On 2000100408 (Win98, no talkback), I can't open a new window and then *immediately* take newWin.document in order to write to the new window. If I wait, I can get it without any problem. Works fine in 2000100208 and IE. Additionally, I get this in my console: JavaScript error: line 0: uncaught exception: [Exception... "Failure" code: "-2147467259" nsresu lt: "0x80004005 (NS_ERROR_FAILURE)" location: "file:///C:/WINDOWS/Desktop/Copy% 20of%20bleh.html Line: 6"]

Dan, here's the problem I was talking to you about today, I didn't have a chance to file the bug before I saw this one, but this is exactly the problem.

Adding a bunch of keywords. This is IMO a definite MUST FIX for rtm.

Is this regression present in the branch?

I now get this error when attempting to install the PSM on 2000100508/Linux. It appears to install and I get a new window which is blank instead of displaying a success message. On the console, I get the following: JavaScript error: line 0: uncaught exception: [Exception... "Failure" code: "-2147467259" nsresult: "0x80004005 (NS_ERROR_FAILURE)" location: "http://docs.iplanet.com/docs/manuals/psm/psm-mozilla/index.html Line: 33"] Error was suppressed by event handler

Never mind. That seems to be different from this. Will open new bug... sorry about spam.

Or maybe it is the same. PSM install is successful but pop-up confirmation window fails with error code I mentioned. I didn't see a window.open() call at first, but there actually is one on line 32 of the page source. I'll assume this is another manifestation of this bug unless told otherwise.

cc: Tim (this is a Beatnik eMix blocker)

Additional notes: it seems the window isn't quite ready by the time it returns from window.open: the document is missing. For example this function makew() { w=window.open("about:blank", "_blank"); w.location.href="http://www.mozilla.org/"; } doesn't work: it loads about:blank. However this function makew() { w=window.open("about:blank", "_blank"); setTimeout("loadw()", 0); } function loadw() { w.location.href="http://www.mozilla.org/"; } correctly loads mozilla.org. So it's a timing thing. Looking into it.

Got it. It was caused by basically a typo in the checkin for bug 44437. Trunk only.

This is somewhat unrelated, but the testcase brings up another question. Are the buttons supposed to be so tall? IE5.5 displays the code in those buttons all on one line which makes the buttons horizontally long and thin. Mozilla makes these buttons horizontally short and fat. What is the correct behavior??? Jake

hoju - i noticed that too :) bug 55285

Verified with 2000-11-08-01-MN6. Working perfect.