Closed Bug 53587 Opened 24 years ago Closed 24 years ago

Persistent search results keep reloading

Categories

(SeaMonkey :: Search, defect, P1)

x86
Windows NT

Tracking

(Not tracked)

VERIFIED DUPLICATE of bug 51211

People

(Reporter: michaell, Assigned: adamlock)

Details

(Whiteboard: [nsbeta3+])

Steps:
1)  Type in a series of terms into the location bar.  I used "greyhound rescue".
2)  Hit return.

Expected result:  Search page loads, Search Sidebar tab opens and displays list
of results.

Actual result: Search page loads, but the sidebar tab keeps loading and
reloading an additional 5 to 10 times.  Even if you open another tab, the search
tab will pop up again with a new reload.  Extremely confusing and annoying.

Using 2000092108 on NT.
Nominating for nsbeta 3.  This is a major problem with high profile functionality.
Keywords: nsbeta3
Marking "nsbeta3+".  I have to agree with LaGuardia on this.  Matt says he saw
this happening in yesterday's build.  He's bugging rjc for help ...
Priority: P3 → P1
Whiteboard: [nsbeta3+]
rcj thinks this is a tree widget problem.  I did some debugging
and saw that it looks during the reload of the frame to checkForDirectoryListing
in navigator.js and remembersearchText in search-panel.js .
I'm giving it to him to see if he knows more.
Assignee: matt → rjc
(Yep, apparently not a <tree> problem at all!)

In navigator.js around line # 425 you'll see that the content area gets various 
load listeners added to it, such as:
   contentArea.addEventListener("load", UpdateInternetSearchResults, true);

The problem is that these are firing several times per page load (which of course 
happens to screw up the search sidebar results panel;  it also slows down the 
browser in general because a lot of extra work is being done).

Who fires these, the docshell maybe?  Matt, can you find out and assign this bug 
to whoever works on it?
Assignee: rjc → matt
this is either a sherlock file problem
or a server problem
Either way rjc will know better
Assignee: matt → rjc
[This is neither a Sherlock file problem or a server problem... see above 
comments.  :) ]

Who fires these, the docshell maybe?  Matt, can you find out and assign this bug 
to whoever works on it?
Assignee: rjc → matt
Robert,

Right now Matt is way overloaded.  We're asking for your help because he hasn't
been able to figure this out.
Assignee: matt → rjc
Don, yes, I know... and I looked into it a bit. Now I'm asking for your help... 
*please* assign this to someone who works on the docshell so they can 
investigate why multiple "load" events are being fired.  :)
Assignee: rjc → don
Adam, can you help us out here?  (See previous comments.)
Assignee: don → adamlock
There has been a historical problem with onload being fired too early when 
progressive layout is enabled.

Wasn't it enabled recently?
Not sure.  Who would have made this change anyway?
Looks like a dup of bug 51211. Images in the search results page may be firing 
onload events which the search onload handler is being called to handle.

*** This bug has been marked as a duplicate of 51211 ***
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → DUPLICATE
Status: RESOLVED → VERIFIED
i agree. vrfy dup, and will be sure to recheck this once 51211 is fixed.
Product: Core → SeaMonkey
You need to log in before you can comment on or make changes to this bug.