An apparent NN4 page loading bug

Is there a workaround?

[david752]

NN4 (NS4) usually loads HTML files with no problem. However, when the file involves server-side processing (using the PHP language), and it does not already appear in the memory or disk caches, I have observed that sometimes the 'onload' event is not signalled. This results in only some frames being loaded. Rarely, NN4 will hang and must be killed.

Of course, IE5, NS6, and Opera work correctly. However, I cannot ignore NN4, especially since my client is devoted to it.

Over the past month I have tried many schemes involving timeouts and sequential loading (each frame's document loading the next in its 'onload' function), with no real success.

I am using Apache on Windows 98 and Unix, frames, Javascript, PHP, and CSS. And yes, I seem to need all that stuff to make a really nice professional site.

(Note that this question has nothing to do with the infamous NN4 resize bug; this problem is with the 'onload' event.)

I hasten to add that this is not a problem with HTML syntax (see posting "Frames do not completely load.........why?"). Since the problem occurs randomly, and also happens when loading external style sheets, I presume that it is a timing problem related to overlapping page loading.

Do you have any ideas if this is really an NN4 bug and, if so, how to work around it?

Thanks, David

[tedster]

I'm not clear about this. Since onload is a JavaScript event, how does it interfere with the complete loading of a frameset? Or is this a PHP thing (I'm not conversant with PHP).

[david752]

tedster, I am not sure why you found the question unclear. I tried very hard to be very specific.

Yes, onload is a Javascript event, but NN 4.08 stops loading pages into frames, leaving the frames empty (or sometimes it hangs).

PHP is simply a server-side preprocessing language, very much like ASP, JSP, SSI, etc. The only reason I mentioned PHP was that it introduces a sizeable time delay while the server constructs the output and sends it back to the browser.

It looks like this time delay allows loading to begin in subsequent frames before the loading in previous frames completes. Then perhaps a limited tolerance for concurrent I/O, or a race condition, causes NN4 to abandon the page loading.

I can't believe I am the only person ever to have encountered this problem. Of course, few Web designers use six frames as I do...

(The reason I use six is that I want a frame that is surrounded by other frames. If you like, see the site (not using NN4!) at www.accesstherapy.com, and experiment with making the browser window smaller. The site currently has so many kludge/workaround timeouts that you may not see a problem most times you view it using NN4, just some unexpected reloads.)

David

[david752]

I find it strange that there is no interest in this topic, but anyway...

I think I have come up with a reliable solution, although it is not elegant. Since there is no interest here, just contact me and I will give the details.

It is always such a relief to solve a problem that has taken weeks of one's time!

David