179713 - flash makes Mozilla (windowsNT) crash

Status: RESOLVED DUPLICATE of bug 179822

(Core Graveyard :: Plug-ins, defect)

Description

[oliv]

User-Agent:       Mozilla/5.0 (Windows; U; WinNT4.0; en-US; rv:1.3a) Gecko/20021111
Build Identifier: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; rv:1.3a) Gecko/20021111

It seems recent builds have this regression: Flash plugin makes mozilla crash
(with drwatson poping up and mentionning an access violation).
Originally, I found this on this page:
http://cplus.fr/animation/video/Frame_video.asp?emission=guignols&inter=0
But searching on google for flash, I clicked found that
http://macromedia.com/software/flash/about makes it crash also, as is 
http://macromedia.com


Reproducible: Always

Steps to Reproduce:
1.go to macromedia.com

Actual Results:  
Mozilla crashes

Expected Results:  
well.... Mozilla should not crash :)

a copy/paste from the about plugins page:

Shockwave Flash
    File name: npswf32.dll
    Shockwave Flash 5.0 r30

MIME Type                     Description    Suffixes Enabled
application/x-shockwave-flash Flash          swf      Yes
application/futuresplash FutureSplash Player spl      Yes

Shockwave for Director

    File name: np32dsw.dll
    Macromedia Shockwave for Director Netscape plug-in, version 7.0.3

MIME Type                     Description      Suffixes       Enabled
application/x-director        Shockwave Movie  dir,dxr,dcr    Yes


System is : Windows NT 4sp5, PII 300 MHz, 384 MiB ram
Mozilla nightly is installed from the zip archive, not the installer (I still
use NS 4).
The last version I was using was from end of october, IIRC, and didn't have such
a crash.

Comment 1

[Matthias Versen [:Matti]]

please update to flash 6

Comment 2

[oliv]

I've tested old version of the nightlies:
20/10 works ok
30/10 works ok
8/11 works ok
11/11 crashes
So this is a recent regression.
I could not test 9/11 and 10/11, as I've found no Windows nightly builds of the
1.2 on the ftp server (the only 10/11 I've found is the 1.0 mozilla trunk).
If there are any, please give me a link.

To Matti:
Please, Matti, If you want to convince me to upgrade to flash 6, you will need
to give me a reason why I would do this.

First, if I did, this would imply I could not help Mozilla developpers to
correct this bug.

Second, there is a bug in Mozilla related to the flash plugin. That's a fact. 3
days ago, Mozilla did not crash. Now it does. period. The solution is not to
close the eyes and act like Mandrakesoft or Microsoft (i.e. "we don't have bugs,
it's impossible" for the first ones and "Please upgrade to the new version with
new bugs"). I've not been used to that by Mozilla people. I'd rather it not to
change.

If one wants me to change my computer settings, he will have to argue a little
more than 5 words (counting "6" as a word).

Comment 3

[Wesha]

oliv, Matti checks as much as 150 bugs a day. Multiplied by 150, even 6 words is
a little too much, I'm afraid.

And you never can say for sure what program of two interacting ones "have bugs"
if the tandem stops working. Maybe the new, legitimate operation by Mozilla just
hit some "sweet spot" of Flash 5. At least that's what happened with JRE 1.3.*,
and SUN confessed that it's their bug.

Comment 4

[oliv]

to wesha:

I agree, but this is not by going to flash 6 (which by the way seems to be
really buggy if I believe their support forums) that I will help in finding an
eventual bug of Mozilla. The odds are that Mozilla-nightly is at fault here.

Addon: NS 4.7 works fine with the very same plugins too... so A works with B for
ages, C (a completely different program) works with B, but A', a developement
snapshot does not work with B leads me to think B is not the problem.

to Matti:
I didn't mean to offend you. If you give me a "reason" (like "Mozilla 1.2 has
been decided not to work with Flash 5 because .... bla"), I'll make the jump.
Until then, I legitimatelly suspect a Mozilla bug, and want to help find it a
test fixes.

Comment 5

[Wesha]

Do you have Talkback installed? If no, please install Mozilla with Talkback,
reproduce the crash, click through Talkback to submit the incident, and post the
incident ID here. That would help a lot in determining where the crash actually
happens.

Comment 6

[oliv]

I've just submitted an incident with the 20021112 tallback version. But sorry, I
don't know what you mean by incident ID. Where can I have this ID? The tallback
never mentioned such thing :-/

Comment 7

[Matthias Versen [:Matti]]

There are known crashers with flash5 and recent version. AFAIK we don't suppoort
flash5 if it works in flash6.

Please run mozilla/components/talkback.exe to get the ID.
If this stack shows that the crash is inside the flash plugin and flash6 works
we will possible not fix this.

Comment 8

[rubydoo123]

Upgrading to Flash 6 is probably the right thing to do here. The sites you are
encountering (especially macromedis site) are probably using scripting. Flash 5
does not support scripting. When you go to a site, most sites sniff to see if a
particualr function is available, depending on the response, you will get served
a specific set of pages/functions. NS4.x does not support many of teh current
DOM functions for example, so you would be served a different set of
pages/funcitons.

Flash 6 has actually been very stable and Macromedia has been extremely
proactive in resolving the issues that have been reported. You may want to try
their beta version -- which I have been using for over a week and it works great:
http://www.macromedia.com/software/flashplayer/special/beta/

Since Flash 6 is the supported version, we will not focus on Flash 5 issues. I
will look at your talkback report and try and see what the issue is though.

Comment 9

[oliv]

TB13840327M

There should be also two others incident reports I did a few minutes ago, before
knowing how to get an ID (thanks google and Matti).

Comment 10

[rubydoo123]

Attachment: stacktrace

incident ID: 13839805

this does not look like a plug-in crash, rather it looks like it has something
to do with reflow and layout

Comment 11

[Matthias Versen [:Matti]]

maybe i can learn something :
0xffff0438
npswf32.dll + 0xb904 (0x1000b904)
npswf32.dll + 0xabd1 (0x1000abd1)
USER32.dll + 0x3067 (0x77e73067)
ntdll.dll + 0x163a3 (0x77f763a3)
nsView::SetDimensions [c:/builds/seamonkey/mozilla/view/src/nsView.cpp, line 530]

isn't this a clear crash in the plugin during a reflow ?

Comment 12

[rubydoo123]

yes, I was looking at two different stack traces, I got them mixed up, this is
clearly a flash crash

reporter: please try a current version of flash

Comment 13

[oliv]

No more crash with Flash version 6.0r47.

But unfortunately, my machine cannot be used to test the flash 5 crash, as I
don't have flash 5 anymore, I guess :-/

Anyway, thanks everyone for your reactivity (is this an english word?).

I'll let you decide if you want to close the bug or go on with it, but maybe
this issue should be mentioned in the release notes of the upcoming 1.2 version.

Comment 14

[rubydoo123]

I am marking this as Fixed in that the reporter installed Flash 6

Comment 15

[Wesha]

As per bugzilla guidelnis, only the bugs where some action was taken by mozilla
team to resolve, such as patch, bugfix, etc. are to be marked FIXED. This one is
INVALID, rather.
Reopening to mark so.

Comment 16

[Wesha]

Yeah, I know that I like accuracy maybe a little too much...

Comment 17

[rubydoo123]

whatever makes you happy

Comment 18

[(not reading, please use [e-mail] instead)]

*** Bug 179743 has been marked as a duplicate of this bug. ***

Comment 19

[(not reading, please use [e-mail] instead)]

here's how I got this:

1)  have old version of 4.x with flash
2)  install mozilla without flash
3)  go to page with mozilla that uses flash
4)  mozilla searches for the plugin, finds 4.x
5)  boom

is there anything we can do to avoid that?

it's going to be common, and users won't know they have to update.

(unless, do we ship Netscape 7.x with Flash 6.0?)

Comment 20

[Matthias Versen [:Matti]]

Seth: you should also read bug 179822 :-)

Comment 21

[(not reading, please use [e-mail] instead)]

let me re-open, and dup against #179822

Comment 22

[(not reading, please use [e-mail] instead)]

dup

*** This bug has been marked as a duplicate of 179822 ***