SWFObject 2.0 dynamic publishing and Firefox on Mac OSX express install issue

[GoogleCodeExporter]

Report copied over + summarized from:
http://groups.google.com/group/swfobject/browse_thread/thread/4f91e8e87037060b

SWFObject 2.0 RC1
Firefox, Mac OSX - after clicking on the button to agree to update
the player, the browser just hangs

The problem is that after the user clicks OK (or whatever it is) to
agree to update their Flash player in Firefox the browser just hangs.
No message appears. IIRC "transferring data from
fpdownload.macromedia.com" (or something to that effect) appears in
the browser status bar at the bottom, but that's all. 

I wondered about pop-up blocking but that doesn't appear to be the
issue. 

swf window used is 320 x 240 plus an few extra pixels

================

Report copied over + summarized from:
http://groups.google.com/group/swfobject/browse_thread/thread/4f91e8e87037060b

I too am having this problem, but with Firefox on OS X. It works fine
with Safari on OS X and Firefox with Linux and Windows, but has the
mentioned problem with Firefox on OS X.

================ 

Report sent via email:

I chose to go with SWFObject 2.0rc1 for our launch, and we have been mostly
successful.

There are some issues we are seeing that I would like to ask your opinion
on, however. I am not clear if it is something we are doing in our
publishing code that could be affecting this, but there are several people
who claim our site isnt working for them in firefox. We have a large swf
that we embed into the page for the main functionality. We are using
dynamic publishing, with the following code:

swfobject.flashvars = { permalink:'dave', vol:'0.3' };
swfobject.params = {};
swfobject.attributes = { id:'channel_swf', name:'channel_swf' };
swfobject.embedSWF('/swf/v1/channel.swf', 'bd', '974', '532', '9.0.47',
'/swf/expressInstall.swf', swfobject.flashvars, swfobject.params,
swfobject.attributes);

Do you see any problems with that?

We are trying to set focus for the object for disabled users as well with
the following code:

swfobject.addLoadEvent( function() {
    var channelSwf = YD.get('channel_swf'); // YAHOO.util.Dom.get() returns
an instance of the HTMLElement
    channelSwf.focus();
});

The main issues we notice are users unable to upgrade their Flash Players
on Firefox Mac (hanging connection to fpdownload.adobe.com) and/or white
boxes where the flash is supposed to be that hang indefinitely.

I have been unable to replicate these issues, however the bug reports are
very consistent.  Maybe you know something off the bat that would help.

Original issue reported on code.google.com by bobbyvandersluis on 19 Feb 2008 at 10:05

[GoogleCodeExporter]

What would be nice to know from everyone who encounters this issue is:
1. Mac OS X version
2. Intel or PPC processor
3. Available Flash Player version
4. Firefox version
5. Installed Firefox extensions
6. Any more information: exact status bar message or possible error messages

Original comment by bobbyvandersluis on 19 Feb 2008 at 10:30

[GoogleCodeExporter]

To avoid authoring errors when retesting:
- Express install is only supported by Flash Player 6.0.65 or higher on Win or 
Mac
platforms
- It that it requires a minimal SWF size of 310x137px
- Ensure that the path to expressInstall.swf is set correctly

Original comment by bobbyvandersluis on 19 Feb 2008 at 1:26

[GoogleCodeExporter]

I've tried to reproduce the issue, however had no luck so far.

I've been testing with:
- Mac OS X 10.4.11
- PowerPC
- Firefox 2.0.0.12
- Install of older fp version, update via express install to latest fp version,
uninstall with uninstaller:
* fp8r24
* fp9r16
* fp9r20
* fp9r28
* fp9r45
* fp9r47

With only positive results.

Furthermore, Firefox has a built-in pop-up blocker (I always have it on, no 
exception
for the Adobe domains), which doesn't seem to hinder the download process.

Now I wonder, did the people that encounter this issue really close all their 
browser
windows (including in their dock)?

Can anyone provide more info that can lead us to reproduce this issue? 

Original comment by bobbyvandersluis on 19 Feb 2008 at 1:37

[GoogleCodeExporter]

Why is this specific to Firefox? I have seen the white box behaviour after 
upgrading in Safari Mac and it seems to 
happen in (unknown) Windows browsers as well.

Original comment by jboeij...@gmail.com on 19 Feb 2008 at 2:57

[GoogleCodeExporter]

'I have seen...' and 'seems to...' isn't really helpful information. If you can
reproduce an issues under certain circumstances, please submit a test page or 
provide
enough details, so we can track the issue down.

Original comment by bobbyvandersluis on 19 Feb 2008 at 4:26

[GoogleCodeExporter]

Is everyone experiencing this issue using dynamic publishing? It would be 
interesting
to see if that is a factor. Also bobby, have you tried testing with the newer 
version
of Flash player? v115. Some of my users have reported issues with that. Also, it
still is trying to perform the upgrade even if they already have the latest 
version.

I would be interested in hearing any workarounds for the expressInstall crash. 
Maybe
manually detecting version?

Here are some user agent strings from people experiencing the issue on our site:

Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en; rv:1.8.1.12) Gecko/20080206 
Camino/1.5.5

Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-us) AppleWebKit/422 (KHTML, like 
Gecko)
Version/3.0.4 Safari/523.15

Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en) AppleWebKit/422 (KHTML, like 
Gecko)
Version/3.0.4 Safari/523.12.2

Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en-us) AppleWebKit/422 (KHTML, like 
Gecko)
Version/3.0.4 Safari/523.10.6

Mac Book Pro 2 Gig RAM
FIrefox 2
the latest flash player that works on you tube, seesmic etc etc.  
OSX 10.4 
Intel chip set

Mozilla/5.0 (Macintosh; U; Intel Mac OS X; en-US; rv:1.8.1.12) Gecko/20080201
Firefox/2.0.0.12

Original comment by dave.m...@gmail.com on 19 Feb 2008 at 7:13

[GoogleCodeExporter]

Analyzing your UA strings, we get the following browsers:
- Camino 1.5.5 on Mac OS X Intel
- Safari 3.0.4 on Mac OS X PPC
- Safari 3.0.4 on Mac OS X Intel
- Safari 3.0.4 on Mac OS X Intel
- Firefox 2.0 on Mac OS X 10.4 Intel
- Firefox 2.0.0.12 on Mac OS X Intel

So we can conclude that it is not Firefox on Mac OS X only, and that the only 
things
in common that they are all relatively new browsers on Mac OS X.

Here's my theory of what is happening:

As docmented in the SWFObject web authors FAQ [
http://code.google.com/p/swfobject/wiki/SWFObject_2_0_faq_web_authors ], Q6, in 
the
past years there have been 2 occasions where a Flash player didn't expose the 
right
version number externally (as can be retrieved with JavaScript) in Safari:
- 9.0.47 exposes 9.0.19
- 8.0.24 exposes 8.0.23

We know something was wrong with the Mac installer for fp9r47, and my suspicion 
is
that this doesn't affect Safari only, but all Mac OS X browsers. Because the
installer was fixed as soon as the bug was known, this issue has become
non-reproduceable using the archived fp installers at Adobe.com.

Looking at your test page you check for fp9r47 explicitly, which causes an 
infinite
express install loop. So the best way to avoid it is to check for any fp9 
release
version EXCEPT for the range fp9r20 - fp9r47 (so fp9r48+), and all issues 
should be over.

When you test for version 10 and you have fp9r115 installed, some strange 
express
install behavior is happening on Mac OS X too. Because there isn't a later 
version of
the Flash Player available, it is just guessing if this will result in possible 
bugs. 

I will forward this info to the team at Adobe.

Because it is not caused by SWFObject I will close this issue report. If you 
have any
additional info, please add it. It is always helpful to have as much detailed
information about issues as possible. Furthermore I will star this issue; if new
issues may come to light, we might reopen it again.

Original comment by bobbyvandersluis on 20 Feb 2008 at 10:37

• Changed state: Verified

[GoogleCodeExporter]

Thanks Bobby,
I will change our version check to 9.0.48 and let you know if we see any more 
crash
reports.
Best regards!
Dave

Original comment by dave.m...@gmail.com on 20 Feb 2008 at 8:47

[GoogleCodeExporter]

After some second thought, my theory about the version issue mismatch doesn't 
really
cut it, because this is only an issue when that fp version is the latest version
released, otherwise it will just trigger express install and install a later fp
version, and all should be fine.

I can reproduce the issue as described with fp9r115 in Firefox or Safari on Mac 
OS X.
Using either static or dynamic publishing and e.g. test for version "9.0.200", 
you
can see that the express install doesn't pop-up, instead you see a white box and
there is a hanging browser connection to fpdownload.macromedia.com.

Original comment by bobbyvandersluis on 20 Feb 2008 at 8:49

• Changed state: Accepted

[GoogleCodeExporter]

Yes, I see that issue when testing for 9.0.200 using fp9r115 on Mac using 
Safari and
FF. Can you think of any workarounds to prevent this from happening?

Original comment by dave.m...@gmail.com on 20 Feb 2008 at 9:37

[GoogleCodeExporter]

It appears that you can't require a version of the player that hasn't been 
released
yet, this would be a developer misuse of the feature (makes kind of sense).

So we are back to square one. Does anyone has more info that might help us to
reproduce the issue?

Original comment by bobbyvandersluis on 21 Feb 2008 at 9:39

[GoogleCodeExporter]

This seems to occur on Firefox 2 on OS X, Intel, 10.4 after updating from the 
version
before 9.0.115 to 9.0.115.

Original comment by Jason.Donenfeld on 21 Feb 2008 at 11:18

[GoogleCodeExporter]

Also: I can confirm comment 9:
"you can see that the express install doesn't pop-up, instead you see a white 
box and
there is a hanging browser connection to fpdownload.macromedia.com."

Original comment by Jason.Donenfeld on 21 Feb 2008 at 11:20

[GoogleCodeExporter]

Jason, just to verify:

- Did you detect the 'before' version with Flash using this test page [
http://www.swffix.org/testsuite/15_nested_iecc.html ] ?

- "9.0.115" is the version you test for in your SWFObject definition?

- If you retrieve the Flash version with SWFObject [
http://www.swffix.org/swfobject/testsuite/test_api_getflashplayerversion.html 
], what
does it say?

Original comment by bobbyvandersluis on 22 Feb 2008 at 10:16

[GoogleCodeExporter]

Just to add an FYI to this.  I was getting this same error.  Firefox 2, OS X 
10.4,
Intel Core Duo.  I had the movie set to update to 9.0.115 (need the mp4 
playback).  I
kept getting the white screen and the transferring from 
fpdownload.macromedia.com.  I
finally went to the link in comment 14 (bobbyvandersluis) and it told me my 
version
was 9.0.64.  That was strange, because I had just gone and downloaded the 
"latest"
from adobe a week or two ago.  So I went back to adobe's site and downloaded 
the real
install again, and now it all works.  I think there was a bad install on 
adobe's site.  

But a further question is why does the failing download look for
fpdownload.macromedia.com instead of fpdownload.adobe.com?  Is that the 
swfobject
script looking for that or the flash player itself?  

Original comment by jmata...@gmail.com on 23 Feb 2008 at 2:39

[GoogleCodeExporter]

Thanks, that's exactly the info we need to know, this confirms that there 
was/is a
bad install, that results in the triggering of an endless express install loop.

Regarding your question about fpdownload.macromedia.com, I don't think that for
express install this was ever updated to fpdownload.adobe.com.

This information will be forwarded to the Flash player team, and the issue 
report
will be closed.

Original comment by bobbyvandersluis on 23 Feb 2008 at 4:33

• Changed state: Invalid

[GoogleCodeExporter]

One potential work around in case Adobe errors in the future could be as 
follows:

Expressinstall.swf would be loaded under the ordinary conditions and methods. In
addition, the script would pass in the minimum flash version as an argument to 
the
expressinstall.swf script. Expressinstall would double check the minimum in the
beginning of the code against flash's version function. If expressinstall was 
indeed
loaded by accident (due to a disconnect between the javascript version 
reporting and
the flash version reporting), expressinstall.swf would call a javascript 
function in
the script that replaced expressinstall with the original movie, OR
expressinstall.swf would load the correct movie within itself (much like the 
way it
calls adobe's express install script external to itself).

Original comment by Jason.Donenfeld on 24 Feb 2008 at 8:33

[GoogleCodeExporter]

Jason, thanks for the good suggestion, we will keep it in mind as a possible 
future
enhancement.

Original comment by bobbyvandersluis on 28 Feb 2008 at 12:26

[GoogleCodeExporter]

I'm not sure that would solve the latest problem, which seems to be that 
sometimes
users are getting a lower version than the latest in the express install 
bundle, so
they should be getting .115, but they are getting .47 or .64 or something, so 
this
check wouldn't help in that case.

We may be able to detect an endless loop in the intstalls - by checking how many
times an expressinstall has been started with X hours/days using a 
SharedObject, and
then calling back to the js to show the alternate content, or add some sort of 
error
method to detect such a thing.

Original comment by TenSafeF...@gmail.com on 28 Feb 2008 at 5:30

[GoogleCodeExporter]

Any success?

Original comment by Jason.Donenfeld on 15 Mar 2008 at 9:18

[GoogleCodeExporter]

For now we are not going to anticipate on corrupt installers. We forwarded our
findings to Adobe and added Q2 and Q3 to our FAQ:
http://code.google.com/p/swfobject/wiki/SWFObject_2_0_faq_web_authors

Original comment by bobbyvandersluis on 24 Mar 2008 at 10:08

[GoogleCodeExporter]

i am still having this problem, almost a year later. has adobe not updated the 
corrupt installer yet?

Original comment by eatplast...@gmail.com on 20 Jan 2009 at 8:27