuilt Xvfb for solaris, when I click on any one of the three xvfb download items, it keeps returning a error page "The page cannot be displayed"
This is for an urgent oracle applications project, any prompt help is appreicated!
-Tim

Oops... forgot to mention...

This seems to work just fine for SPARC, its on x86 (SUN v40z) that there seems to be some sort of issue?

thanks, 

-- MikeE
(truncated truss output for those interested...)
sigaction(SIGXCPU, 0x08047CE0, 0x00000000) = 0 
sigaction(SIGXFSZ, 0x08047CE0, 0x00000000) = 0 
sigaction(SIGTHAW, 0x08047CE0, 0x00000000) = 0 
sysi86(SI86V86, 0x00000004, 0xFFFFFFFF, 0x08415378) Err#1 EPERM [sys_config]
fstat64(2, 0x08046FE0) = 0 

Fatal server error: 
write(2, "\n F a t a l s e r v e".., 21) = 21 
The server can't access IO registers. 
write(2, " T h e s e r v e r c".., 38) = 38

Solaris 8 (x86) Install Problem | No frame buffer console

Hi,
I am trying to install Solaris 8 (10/00) on my AMD K7 system. After I have selected the installation media, it will say something like "No usable keyboard", "No frame buffer console" then the installation will stop. I'm booting from the DCA and ACPI is disabled.
The configuration of my system is as follow:
Thunderbird 1GHz on ASUS A7V
384MB RAM
Adaptec 29160 (Driver downloaded from Adaptec)
Seagate U160 Disks
Plextor UltraPlex CD
MGA 400 Display card
104 keyboard & Logitech mouse.
I have checked the HCL, and they are on the list. Please help. Thanks.
Best regards,
Jason

Hi,
I am now trying to install Solaris to my other machine... and hoping it will work. This is a K6-2 PC with 192MB RAM and 2 x 8GB harddisks.
The first disk is occupied by Trustix and Slackware (Both are Linux), and I have the entire second dirve for Solaris. Then when I boot from DCA, my machine will keep rebooting itself. IF I boot from CD, it will say cannot find drive for installing Solaris. I have used LBA/CHS/Normal mode for my harddisk - none of them work. Any idea?
Please help.
Thanks.
Jason

Possible to build PC based frame buffer using PCI 6111?

I'm considering building a variable scan rate, passive capture frame buffer for Scanning Electron Microscopes. There are no off the shelf frame grabbers that support the clock rates I need and so I'm considering building a line grabber with a DAQ style device and building a frame in PC memory. Has anyone done this. Does the 'setup and run' support of a a PCI-6111 support grabbing video lines in rapid succession and then DMA'ing them out before the next video line comes in? Pixel clock is about 38KHz and 640 samples per line.

Rick Walton,
I noticed that I failed to post the link to the product spec page for the PCI-6111 (it can be found here).
The PCI-6111 can handle sampling rates up to a maximum of 5MS/s (see NI PCI-6111 User Manual).
The sampling rates are indeed programmable at your desired range.
If you have any further questions please post back. You can also look at the NI PCI-6111 page for specific pricing and purchasing inquiries.
Craig H.
National Instruments

Elite 3D frame buffer boot error.

Hi there,
I have a SUN Ultra 10 with a Elite 3D AFB Graphics card, that I recently tried to install.
Without the Elite 3D card (using the inbuilt PGX24 PCI card) the system boots fine and loads X Windows ok. after I install the Elite 3D card I notice a frame buffer error in the boot process and the system freezes just before loading X windows login.
Any ideas what is causing this? Do I need to install any drivers? I am using Solaris 8 and apparantly it should auto detect and configure the card.
Also please note that I use a Belkin KVM with a SUN Adapter to convert SUN I/O to a standard VGA and PS2 keyboard/Mouse. I have set the DIP switches on the Adapter to simulate 1280 x 1024 res. I don't think the adapter is the cause of the error (as I can see the display while booting etc), but thought I'd mentioned it anyways.
Thank you for your help. Please copy your answer to my [email protected]
Regards,
David

I have no idea about this framebuffer memory issue... but are you trying to use the mdadm hook or the mdamd_udev hook? If you are using the mdadm hook, you should really be using the mdadm_udev hook. The lvm2 hook requires that you have udev in your initramfs anyway, so you might as well use that one. It is apparently faster and more reliable anyway.

Frame buffer driver?

I am currently trying to port a program to Solaris, which uses the pixrect library. I know that the library is not supported under Solaris 2.x systems. However I downloaded a pixrect library patch that someone has wrote. After installing it and compiling it, everything checked out ok. But when I try to run the test files which came with the patch they wont run. Yes I tried to contact the creator 4 times!! After being ignored I figured I would post it and see if anyone out there has tried it or has any ideas. Here is the problem: within the program it declares a char* fb well then it says �fb=/dev/fb� After trying to execute the program it prints the error: �fb_type 22 is unknown� Almost as if it doesn�t know what fb is. I know that the max min on /dev/fb is 22 0. Also it is symbolically linked. So why this error? I�m running it on an Ultra 5 within both Solaris 2.7 and 2.8. I was told that my frame buffer is a PGX24. I am a newbe to all this, (recent computer science grad) however I have talked with many experienced people and they didn�t have many answers either. Please if anyone has a better clue than I please email me at [email protected] If you need more to go on I will provide anything you need. Thank you!!!

Did you try without the -s?

Images - are there frame buffer copies? How to release those resources?

Hello,
I'm a bit confused about the memory usage of images (BufferedImage class) in a Swing application.
I have about 50 Buttons (my own class derived from JButton), which use some images to render themselves. The images are cached, of course - although there are 50 button instances, they all use the same set of image instances.
Theses images instances use not more than about 2 mb of heap space (uncompressed images). But when I look at the heap usage via a memory profiler, the memory usage of those buttons is much more than 2 MB.
Now, I'm not a swing expert knowing about all internal structures, so I can't imagine why the buttons instances use a lot more memory than the sum of those image instances (I mean a lot more, not only the memory overhead that comes with the button instances), so maybe you can help me. Is there anything like a frame or drawing buffer, where those images data is copied to and internally used for rendering? I could imagine, that it is much more faster to draw from the graphics card memory than from the normal heap memory, is this correct? Maybe this is the answer, why the memory usage is much higher than those 2 MB?
Additionally, is there anything I have to know about releasing image resources, is there something special? Or is the GC able to free those BufferedImages (and perhaps connected frame buffer copies?) as usual?
I really appreciate your help
Thanks a lot

The profiler specifically said there were x+ instances of BufferedImages and that they were using more then 2MB?

How to get frame-buffer working?

Hi all.
Sorry if this is something simple, but I'm at a loss. I'm try to get an older PC to run some video apps (Mplayer, VLC, Etc.) at a decent speed, with frame-buffer support.
I've booted into Arch using after appending the 'vga=n' command to the kernel. I then ran mplayer without any video options specified, but it ran way too slow in comparison with Xorg. I also specified the 'fbdev2' option, but I couldn't go fullscreen. Is there something I need to do, or is the frame-buffer just slower in general?
The card is a Matrox G200 (8 MB).
In the end, I guess these would be the applicable questions:
- Is the Arch kernel built with 'matroxfb' support by default, and if it is, how do I use it?
- Do I need the 'directfb' package for anything?
- Do I need to build the Matrox drivers myself?
- Do I need to build Mplayer, or VLC, myself with Matrox FB support?

There is special support for Matrox in mplayer: -vo mga and -vo xmga
Run 'mplayer -vo help', see if those mentioned above are on the list. If not, you'll need to compile mplayer yourself. Though I have no idea what exactly the requisites are to get those working. If you can't get those to work, you're limited to fbdev which is indeed much slower than Xv, because colorspace conversion and scalign need to be done in software. So get vo mga working.

Is DRM needed to configure a VDMA frame buffer?

i have TPG->VDMA->AXIS to video out with a TC configured with constants.
I've been working through the ZC702 TRD
http://www.wiki.xilinx.com/Zynq+Base+TRD+2015.2
My TPG/VDMA drivers seem to load ok by looking at the kernel log. And, I have a /dev/v4l-subdev0 that shows up.
I seem to be able to configure the TPG fine and I think I understand what is going on with the TPG driver.
But, for VDMA if I just have it set up as a frame buffer what's the easiest way to configure/enable it? The TRD code is really large/conplicated and trying to figure out what the general methodology is for accessing the VDMA driver from a user space program.

Configuration of VDMA is done in Xilinx Video pipeline driver (using linux dma engine APIs).
Take a look at drivers/media/platform/xilinx/xilinx-dma.c
In devicetree (software/petalinux/subsystems/linux/configs/device-tree/video-cap.dtsi a video node is instantiated -
Userspace application will use this video node to query/queue/dequeue buffers.
video_cap {
compatible = "xlnx,video";
dmas = <&vdma_1 1>;
dma-names = "port0";
<snip>
In above example video node is using vdma_1 1 instance which is VDMA S2MM channel - it takes TPG streaming input and write it to mem.
In V4L2 framework vdma programming is encapsulated and it is not recomended to configure VDMA directly from userspace.
One question: Is you end goal to capture frames from TPG and displaying those on display( using DRM) ?
For reference take a look at video_lib (rdf0286-zc702-zvik-base-trd-2015-2\software\xsdk\projects\video_lib\src\s2m_pipeline.c) source.
Few other relevant links:
https://lwn.net/Articles/447435
http://lists.freedesktop.org/archives/dri-devel/2012-March/019778.html
http://www.wiki.xilinx.com/Xilinx+V4L2+driver
https://archive.fosdem.org/2014/schedule/.../v4l2_frameworks.pdf
-Radhey

Thinkpad T400 and frame buffer 1440x900

Hello!
I can't get the frame buffer working at the native resolution 1440x900 on my Thinkpad T400.
Have added the vga option at the kernel line in Grub.
kernel /vmlinuz26 root=/dev/sda6 ro devfs=nomount vga=0x355
Have tested with 354, 355 and 356. The screen just goes black.
Here is some info I used.
http://en.wikipedia.org/wiki/VESA_BIOS_Extensions
Orjanp

If you have an nvidia card, try vga=0x365. That's what I use on my Dell laptop with a Quadro NVS 135M driving a 1440x900 screen.
Failing that, use vga=ask and it will give you a list of all the modes the card you have supports.

Flex with php air application

Hi ,
I am new in flex.
Actually i make a flex with php desktop application. There are two problems which i am not getting:
a.) I make three module which i load on tab navigator click . These mudules working fine . But when i make this project export relese build to make .exe file . These modules not added in installation file.
i.e after installation i have to copy these modules .swf file in program files where the application installed. So i want to know is there is any method so that module added in the .exe file so that they by default goes at the
time of installation of exe.
b.) I use php services in this project and when i insatlled the application to system its not working untill i add the the project service php file and zend frame work folder on user wamp . Is there any method that all these happen by default at the time of insatllation.

Hi,
AIR is best suited to be a client application.
P.S: Since you are using PHP, you may want to take a look at
amfphp so that you can use AMF to speed up data transfer between
your AIR app and the PHP server.

Please help me, I need to use Flex with CF

Hi everyone,
I am starting to retract my brain from an intense 2 days to
figure out how to use Flex with Coldfusion. I am using Coldfusion 8
Beta (this part is good, don't need help for CF itself). I am also
using Flex Builder 2 (with chart but this not the topic). I
installed FDS 2.5 before to install ColdFusion 8 but I think CF8
have is own FDS or something like that... anyway, it might help to
mention it!
Well, I tried HTTPService, RemoteObject, WebSer vice -
NOTHING WORKS!!! Argh! Did I mention I spent 2 days? ;-)))
Well, I start with the basic: a login form with an email (as
a username) and a password to be validated. I have a CFC to do the
validation and return a simple message (string): "OK' when it is
valid and a custom message when it is not valid; depending if it is
the password and/or the email which is not valid.
I always got an error. Since the error is different depending
of the method I use, I will explain the latest method I used in
this message and try to concentrate on that method specifically!
The method is WebService and here is the error:
faultCode:Server.Error.Request
faultString:'HTTP request error'
faultDetail: 'Error: [IOErrorEvent type="ioError"
bubbles:false cancelable=false eventPhase=2
text="Error #2032: Stream Error .
URL=http://localhost:8501/iDashboard/login.cfc"].
URL: http://localhost:8501/iDashboard/login.cfc'
Any help will be very useful!!!
Thank's

Cyber,
This may be what you are looking for. It worked well for me.
Flex is sweet, but with the number of people confused about setup
with CF they sure aren't making any friends.
Flex/ColdFusion
setup

Having trouble with creation of a cloud service with multiple virtual machines using java sdk.

I am creating a cloud service on azure with 2 virtual machines using java sdk API. Service created successfully.
My input endpoint details are as follows.
INPUT ENDPOINTS
Head : 191.238.144.47:2400
Head : 191.238.144.47:22
Node0 : 191.238.144.47:43211
For Head instance port 2400 is for HTTPS and port 22 is for SSH and for Node0 instance port 43211 is for SSH.
But I am having problem with doing a ssh on Head instance. Sometimes it works sometimes doesn't. Same problem with HTTPS also.
I have some application running over there but when i try to access it thru browser sometimes it works but most of the time doesn't. When I restart the instances from azure portal, its works after
that(not always but most of the time).
Now I am confused what is going on there. I am creating cloud service and virtual machines using java sdk and setting input endpoints also. After creation of all instances i restart every instance programmatically .
I am not sure whether restart is required or not. It must be something to do with input endpoints only but not able to get the right thing i guess. When i do the same thing thru azure portal(creation of cloud service with virtual machines and setting up input
end points) everything works fine but not achieving the result when implementing it by java sdk API. Please help me.

HI Nithin,
Thanks for your reply. I am setting the endpoints after creating my instances using update call. Here's the code snippet.
AzureService aServ = new AzureService(session);
if(aServ.checkNameAvailability(clusterName)) {
aServ.createHostedService(clusterName, "dbX cluster");
// Creating head instance
aServ.createHead(clusterName, imgName, headType, userName, pswd);
// Setting end points for head node
String name = "ssh";
int port = 22;
aServ.updateVMInputEndpoint(clusterName, "Head", name, port);
// Restarting head instance
aServ.restartVM(clusterName, "Head");
String roleName = "Node";
String tmpRoleName = "";
for(int i=0; i<noi; i++) {
port = 43210+(i+1);
tmpRoleName = roleName + i;
// Creating node instance
aServ.createVM(clusterName, tmpRoleName, imgName, nodeType, userName, pswd);
// Setting end points for node instance
aServ.updateVMInputEndpoint(clusterName, tmpRoleName, name, port);
// Restarting node instance
aServ.restartVM(clusterName, tmpRoleName);
// Method to update the input endpoint details
public void updateVMInputEndpoint(String clusterName, String vmName, String name, int port)
throws Exception {
VirtualMachineGetResponse resp = computeManagementClient.getVirtualMachinesOperations().
get(clusterName, clusterName, vmName);
VirtualMachineUpdateParameters updateParameters = new VirtualMachineUpdateParameters();
//get the configuration list
ArrayList<ConfigurationSet> configlist = resp.getConfigurationSets();
//get inputendpoint list and update it
ArrayList<InputEndpoint> endpointlist = configlist.get(0).getInputEndpoints();
InputEndpoint inputEndpoint = new InputEndpoint();
inputEndpoint.setEnableDirectServerReturn(false);
inputEndpoint.setPort(port);
inputEndpoint.setLocalPort(port);
inputEndpoint.setName(name);
inputEndpoint.setProtocol(InputEndpointTransportProtocol.TCP);
endpointlist.add(inputEndpoint);
// Open port for https on head node
if(vmName.equals("Head")) {
inputEndpoint = new InputEndpoint();
inputEndpoint.setEnableDirectServerReturn(false);
inputEndpoint.setPort(2400);
inputEndpoint.setLocalPort(2400);
inputEndpoint.setName("https");
inputEndpoint.setProtocol(InputEndpointTransportProtocol.TCP);
endpointlist.add(inputEndpoint);
updateParameters.setConfigurationSets(configlist);
//required for update
OSVirtualHardDisk osVirtualHardDisk = resp.getOSVirtualHardDisk();
updateParameters.setOSVirtualHardDisk(osVirtualHardDisk);
updateParameters.setRoleName(resp.getRoleName());
OperationResponse updtResp = computeManagementClient.getVirtualMachinesOperations().update(clusterName, clusterName, resp.getRoleName(), updateParameters);
And every time i am creating a new cloud service along with head and node instances. Region is "South Central US".
I am setting ProvisionGuestAgent field to true at instance creation time. Thank you.

Problem with rendering still frame capture in FCP7

I'm a hack, so let's keep that in mind as I ask this question....but I *did* try to look through the forum for a similar issue before I created this post.
My situation: I am pulling some HD video from a Canon Vixia into FCP and basically just piecing it together with some minor editing to match music beats, etc. for some local sports leagues. Generally I find no issues with FCP, although I've only been dabbling with it for a few months.
My problem: My latest creation here: http://anothermotherproductions.smugmug.com/Movies/2011-Lacrosse-sample-movie/17 693649_L5rr8X does something annoying. I pulled a still frame from the video clip to let it hang there for a few seconds. See the 1:01 mark. The still frame looks clear in FCP, but when exported (and I've tried a few different formats, like h.264, pro-res, and "sharing" it for iPod) the image is ugly - looks like a motion blur from an old TV or something.
I've checked that the last frame of the video clip is the frame BEFORE the still frame, so walking step by step through the sequence reveals a normal sequence of motion, with the final frame "hanging" there for a while because its the still frame.
Make sense?
Anyone have constructive input?
Thanks.

Thank you so much - that helped tremendously. Image looks great now, AND I learned something.

Flex with a virtual frame buffer

Similar Messages

Maybe you are looking for