67133 - Alternate contents of an empty <OBJECT> (data=.html file) is not being rendered.

Status: RESOLVED DUPLICATE of bug 65455

(Core :: Layout: Form Controls, defect, P4)

Description

[illinar]

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; 0.7) Gecko/20010109
BuildID:    2001010901

Mozilla 0.6 worked absolutely correctly with this page, Internet Explorer (all
versions) works with it correctly as well, but 0.7 suddenly crashes trying to
render this page, probably that is caused by <form> elements of the page, since
page is displayed, except the <form> part and afterwards Mozilla dies.

Reproducible: Always
Steps to Reproduce:
just open the url

Actual Results:  crash with GPF

Expected Results:  render the page correctly as Mozilla 0.6 does

probably that is caused by <form> tag without </form>, also
tag <object>...</object> is there...

Comment 1

[Boris Zbarsky [:bzbarsky]]

no crash, linux build 2001-01-29-08.

Could you please try a more recent Mozilla build?

Comment 2

[Keyser Sose]

WORKSFORME
Platform: PC
OS: Windows 98
Mozilla Build: 2001012904

Comment 3

[Stephen Walker]

marking worksforme using the 2001020104 nightly on win2k.

Comment 4

[timeless]

vrfy wfm w2k 1/24-20

Comment 5

[illinar]

Mozilla build 20001020104 stops crashing but it doesn't render the page 
properly, in fact - the whole <object><form>...</object> section disappears - 
there is a large gap there

again Mozilla at W2K platform

Comment 6

[Brian 'netdragon' Bober]

If I am reading correctly, an <object> ... </object>, the ... should be 
rendered if <object> isn't loaded correctly. 
http://www.w3.org/TR/html4/struct/objects.html#h-13.1
"If the user agent is not able to render the object for whatever reason 
(configured not to, lack of resources, wrong architecture, etc.), it must try 
to render its contents." 

Therefore, I believe this bug is valid. As for the chance of there being a 
duplicate, I find it rather likely.

Comment 7

[Brian 'netdragon' Bober]

Although I think the Summary should be changed to "Form inclosed in an an empty 
object not being rendered". I think the browser is acting as if this object is 
valid even though it isn't being loaded.

Comment 8

[Keyser Sose]

Confirmed on the OBJECT/Form not being rendered.
Platform: PC
OS: Windows 98
Mozilla Build: 2001020320

Marking NEW.

Comment 9

[rods (gone)]

This seems to be an object frame issue, it does seem that the button in the form 
should be rendered and it doesn't appears any frames are being created. Over to 
core layout

Comment 10

[rods (gone)]

Attachment: simplified testcase (renders correctly in Nav 4.x)

Comment 11

[Marc Attinasi]

The testcase works fin in Nav6 RTM. Building with fresh source to check on
recent build...

Comment 12

[Marc Attinasi]

By setting processChildren to PR_TRUE when the ObjectFrame is created I can get 
all of the other frmaes to be created (form,input) but nothing is displayed, 
probably because all of the frame rects are 0,0,0,0. I'm not sure what is up... 
Here is the frame dump:

ObjectFrame(object)(1)@00DC1E30 next=00DC214C {0,0,0,0} [state=00000024] 
sc=0272B130<
  Text(0)@00DC1E78[0,1,T]  next=00DC1EB4 {0,0,0,0} [state=41000024] sc=02729F60<
    "\n"
  >
  Block(form)(1)@00DC1EB4 next=00DC2110 {0,0,0,0} [state=00000406] 
sc=02729B60(i=1,b=0)<
    line 00DC20E8: count=3 state=inline,dirty,NOT Impacted,[0x3001] {0,0,0,0} <
      Text(0)@00DC1F1C[0,0,F]  next=00DC1F58 {0,0,0,0} [state=00000426] 
sc=0272DE60<
        ""
      >
      ButtonControl(input)(1)@00DC1F58 next=00DC20AC {0,0,0,0} [state=00000426] 
sc=0272DDF0<
        Area(input)(1)@00DC1FFC {0,0,0,0} [state=00090406] sc=0272D230(i=1,b=0)<
          line 00DC2084: count=1 state=inline,dirty,NOT Impacted,[0x1001] 
{0,0,0,0} <
            Text(-1)@00DC2048[0,0,F]  {0,0,0,0} [state=00000406]sc=0272AB10<
              ""
            >
          >
        >
      >
      Text(2)@00DC20AC[0,0,F]  {0,0,0,0} [state=00000426] sc=0272DE60<
        ""
      >
    >
  >
  Text(2)@00DC2110[0,0,F]  {0,0,0,0} [state=00000426] sc=02729F60<
    ""
  >
>

Comment 13

[Marc Attinasi]

Hmm. The Object frame is never being reflowed, probably the reason for the empty 
rects. Must ahve something to do with the fact that it is marked as a replaced 
element but that it has no data coming in... maybe?

Comment 14

[Marc Attinasi]

This is looking more and more like an Object Frame issue. Note that you can get
the child frames to be created by setting processChildren to PR_TRUE when the
ObjectFrame is created in CSSFrameConstructor CreateByTag, but there are still
other problems with the way the Object frame handles itself in Reflow when it
has children (it thinks it is an image).

Over to the PluginMaster to look at, thought the priority sems pretty low for
this kind of markup...

Comment 15

[Vladimir Ermakov]

QA Contact Update

Comment 16

[Alexey Chernyak]

This actually has nothing to do with Forms.
Mozilla doesn't render contents of an empty <object>, which means it treats it
as a valid object?

Attaching a new testcase with 2 objects, one empty, one with missing data.

Old summary: Form inclosed in an an empty <OBJECT> not being rendered.
New summary: Alternate contents of an empty <OBJECT> is not being rendered.

CCing Ian Hixie and David Baron for their evaluation.

Comment 17

[Alexey Chernyak]

Attachment: testcase

Comment 18

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Marking duplicate of a bug that's a lot clearer since it starts off with that
description of the problem.

*** This bug has been marked as a duplicate of 65455 ***

Comment 19

[Alexey Chernyak]

verified dupe

Comment 20

[Daniel Wang]

*** Bug 168866 has been marked as a duplicate of this bug. ***

Comment 21

[Daniel Wang]

reopening re: bug 168866

Comment 22

[rubydoo123]

I am not sure why this was reopened, but am remarking as a dup of bug 65455

*** This bug has been marked as a duplicate of 65455 ***