Closed Bug 72427 Opened 24 years ago Closed 23 years ago

Harry Potter & Shockwave.com (and other) pages with flash are blank

Categories

(Core :: Layout, defect, P1)

defect

Tracking

()

RESOLVED WORKSFORME
mozilla0.9.1

People

(Reporter: yminsky, Assigned: peterlubczynski-bugs)

References

()

Details

(Keywords: shockwave, testcase, Whiteboard: [seeking reviews])

Attachments

(5 files)

The page in question (www.harrypotter.com) has a flash animation in it. For some reason, the page is displayed ridiculously wide, and not centered on the animation in question. As such you have to scroll VERY carefully to even find the damn thing... Hard to know if this is a mozilla bug or a bug in the page. But it displays fine netscape and IE
On 2001031604 Win2k, it works fine for me.
Forgot to mention the version. I'm running mozilla-pre0.8.1-0, the RPM version distributed by Chris Blizzard.
Worksforme, linux build 2001-03-14-08. Yaron, could you retest this with a newer build (maybe 0.8.1 once that comes out in a few days)? I recall a layout fix to deal with problems such as the one you describe that went in after 0.8
here a sample of another page doing probably same thing. Observerd on linux 20010319 http://www.webpagesthatsuck.com/sucker.html
I tested the harrypotter site and couldn't see the bug. Then went back to the URL i pasted above, and the bug was gone there too. So whatever triggers it, the bug seems to happen only the first time flash is loaded within a session, and subsequent loads displays page OK.
Interestingly, this doesn't come up on the "web pages that suck" page that was given as an instance of the bug. This bug seems harder to replicate than I would have expected. It has reappeared on multiple builds, including some fairly recent pre 0.8.1 builds. Quite odd. y
Here is the url that I have problems with: http://harrypotter.warnerbros.com/intro.html The flash animation, which is used in a table is not displayed but I hear the audio portion. The page is wide and the scrollbars are present.
Status: UNCONFIRMED → NEW
Ever confirmed: true
Changing OS/Platform to all. This reproduces on the Win 98 and Mac OS 9.
OS: Linux → All
Hardware: PC → All
Assignee: karnaze → peterl
Target Milestone: --- → mozilla0.9.1
Reassigning to peterl and moving to m0.9.1
Ok, I created a reduced sample of the problem. It contains a table with a object element with a child embed tag. The OBJECT and EMBED elements have both the same width and height specified. If the OBJECT (which contains the embed element) has different width and height values, the flash animation is displayed. See my second test case for this example.
The Object's width and height values are different than the EMBED element. In this case, WIDTH="501" HEIGHT="216" has been specified.
Hmmm..Peter, this may have been the problem we saw at the old "subaru" page.
Good catch petersen. could be the issue that sometimes we use one measurment from one tag over the other assuming they are the same. Should we be resolving from style everywhere?
*** Bug 72243 has been marked as a duplicate of this bug. ***
Another good testcase going the other direction. http://bugzilla.mozilla.org/showattachment.cgi?attach_id=27956
Frame dump: ObjectFrame(embed)(7)@00CB4FF0 [view=02710D68] {0,-3000,7500,3225} [state=00002014] sc=00E0AF88<> Content dump of testcase object@00E756F8 classid=clsid:D27CDB6E-AE6D-11cf-96B8-444553540000 codebase=http://active.macromedia.com/flash2/cabs/swflash.cab#version=4,0,0,0 width=500 height=215 refcount=4< Text@00E75900 refcount=3<\n> param@00E75CB0 name=movie value=http://cache.harrypotter.warnerbros.com/intro/swf/hp_train.swf refcount=2<> Text@027B0070 refcount=3<\n> param@027B00D8 name=quality value=high refcount=2<> param@027B0260 name=bgcolor value=#000000 refcount=2<> param@027B0498 name=BASE value=http://harrypotter.warnerbros.com/ refcount=2<> Text@027B0650 refcount=3<\n> embed@027B0790 src=http://cache.harrypotter.warnerbros.com/intro/swf/hp_train.swf quality=high bgcolor=#000000 width=500 height=215 type=application/x-shockwave-flash base=http Version=ShockwaveFlash refcount=3<>
Status: NEW → ASSIGNED
Priority: -- → P1
*** Bug 70325 has been marked as a duplicate of this bug. ***
Marc, This looks more up your alley in core layout. Note: console says we didn't set the max-element sixe correctly.
Assignee: peterl → attinasi
Status: ASSIGNED → NEW
*** Bug 73367 has been marked as a duplicate of this bug. ***
*** Bug 70880 has been marked as a duplicate of this bug. ***
*** Bug 73771 has been marked as a duplicate of this bug. ***
*** Bug 72988 has been marked as a duplicate of this bug. ***
*** Bug 75177 has been marked as a duplicate of this bug. ***
Setting mostfreq kwd since there a loads of dups. Setting P2 since it is not a stability issue (no crash)
Status: NEW → ASSIGNED
Keywords: mostfreq
Priority: P1 → P2
Karnaze/Arun: This may be a dup of bug 5522. This bug has good testcases. Nominating mozilla0.9.
Keywords: mozilla0.9
I don't think this can make it for 0.9... Of course, if somebody BIG tells me to do it then I will ;)
Does anyone think that Bug 76769 is the actual dupe of Bug 5522 ? Or is it an additional dupe? Marc, if peterl is right, and this is in fact a dupe of Bug 5522, then fixing it in time for 0.9 would go a LOOONG way towards making Shockwave Installers work. I'm pretty BIG: 5 ft. 10' :-)
The problem seems to appear when the height of the flash animation is scaled down. I've submitted attachment 32416 [details] - another offending page with the following embedding code: <embed src="http://www.2cpu.com/Images/zzqbanner2.swf" scale="exactfit" pluginspage="http://www.macromedia.com/shockwave/download/index.cgi?P1_Prod_Version=ShockwaveFlash" type="application/x-shockwave-flash" width="125" height="125"> </embed> Removing the width and height parameters from this code causes the page to render to the correct width, but the animation is about double size (original size i suspect.) The problem seems to occur only when height is exactly 125. As far as I can tell, width has no effect on the problem. This problems has been observed in builds pulled off CVS on the 25th and the 26th April. It does not occur with mozilla 0.8.
Karnaze: I confirm that I can reproduce this pretty easily in an optimized build but can't in a debug build.
*** Bug 78097 has been marked as a duplicate of this bug. ***
Marking nsbeta1.
Severity: normal → critical
Keywords: nsbeta1
Priority: P2 → P1
Blocks: 74980
Reassigning to myself.
Assignee: attinasi → karnaze
No longer blocks: 74980
Status: ASSIGNED → NEW
*** This bug has been marked as a duplicate of 5522 ***
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → DUPLICATE
I can't duplicate this bug anymore, and the duplicate was marked fixed. Maybe we should mark this one fixed and get it off the mostfreq list. One down, a whole bunch to go...
Whiteboard: [THIS BUG IS FIXED AS OF 20010502]
*** Bug 81812 has been marked as a duplicate of this bug. ***
Reporter of bug 81812 still see this bug on Win2000 with Mozilla 0.9 (20010505) Shockwave Flash 5.0 r41, Shockwave 8. The fix for bug 5522 made it into 0.9. This bug should not be verified dup/fixed without further testing.
R.K.Aa: Please duplicate this behavior with a recent daily build. I'm not seeing this in 2001052204 Win98.
Marking verified in the trunk and branch june 04th builds.
Status: RESOLVED → VERIFIED
Still happened on my build.
Status: VERIFIED → REOPENED
Resolution: DUPLICATE → ---
Whiteboard: [THIS BUG IS FIXED AS OF 20010502]
--->peterlubczynski I think it's another variable that's not initilized. Attaching patch and cc:ing layout god Marc. Chaning summary from: Page with shockwave flash displays ridiculously wide
Assignee: karnaze → peterlubczynski
Status: REOPENED → NEW
Summary: Page with shockwave flash displays ridiculously wide → Harry Potter (and other) pages with flash are blank
Whiteboard: [fix-in-hand]
This is a regression, I'm seeking reviews for my patch.
Status: NEW → ASSIGNED
Keywords: patch, regression
Whiteboard: [fix-in-hand] → [seeking reviews]
Patch looks fine, but I don't see how it is a regression?
Well, I'm a bit confused. It's not a regression, now I think. It was makred a dup of bug 5522. It was orriginally about http://harrypotter.com but got more attention when it blocked Shockwave.com. When bug 5522 got fixed, it spawned bug 86075 which now has a comment that it is not the same as this. I haven't tried Shockwave.com lately in a non-debug build, so the problem may still be there as well (but I think that's really bug 86075). In any case, initilizing variables seems like a very safe way to fix this bug. Can I get your [s/r]=?
Keywords: regression
Assuming that you have tried Shockwave.com with your DEBUG build and it is fine, then [s]r=attinasi (I think the initialization is a good idea, but I am paranoid that the value it is being initialized to is OK, and testing it in a DEBUG build is good enough for me).
Assuming that you have followed peter's one condition the patch does look very safe to me. sr=blizzard
*** Bug 87324 has been marked as a duplicate of this bug. ***
Yup, happening on Shockwav.com too. I need to test this patch. Arun, this is an RTM stopper.
Summary: Harry Potter (and other) pages with flash are blank → Harry Potter & Shockwave.com (and other) pages with flash are blank
Build 2001-06-22-04, Win2k sp2 WORKSFORME for all testcases and URLs. I really believe this is caused by bug 86075. Also, a VERY important note: I went to test shockwave.com and got a shockwave.com HTML message saying Netscape 6 (and flavors) not supported. You can't get around this; you CANNOT visit shockwave.com.
This is not fixed. http://www.digitoday.no/dtno.nsf/pub/te20010622095039_pcs_72599679 Seeing this with a current linux CVS build - around 200 pages wide. If the flash doesn't load at first attempt - force a reload. The most frequently loaded commercial is a Compaq flash banner with heading "golden offer". It will appear in upper right corner.
This may be fixed by bug 46029 or by writing to the webmaster(s) of shockwave.com
oops: regarding my comment 2001-06-25 13:58 I discovered that those pages lay out too wide regardless of whether there is flash on them or not. It's another bug.
WFM Linux 2001062421
Status: ASSIGNED → RESOLVED
Closed: 24 years ago23 years ago
Resolution: --- → WORKSFORME
Verified 2001062504 Win98
not for me..go to shockwave.com use this pref to override the shockwave block : paste this line in 'allns.js' under 'default/prefs' folder pref("general.useragent.override","Mozilla/4.75 [en] (Win95; U)"); restart the browser..and go to shockwave.com and click 'games' 'shows' ..'music' 'games' 'shows' ..something like that in that order..and see that the page appears messed up...if u agree pls repoen this bug..
Try this URL instead: http://www.shockwave.com/bin/shockwave/entry.jsp?skip=net6 Both Games and Shows show up fine in my debug build from yeterday. What build are you using?
it's today's branch 0628...try to alternate between those tabs(3-4 times)..and see if it works... (games->shows->music->show->games)
**wakes up and...first thing in the morning ** This is still happening !!!
*** Bug 88378 has been marked as a duplicate of this bug. ***
*** Bug 88252 has been marked as a duplicate of this bug. ***
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: