18495 - Large table loading performance is much slower than NS 4.x

Status: VERIFIED DUPLICATE of bug 5588

(Core :: Networking, defect, P3)

Description

[omega2]

On my system (NT 4.0, SP3), the M10 build (and previous builds) load the
URL above very slowly.  The lower left pane (marked "All Classes") takes
much longer to load than on NS 4.x versions.

For a better test, download the Java SDK 1.2 API documentation locally
and try loading the API documentation to eliminate dependency on network access
time.

Comment 1

[leger]

This should be a lot better now, the fix for 6048 should have helped.  Can you
try the latest M11 milestone build
at:http://www.mozilla.org/projects/seamonkey/release-notes/

Also, don't forget to check: http://www.mozilla.org/quality/most-frequent-bugs/
for known problems.

Also...many other tables problems already reported...feel free to add comments:
http://bugzilla.mozilla.org/show_bug.cgi?id=12598
http://bugzilla.mozilla.org/show_bug.cgi?id=18034

Try http://bugzilla.mozilla.org/query.cgi to look for dups.

*** This bug has been marked as a duplicate of 6048 ***

Comment 2

[omega2]

I've looked at the other bugs, and none of them seem to cover exactly the
slowness in this document.  The page is large but not "pathologically large",
as NS 4.x handles it quite well.  The interactive speed browsing the JDK 1.2
documentation remains very problematic.

The bug remains about the same in M11.  Actual load time remains much slower
than NS 4.x.  A load time of ~20 sec below was obtained using M11.
 Note that under NS 4.6, the load times <= 2 secs.  This is a lot faster!

More specifically, I tried the table by itself and got this result.

*** Timing layout processes on url: 'file:///F|/jdk1.2/docs/api/allclasses-frame
.html', webshell: 019F8B50
failed to set the page title.
Content creation time (this=01A6AB70): Real time 0:0:2, CP time 2.253
Reflow time (this=01A1B220): Real time 0:0:15, CP time 14.531
Frame construction plus style resolution time (this=01A1B220): Real time 0:0:2,
CP time 2.103
Style resolution time (this=01A1B220): Real time 0:0:2, CP time 1.943
Parse Time (this=01A6AC90): Real time 0:0:0, CP time 0.661
DTD Time: Real time 0:0:1, CP time 1.712
Tokenize Time: Real time 0:0:0, CP time 0.150
Total (Layout + Page Load) Time (webshell=019F8B50): Real time 0:0:20, CP time 1
9.418
failed to set the page title.
Document file:///F|/jdk1.2/docs/api/allclasses-frame.html loaded successfully
Document: Done (20.279 secs)
Total (Layout + Page Load) Time (webshell=018F0830): Real time 0:0:0, CP time 0.
611

To verify, I suggest obtaining a local copy of the JDK 1.2 SDK API documentation
from www.javasoft.com.

Comment 3

[leger]

Fine.  When performance gets better, this will get better.  But if you want thie
page checked once this happens, you can do so.

Clearing DUPLICATE resolution due to reopen.  Putting on [PERF] radar.

Comment 4

[Michael Lowe]

*** This bug has been marked as a duplicate of 5588 ***

Comment 5

[leger]

Bulk move of all Necko (to be deleted component) bugs to new Networking

component.

Comment 6

[Tom Everingham]

re-assigning QA contact to myself

Comment 7

[Tom Everingham]

verified dupe of 5588