If you are referring to the file-size issues with IE5.5+, yes, I am aware that the error file must be larger than 512KB ... and just to be certain, I padded the file with some deroggatory comments about MSIE that I found when researching the issue.
Even my host provider is scratching their heads on this one, but steadfastly refusing to spend precious resources to identify the issue.
My .htaccess file is sparce - (as we are new domain-name owners, though we've been hosting and writing code since 1999, using GeoCities).
Code: Select all
# Server-side includes
XBitHack full
# Parse php commands
AddHandler application/x-httpd-php .htm .html
# Directory Indexes
Options -Indexes
# Custom Error Pages
ErrorDocument 403 /_errors/403.php
ErrorDocument 404 /_errors/404.php
The 403.php file is not a bother. Works 100% of the time in MSIE and FFox. It's only the 404.php file that seems to be affected.
The way it APPEARS to me is that it's a race to see which makes it first to standard output ... friendly default page, or php-parsed custom page. Generally, the friendly page makes it first. But if the server is busy doing something else, the php-parsing engine ends up being faster and then we actually SEE our custom page on screen. Either way ... the php code is being executed, because the success rate on emailed information is 100%.
EDIT: I have removed the test that was here, that demonstrated the problem, because we have replaced it with the PHP solution that is a work-a-round to the MSIE failure. It is only a hack, but it works reliably, so we are calling it quits and sticking with the UGLY, but effective, work-a-round.
Anyway ... this isn't something that you see with computers very often. Usually, if it fails, it fails 100% of the time. If it works, it works 100% of the time. Here we have a situation ... same .htaccess file, same 404.php file ... yet you end up with different results.
The host has even tried using just a text error in the .htaccess file (rather than a HTML or PHP file ... with the SAME results).
I tried just a plain HTML file and the success rate improved ... more like 80/20 instead of 20/80 ... which leads me to believe it's a SPEED thing.
Anyway ... why try to swim uphill? While it's an interesting problem, my goal is to simple catch 404 traffic. Redirect works. Like to know why, but it's academic at this point.
BTW .. doing the Output buffering flush on the redirected file 404new.php, fixed the problem.
Thanks for your help. This board is an excellent resource. Maybe some day I'll gain enough php knowledge to reciprocate and help others
-Scott