Hello once more,

Tuesday, July 8, 2003, 3:25:31 PM, you wrote:
MO> Hello too,

MO> Well yes, IE is not that wrong this time (though it changes nothing at all :))
MO> Take a look into the <head> - section of your document, maybe there is
MO> some content-encoding sent or your apache config (maybe some default-encoding,
MO> not sure about, but's a hint).

My <head> looks like that - nothing special..

<head>
<LINK href="/css/main.css" rel=stylesheet type=text/css>
<TITLE>Some title</TITLE>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-2">
<script language="JavaScript" src="/javascripts/date_picker/date_picker.js"></script>
</head>

And nothing set in Apache.. :-/

MO> As mentioned, the bug exactly is, that ie get's a wrong content-type and therefore
MO> can't decompress the g-zip stuff. i'd say sending correct g-zip headers may
MO> not be wrong at all, though i must say you have to relay on google on more
MO> info on that. don't sure about them.

I tried to find something about that with google, but no effects so
far..

MO> but why the heck the ie-thingy can decode on a refresh-hit is not logical.

ie is not logical at all ;) And more serious - ie shows correct page
if user click the link to page being a problematic, but redirecting or
moving to such page from <form> with "action" option doesn't work and
we have blank page.

MO> The 4k-thing is becaus mod_gzip doesn't compress under this size.

I can't agree with you here. I checked with telnet what I get. And I
what I saw was surely zipped content. No matter less or more 4kB page
was. But somehow over 4kB IE understands what to do.

MO> (don't get me wrong but in some way i'm happy about someone having the same bug - 
MO> so the time to figure this one out isn't lost one :) - definitly nothing personal)

Sure, I can say the same ;) Good to have someone to talk to about that
problem. :)

MO> regards
MO> mario


--
Best regards,
Sebastian


Tuesday, July 8, 2003, 3:37:27 PM, you wrote:
sdcp> Hello,

sdcp> Good to see someone else met that "feature" of MS Internet
sdcp> Exploder ;)

sdcp> At really I don't put any headers with content info.
sdcp> Only headers are:

sdcp>  header ('Last-Modified: '.gmdate("D, d M Y H:i:s").' GMT');
sdcp>  header ('Expires: '.gmdate("D, d M Y H:i:s").' GMT');
sdcp>  header ('Cache-Control: no-cache, must-revalidate');
sdcp>  header ('Pragma: no-cache');

sdcp> And output is simple HTML. What should I do so? Should I send any
sdcp> other header to fix that?

sdcp> And I also spent a lot of time wondering "why the hell I get blank page
sdcp> sometimes?". Especially it was working good on one server and buggy
sdcp> on another (now I now, another server had compression turned off).

sdcp> --
sdcp> Best regards,
sdcp> Sebastian

sdcp> Tuesday, July 8, 2003, 2:41:28 PM, you wrote:
>>> Hello,
>>> 
>>> I've found problem with MSIE showing sometimes blank page when
>>> compression is turned on (no matter using zlib.output_compression or
>>> gzip handler). I've found that happens when page is <=4096 bytes long
>>> (output in HTML). And this suggested me to try to change buffer size
>>> for compression (for zlib, that I was using) cause it was set to
>>> exactly 4KB. But that didn't help at all :-( If I add space characters
>>> at the end of output to let page grow ower 4096 bytes - everything
>>> works fine, else MSIE shows blank page and user has to hit "Refresh"
>>> button to see correct page. With other browsers that problem doesn't
>>> happen.
>>> 
>>> Do you have any idea how to solve that? BTW - changing browser or
>>> doing anything with user system is not good idea for me, cause I can't
>>> force users of my application to do anything - they use (and will use)
>>> windows+IE and so that's my problem to find solution. And output
>>> compression is a MUST for me cause I have to cut network traffic to
>>> lower users costs of using internet (GPRS transmission using mobile
>>> phones).
>>> 
>>> Big thanks in advance for any help
>>> 
>>> Sebastian Baran

MO>> heyho lucky one

MO>> thought not anyone else would have that "bug". it's a very nasty one (at
MO>> least it was in my case) to figure out. i'd suggest looking for the
MO>> content-type you are sending with. may it be it's a text/html or
MO>> somewhat thislike?

MO>> regards
MO>> mario


-- 
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php

Reply via email to