Cool VL Viewer forum

Sorry, should rtfm better.

After running the client with gdb I get the same behavior and the following in the terminal:


2011-08-01T17:04:38Z INFO: loadFile: Loading login history file at /home/mooreted/.secondlife/user_settings/saved_grids_login.xml
2011-08-01T17:04:38Z INFO: login_show: Initializing Login Screen
2011-08-01T17:04:38Z INFO: LLViewerMedia::getCurrentUserAgent: SecondLife/1.26.0.12 (Cool VL Viewer; default skin)
2011-08-01T17:04:38Z INFO: setStartupState: Startup state changing from STATE_LOGIN_SHOW to STATE_LOGIN_WAIT
2011-08-01T17:04:38Z INFO: LLPluginProcessParent::receiveMessage: plugin version string: Webkit media plugin, Webkit version 2.02.43315 (QtWebKit version 4.7.1)
2011-08-01T17:04:38Z INFO: LLPluginProcessParent::receiveMessage: message class: base -> version: 1.0
2011-08-01T17:04:38Z INFO: LLPluginProcessParent::receiveMessage: message class: media -> version: 1.0
2011-08-01T17:04:38Z INFO: LLPluginProcessParent::receiveMessage: message class: media_browser -> version: 1.0
2011-08-01T17:04:38Z INFO: handleMediaEvent: Media event: MEDIA_EVENT_NAVIGATE_BEGIN, url is data:text/html,%3Chtml%3E%3Cbody%20bgcolor=%22#ffffff%22%3E%3C/body%3E%3C/html%3E
2011-08-01T17:04:38Z INFO: handleMediaEvent: Media event: MEDIA_EVENT_NAVIGATE_COMPLETE, result string is:
2011-08-01T17:04:38Z INFO: handleMediaEvent: Media event: MEDIA_EVENT_NAVIGATE_BEGIN, url is data:text/html,%3Chtml%3E%3Cbody%20bgcolor=%22#ffffff%22%3E%3C/body%3E%3C/html%3E
2011-08-01T17:04:38Z WARNING: LLPluginProcessParent::pluginLockedUpOrQuit: child exited
2011-08-01T17:04:38Z INFO: LLPluginMessagePipe::pumpInput: Got EOF from plugin socket.
2011-08-01T17:04:38Z WARNING: LLFunctorRegistry<boost::function<void ()(LLSD const&, LLSD const&)> >::getFunctor: tried to find 'MediaPluginFailed' in LLFunctorRegistry, but it wasn't there.
2011-08-01T17:04:38Z WARNING: LLFunctorRegistry<boost::function<void ()(LLSD const&, LLSD const&)> >::log_functor: log_functor called with payload: {}
2011-08-01T17:04:38Z INFO: LLPluginMessagePipe::pumpInput: Got EOF from plugin socket.

This could indeed be a problem with plugins that webkit would try to launch and that would crash it... You installed the 32bits version of Flash, of course ?...

Do you mean that you are properly presented with the Cool VL Viewer TOS ?... Or just with the TOS window and a black page ?

Nothing new here: I was hoping for a report of some crash of the webkit plugin...

I'm short of ideas right now... Note that one of the reasons why I never use any 64bits OS on my 64bits computers is that you always run into such weird issues with them...

I usually don't use 64bit, but I now have 6gb of RAM.

Thanks for the help anyway. Everything but browser-based features work. I can always just search for stuff.

32bits kernels can perfectly use 6Gb of RAM...

Only if you use a PAE kernel which is slower than native 64bit since it's a translation rather than direct access to all the registers. In benchmarks PAE kernels often come off badly:

http://www.phoronix.com/scan.php?page=a ... _pae&num=2

It's easy to make benchmarks lie on numbers, or more exactly, to be mislead on the reason why numbers are what they are... If you compare a 32 bits distro such as Ubuntu, which is compiled to run on i386 (!) to another compiled to run only on x86_64, then yes, you will find a difference, not because of PAE (which only slightly affects memory accesses above 4Gb), but because the CPU new instructions, addressing modes and registers are used. It always baffled me to see how Debian-based distros still insist on compiling for i386 CPUs (who still got an i386 nowadays ?....). If you are really concerned about speed, then your best bet is to use another distro. I personally use Mandriva (compiled for i586), and the best distro to use for even more speed would be Gentoo (since it compiles entirely on your computer and you then can use the best compiler optimization options for your CPU).

I'm pretty sure 32bit Ubuntu 11.04 uses the 2.6.38-7-generic-i686 kernel. Not that I'm a big fan of Ubuntu, but I tried ArchLinux, Fedora, Debian and when I tried to set up xbmc with my remote control, after two weeks I could not find a way to do it. It took 17 minutes with Ubuntu. Ubuntu has it's issues for sure, but I find it easier to use than most. Mandrake was one of my first distros; I haven't tried Mandriva yet. Gentoo is too time consuming. I work 6 days a week and just don't have much time to mess with stuff.

Um, wow, Linux Torvald's view is bombastic, but he doesn't like PAE either. I'm far from an expert and you know way more than I do, but I just keep reading bad things about PAE:

http://blog.linuxolution.org/archives/117

The overall speed of a distribution is not so much about the kernel optimizations than about the optimizations that were used to compile the *programs* that make up the distro. Also, if the benchmark program you use was compiled for i386, it's dishonest to compare the results under a 32bits kernel with the results of the same program compiled for x86_64 on a 64bits kernel. So, yes, if you compare the speed of Ubuntu 32bits (compiled for i386) with Ubuntu 64bits, the latter will be much faster, but if you compare Mandriva 32bits with Mandriva 64bits, I doubt you will notice any speed difference (and in fact, some CPUs, such as Core2 ones, are faster for some operations in 32bits mode...).

Huh, interesting. I have an AMD Pheonom quad-core. It seems like rendering video and such is faster with 64bit, but maybe it's just because I doubled my RAM. When I get a less hectic day maybe I'll read deeper into the difference.

Thanks for the help. I'm sure I'll learn more about the subject.

Could you please try and download this file, untar it in the v1.26.0.13 installation directory (just like you install the v1.26.0.13 diff files), and tell me if it works for you ?...

This is a media plugin based on QtWebkit v4.6 which got a bug affecting the showcase search tab teleports, but that Linden Lab still insists on using for Linux (while their Windows and Darwin builds use v4.7.1), even in their newest v2.8 development viewer... There's perhaps a good reason for it !