Groups | Blog | Home
all groups > macromedia players flash > december 2006 >

macromedia players flash : Flash9b.ocx Crash



Erik Engineer
12/12/2006 4:46:53 PM
I am getting a crash in Flash9b.ocx (everytime the control loads) with Internet
Explorer 6 and Windows XP Profession SP2 (and I think I had the same issue with
Flash Player 8 too). It appears to be an "invalid opcode" exception on an
"emms" instruction (at 0x300511F5 when the module is loaded at 0x30000000; I
could tell you more with symbols and/or source). The machine is a Pentium Pro
200 MHz system with 256 MB of RAM. I know this system does not support MMX. I
am also aware this is the below the cited system requirements of a Pentium II
450 MHz or better. Does Flash Player 9 even have an MMX incapable code path?
Even if it does not, it should be noted I received no errors on install or at
run-time (unless you count the crash itself). I think it should probe for MMX
support and give a message at least instead of just crash (Note: a Pentium Pro
is a Family 6 processor without MMX support so you should probe for MMX and not
just processor class).

Besides the obvious bug, if Flash Player does not support my platform at all
(even unofficially) what are my options? I noticed Flash Player 7 did not have
any processor requirements (though I imagine it requires at least a 386 for
Win32 support). Of course I also noticed Flash Player 7 is no longer supported
at all too(only 9 and 8 seem to be currently supported).

Any help would be appreciated. Thank you.
Erik Engineer
12/26/2006 5:07:10 PM
Erik Engineer
1/21/2007 5:39:52 PM
spacejones NO[at]SPAM gmail.com
1/29/2007 11:38:55 AM
I have the same problem. I could not get Flash stable in either
Firefox 2.0.1 or Ie 6 or 7. I am an avid user of Firefox and it kept
crashing with a NPSWF32.dll error. Then after awhile I just wanted a
working browser. I attempted to get it working with IE 6 or 7. Now all
I get is a Flash9b.ocx error. I have been searching the net and usenet
for answers but I am not finding any that are working. I even looked
on Adobe's website and nothing. I feel your pain.

On Jan 21, 12:39 pm, "Erik Engineer" <webforumsu...@macromedia.com>
[quoted text, click to view]
montreal_2
1/30/2007 4:33:10 PM
Hello,

In another thread titled "Flash Player 9 crashes Internet Explorer 6"


[url]http://www.adobe.com/cfusion/webforums/forum/messageview.cfm?forumid=44&cat
id=184&threadid=1236801&enterthread=y[/url]

I described the exact same symptoms as yours.

I too arrived at the same conclusion as you.

I believe that since my Pentium II (200 mhz.) does not support MMX technology,
then that explains why Flash Player 8 and newer always crash the first time I
visit a webpage supporting Flash content.

I am wondering if the reason Flash 8 and 9 get installed without alerting the
user that their CPU does not support MMX is because there is no flag in the
Windows configuration file to indicate that this is the case.

If such a flag exists and is incorrectly set, then perhaps by changing it,
then Flash Player would load with an alternate set of software routines to
mimic the missing hardware instructions.

I am trying to find out more about how the installation of Windows goes about
detecting and advertising the presence or absence of MMX.
galahs.australia NO[at]SPAM gmail.com
1/30/2007 6:03:22 PM
I think Adobe should actually write: Requires MMX some where on the
system requirements page if this is the case. Even that small step
would save alot of heart ache.


I remember once about there being an MMX emulator. Wasn't good for
performance but might be useful if you really really need to use Flash
9 on your old computer.
Erik Engineer
1/31/2007 8:17:57 AM
montreal_2: Good of you to comment. I am glad I am not the only one with this
issue. It is also good to know I can install the older Flash 7 and have that
work for the time being (unless specific sites request higher versions).

I was under the assumption all the Pentium IIs supported MMX but perhaps I am
wrong. Either way, the flash player needs to probe the processor features
properly to use them (even if they do not want to support systems without MMX
they should probe and give a more appropriate message instead of just crashing).

I do not think this is a Windows flag that communicates this to the
application. An application that wants to use such processor features just
probes for this and uses it. I could be wrong (it would be nice if there was
such a fix though). I do know the OS has to probe and know this however as it
must know which registers to save across task switches. MMX reuses the
floating-point registers whereas SSE uses other dedicated registers I believe.
AddThis Social Bookmark Button