From Bugzilla Helper: User-Agent: Mozilla/4.76 [en] (Win98; I) BuildID: 2000100204 When doing a phonebook search, the operation just times out. Reproducible: Always Steps to Reproduce: 1. Start N6 2. Go to http://phonebook 3. In the search box, type a username (e.g. barrettl) 4. Click the "Search" button. 5. You may get a couple of boxs pop up about requesting a secure document. Click OK on these. Actual Results: The browser will just hang, until you get the message "The operation timed out when attempting to contact malastare.mcom.com." Expected Results: The search should complete and display the information page on the user, or a list of matches. This works with no problems on Linux.

Updated bug to block #18687.

I'm seeing this too. I think it may be related to PSM, since PSM must be installed in order to get information from this site. Trying Barrett's test case on Linux and HP-UX when PSM isn't installed gives you the message "The connection was refused when attempting to contact <hostname here>"

is this HP only?

Yes, So far it appears to be HP only. Linux is functioning properly.

I'm assuming you are using netscape's commercial branch builds, correct? if so, this might belong into bugscape.

I've just found an external site that exhibits the same behavior. If you go to http://i-bench.zdnet.com/ibench/performance_tests/loadspeed/benchmark.html a series of pages are supposed to load and then it takes you to the test. On HP-UX, the first page loads and then the browser pauses. After a while you get an alert box that states "The operation timed out when attempting to contact i-bench.zdnet.com." Again, this site works just fine on Linux.

Barrett, it looks like the zdnet site was a different bug (56934). It doesn't work on the Linux branch build from the 24th either. A fix was checked in for that on the evening of the 24th and now this site works on both Linux and my HP-UX build from this morning.

I don't think this is the same bug as #56934. That bug states that running the tests crashes the browser. Here, you couldn't even get to the tests. But, using this mornings build, I can now get there. I'm wondering if the phonebook will work now. But without PSM included we can't check it. Is it possible to patch in the older PSM to try this?

You can try the older PSM http://haathi.mcom.com/hp/m16/moz-install.html but I am guessing it won't work. (Please give it a try and let me know what you find - Shannon had reported a problem using the old PSM which I assumed was due to interface changes... but it could have been build related).

I tried loading the old psm yesterday, no go. Same result as Shannon I believe, it just sat there with commercial and mozilla loaded about 2% and then sat there.

After some further review.... we have found the problem is solved by switching skins. However when you quit and start up again the problem re-manifests itself. It would appear that something is not getting set/initialized during startup, but a change of skins just happens to do the trick. :-) ccing ben goodger to get his thoughts

interstingly enough, Michael La Guardia's sidebar pane works fine!! even when it isn't supposed to. here is the url to install it http://rocknroll/users/michaell/publish/addpanel/addpanel.html

the sidebar panel calls quermia for the phonebook query, whereas the phonebook page calls malastare. Tried running the html page for the sidebar panel in the browser and it called successfully to quermia. doesn't explain why a change in chrome would allow a call to malastare to work.

doesn't look like a server problem afterall, instead it looks like there is a problem with displaying in a frame. Specifically, I have taken the code for both the phonebook pages frame and the code for the sidebar page and loaded them in the browser as seperate pages. Both framed and not. The pages that were framed that had a target of resultFrame ( name of the bottom frame ) showed erratic behavior, incudling partial returns of the phonebook page as well as time-outs. The pages that were not framed with targets of _content worked just fine. I tried with target urls using malastare, quermia and phonebook ( which is aliased to both malastare and quermia and would appear to toggle betweent the two ). This behavior holds a little more water when you start thinking about skin changes activating something that fixes this behavior.

Shannon discovered this problem was showing up on Linux around 10/02 and now is fixed. By patching the PSM in to a recent build she was able to determine this problem has been fixed on HP as well ( we knew linux was not showing this bug already ).

Verifying. Works now.