78188 - jrgm.mcom.com/page-loader/loader.pl unable to load spinner.com

Status: RESOLVED WORKSFORME

(Core Graveyard :: Embedding: APIs, defect)

Description

[bmartin]

latest mfcembed 04-27-01

O/S: Win98
System: HP Vectra VL 266MHz 128MB RAM

Steps:
1. Download latest embed-win32.zip and extract it
2. launch mfcembed.exe
3. goto jrgm.mcom.com/page-loader/loader.pl
4. select submit

Results:

When script completes, www.spinner.com has NaN results for all five 
iterations of the page load.  This has been the result for the past month for 
this URL.

Going directly to 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/index.html
appears to load fine

Comment 1

[Judson Valeski]

-> chak

Comment 2

[Chak Nanga]

Running it via the .pl script or going directly to the URL (as mentioned above) 
loads the page correctly. However, the OnLoad() handler does not seem to get 
fired within the 30 second limit the script extects it in - hence the "NaN" in 
the final results table.

I also noticed the following after the spinner page was loaded/displayed:

1. The progress bar does not completely get cleared even after the page is 
loaded. Placing printf statements in the nsIWebProgressListener callbacks 
showed that either i'm not interpreting the values passed in correctly or i'm 
being sent bogus values. I also noticed several callbacks with the same param 
values in succession.

2. The busy cursor (Arrow+HourGlass) is displayed well after the page is 
displayed.
(The above behavior is just for the spinner page)

I think these are all related.

Cc:ing Rick and Adam for comments

Comment 3

[Judson Valeski]

*** Bug 78186 has been marked as a duplicate of this bug. ***

Comment 4

[Chak Nanga]

Targeting to 0.9.1

Comment 5

[Adam Lock]

This bug is a little bit confusing. If you type 
"http://www.spinner.com/index.html" directly into the address bar, you get 
redirected to a different page than if you go via the perf URL. 

Is it possible that the perl script feeding out the content isn't responding 
correctly to the redirect events?

Comment 6

[Chak Nanga]

I think the perf perl script has all that content "replicated" on jrgm.mcom.com

Here's something else i noticed:

0. Shutdown mfcembed if it's already running. Restart mfcembed
1. Go to http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/index.html and 
you should see the busy cursor.
2. Now go back to the same URL by selecting in the URL bar the page load will 
complete and you won't see the wait cursor forever like in step 1

I put some print statements in nsDocShell::OnStateChange() which look like:
(Lines with START are for the "STATE_START" case)

Set 1 (This is printed out for step 1 above)
-----

*** START : http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/index.html
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_7009755_123456_aol
.html?target=external&width=468&height=60
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000123_123456_ao
l.html?target=_top&width=124&height=191
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000124_123456_ao
l.html?target=_top&width=124&height=121
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000125_123456_ao
l.html?target=_top&width=124&height=73
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_7009755_123456_aol
.html?target=external&width=468&height=60
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000124_123456_ao
l.html?target=_top&width=124&height=121
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000123_123456_ao
l.html?target=_top&width=124&height=191
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000125_123456_ao
l.html?target=_top&width=124&height=73


Set 2 (This is printed out for step 1 above)
-----

*** START : http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/index.html
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_7009755_123456_aol
.html?target=external&width=468&height=60
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000123_123456_ao
l.html?target=_top&width=124&height=191
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000124_123456_ao
l.html?target=_top&width=124&height=121
*** START : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000125_123456_ao
l.html?target=_top&width=124&height=73
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000123_123456_ao
l.html?target=_top&width=124&height=191
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000124_123456_ao
l.html?target=_top&width=124&height=121
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_7009755_123456_aol
.html?target=external&width=468&height=60
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/html_92000125_123456_ao
l.html?target=_top&width=124&height=73
*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/index.html


Notice that in Set1 we never get an 

*** EndPageLoad : 
http://jrgm.mcom.com/perf/loadtime5/base/www.spinner.com/index.html

eventhough a corresponding START was called. While in Set2 we have matching 
START/EndPageLoads

Comment 7

[Chak Nanga]

Correction:

In my previous posting "Set 2 (This is printed out for step 1 above)" should 
actually read "Set 2 (This is printed out for step 2 above)"
                                                  -^-

Comment 8

[Chak Nanga]

->0.9.3 since i'm not sure i can get it done in time for 0.9.1

Comment 9

[John Morrison]

I'd like to note that while it would be good to have this working for testing
purposes, I think the larger point is not whether it can load that particular 
page (or twin failure 'www.zdnet.com_Gamespot.com'), but whether this is 
pointing to a specific defect in how the embedded app handles page loads.

Brent: does this load in another embedded application?

Comment 10

[Chak Nanga]

->0.9.4

Comment 11

[Michael Dunn]

It should be noted that this failure is reproducible in other embedded 
enviorments besides MFCEmbed.

Comment 12

[Asa Dotzler [:asa]]

Bugs targeted at mozilla1.0 without the mozilla1.0 keyword moved to mozilla1.0.1 
(you can query for this string to delete spam or retrieve the list of bugs I've 
moved)

Comment 13

[Michael Dunn]

It now works for me.