Win98 only 11/18 build - Start viewer.exe - Enter any URL in location field and press enter - Sometimes viewer just hangs, other times it crashes with the following msg: VIEWER caused an invalid page fault in module RAPTORHTML.DLL at 015f:005ac1cc. Registers: EAX=00000000 CS=015f EIP=005ac1cc EFLGS=00010246 EBX=00000000 SS=0167 ESP=0090d92c EBP=0090d95c ECX=00abaf20 DS=0167 ESI=00abb528 FS=2b27 EDX=0090d938 ES=0167 EDI=005e38ef GS=0000 Bytes at CS:EIP: a5 a5 a5 a5 5f 5e c2 08 00 8b 44 24 04 8b 4c 24 Stack dump: 00000000 00000000 005e6730 00abb520 005e38ef 00000000 00000000 0090da98 00000000 00000000 005ac0fa 00abaf20 0090d974 005e615b 00abb520 ffffffff

Same thing happening with xpviewer.exe

Putting ss: in Summary for bug tracking.

Jan, how about 11/19 build???

Okay I get this same problem on 11/19 seamonkey build using Win 98....

I take that back. sometimes the URL are loading(very slow though) and sometimes they are hanging. THis is on Win 98 using 11/19 seamonkey.

The OS should be Win98, according to RickG.

Win 98 is not currently an option in bugzilla

Request sent to terry to include win98 as OS option in bugzilla.

there is no problem with the viewer DEBUG build on Win98 using source code as of 11/20/98 at 1:45pm. Next, look at current code no-debug build.

adding sujay to cc list as he is looking at win98 while I am out sick.

for me with code current as of 1:45pm this afternoon, optimized NON-DEBUG build, many URLs show up fine. yahoo, netscape, lycos, etc. However, on some pages I see severe painting problems which lead me to believe there may be a memory smash somewhere in rendering. I'm tentatively assigning it to michael, and I'll continue to do research. One page that looks particularly bad is http://www.hotwired.com. Also, sample3's top table with background images looks really bad until you resize the window.

i finally got my tree cleaned and have the latest code from cvs. unfortunately, i don't have a machine with win98. is everyone *sure* that the problem can't be reproduced under win95? i'm trying to see if i can get it to go bad under win95 and seeing if there is some machine around that i can use for testing or install win98 on...

okay I just got the updated seamonkey build(11/20) for Win 98 and URLS do seem to load fine...But definitely pages come up slower than the builds we got earlier in the week...

i'm not assigning it to kipp, but i'm adding him to the cc: list per his instructions. it looks like layout may be the first order culprit. both nt and win95 have similar problems, it's just that they are more glaring under win95. BTW: all i'm talking about is the painting!

I fixed the problem with www.hotwired.com not laying out properly (it was due to a scroll-frame not propogating the max-element-size information properly). So now the page lays out properly, though it is painfully slow to load.

as far as i know i'm off the hook, so who gets this now and is it still an issue?

Angus, this is the one you wanted to change the Summary test on. Reminder to do so :-)

I'm changing the summary to reflect the nature of the bug. There were (Are?) "severe painting problems" that buster mentioned which Michael said could be attributed to layout. Are these fixed? If so, please mark fixed. Thanks.

Reassingning to Troy so he can answer our layout-related questinon about this bug's status (see comments)

Taking off ss: list per bug mtg today.

Looks fine now. If I remember right there were problems with lots of scroll frames being created because the IsScrollable() code was checking for a scrollable web shell for all elements and not just for the BODY. The problem affected pages with framesets and Chris fixed it to explicitly check for the\ BODY tag

janc, please verify with latest build.

janc, could you please Verify or Reopen today. I have Win95 and this page looks a little weird. Using Jan 20 build. How does NT look?

1-19-99 build (couldn't get the 1-20 build to unzip...) Tested on Win98 and NT and looks fine now. This should have been marked as "fixed" rather than "worksforme". Changing resolution.

The reason it was marked as WORKSFORME is because I didn't actually do anything. If I had explicitly fixed it I would have marked it as FIXED. That's a long standing rule within engr is that you don't mark it FIXED unless you specifically made a change that fixes it