The only way this could cause the failure is if

a) the time delay between the tcp packets is large 

OR

b) the first packet has the PUSH bit set

AND

c) flixster.com's http server implementation is badly broken


Generally speaking tcp applications do not know and cannot determine how 
the data has been split into packets.

The only way I've seen this kind of thing be a problem is with the above 
conditions, where the http server implementation
does something really stupid like assuming that all of the header data 
is in the first read() on the socket, rather than
actually parsing the returned data to see if the header is complete, or 
if more read()'s are required before determining if this is a 1.0 or 1.1 
request.





steve lu wrote:
> Hi All,
>
> While testing my embedded browser (Firefox 2.0.0.6 code base),
> I ran across this site that consistently fails.
>
> The site is Flixster.com and it uses cookie to verify the characters in 
> an image.
> (to counter automatically generated registration requests.).
>
> In my embedded browser, the registration page fails to redirect to the 
> next page.
> and always return me back to the registration page.
>
> In Ethereal, the difference between embedded browser and Firefox 3.0b4 
> is obvious:
>   

> The issue is that the embedded browser's HTTP Request somehow got broken 
> into two packets
> (while Firefox's HTTP request sits in one packet nicely).
>
> I don't know if this explains why Flixster server fails to redirect the 
> embedded browser.
> (But it seems to be the only culprit, since the two POST requests are 
> almost identical except
> the number of IP packets they sit across.)
>
>
> Thanks for any help,
>
> -Steve
>
>
>
> _______________________________________________
> dev-embedding mailing list
> [email protected]
> https://lists.mozilla.org/listinfo/dev-embedding
>   


-- 
Brad Clements,                [EMAIL PROTECTED]    (315)268-1000
http://www.murkworks.com                          
AOL-IM: BKClements

_______________________________________________
dev-embedding mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-embedding

Reply via email to