Closed Bug 5588 Opened 21 years ago Closed 20 years ago

Very slow in viewing JAVA 2 API docs...

Categories

(Core :: Layout: Tables, defect, P3)

x86
Windows NT
defect

Tracking

()

VERIFIED FIXED

People

(Reporter: karnaze, Assigned: troy)

References

()

Details

(Keywords: perf)

This is a spin off of bug 3982. The URL takes a very long time to load and I
suspect the external style sheets in the sub documents.
Assignee: peterl → karnaze
Component: Style System → HTMLTables
No. Style sheet loading is not a significant amount of time in this page.
The delay is in table layout. You can see it by loading the single frame:
http://puck.informatik.med.uni-giessen.de/java/jdk1.2_docs/api/allclasses-frame.
html
Even resizing the window on that page is considerably slow (style is not a
factor there).
Changed URL from
http://puck.informatik.med.uni-giessen.de/java/jdk1.2_docs/api/index.html to
http://puck.informatik.med.uni-giessen.de/java/jdk1.2.1_docs/api/index.html
because I'll delete the old JDK 1.2 docs if all my users have updated their
bookmarks to the new JDK 1.2.1 version.
The slowdown may be caused by the lower-left frame, because index.html links in
the "All Classes" page, which is very huge (141971 bytes in the JDK 1.2.1
version) in size and in the number of lines.
Status: NEW → ASSIGNED
Target Milestone: M7
Moving to M9.
Whiteboard: [Perf]
Blocks: 8691
The url mentioned above loads within what seems to be a reasonable amount of
time, it didn't take that long with the latest build 1999080408 on my Win98
machine.
Target Milestone: M9 → M12
Its still a lot slower than Nav4.6. Moving to M12.
Status: ASSIGNED → RESOLVED
Closed: 20 years ago
Resolution: --- → INVALID
The urls appear to be invalid.
Ouuuchhh, sorry ;-(
I forgot to update the URL - we updated the docs to JDK 1.3 beta.
I fixed the link.
Status: RESOLVED → VERIFIED
Was this mark invalid due to the link? Since the link works, is there still a
problem here?
Status: VERIFIED → REOPENED
Seems to be a mistake.
The bug issue is "performance", only the test URL was broken because we updated
the JDK docs, but I forgot to update this URL.
My pre-M10 SPARC build is still slower than Netscape 4.7 SPARC...
I'll reopen the bug, OK ?
Resolution: INVALID → ---
*** Bug 18495 has been marked as a duplicate of this bug. ***
mass move to m14.
Can someone verify that M11 hangs with these docs ?
It runs quite fine here.
Only the lower left frame hangs for a while, while painting its content.

I am using WIN98 1999122208
> Only the lower left frame hangs for a while, while painting its content

AFAIK it hangs until the CSS has been loaded completely...
Keywords: perf
Bulk add of "perf" to new keyword field.  This will replace the [PERF] we were
using in the Status Summary field.
Status: REOPENED → ASSIGNED
Target Milestone: M14 → M15
Whiteboard: [Perf]
The url loads faster than Nav4.x now on WinNT. Marking worksforme.
Status: ASSIGNED → RESOLVED
Closed: 20 years ago20 years ago
Resolution: --- → WORKSFORME
Can't verify this under Solaris 2.7/MU4 SPARC - but this may be the gerneral
"Unix rendering performance is sluggish/slow" bug (bugid 26502)...
Keywords: verifyme
Can't verify this. Both IE5 and NN4.7 render the lower left frame much faster 
than Mozilla. Reopening...
Status: RESOLVED → REOPENED
Keywords: verifyme
Resolution: WORKSFORME → ---
Waqar, can you take a look.
Assignee: karnaze → waqar
Status: REOPENED → NEW
Troy is looking at this. Reassigning to him.
Assignee: waqar → troy
Fixed the performance problems
Status: ASSIGNED → RESOLVED
Closed: 20 years ago20 years ago
Resolution: --- → FIXED
Performance has greatly improved with this fix, but it still not on par with 
NN4.7.
On my Celeron 550 running Win98 with yesterdays build, this page takes ~1 second 
to open and render with NN4.7.
With Mozilla, it takes ~3 seconds.
There was a specific problem that was causing the page to be extremely slow. 
That problem was fixed and that's why the bug was marked FIXED.
Of course, I meant no offense. The comments in this report where a vague, so I 
thought this bug was more generic, something like a tracker-bug.
The other thing about the page that is still a performance issue is that the 
page has lots of links. That's a known problem because we do too much string 
allocation when determining the link color

However, there were some very specific problems that were nicely demonstrated by 
this page. Problems I haven't seen on any other pages.
Verified fixed.
Status: RESOLVED → VERIFIED
Depends on: CVE-2015-2713
Depends on: 1137723
Depends on: 837007
Depends on: 765409
Depends on: 472950
Depends on: 1001994
Depends on: 401042
You need to log in before you can comment on or make changes to this bug.