rtti for dynamix casts

[yan seiner]

yan seiner wrote:

>Brian Dembinski wrote:
>
>  
>
>>I've included the entire debug output, but the tail of that is :
>>debug: _finished
>>debug: [void KIO::Scheduler::releaseJob(KIO::SimpleJob*, bool)] 
>>called, job:0x84fef10
>>debug: SlaveBase: read() error 2, No such file or directory! exiting!
>>launcher: creating slave for protocol http
>>launcher: error sending launcherSlavePidCmd: Broken pipe
>>debug: Reading Entry Local Protocols/http
>>debug: (850) HTTPProtocol::reparseConfiguration
>>debug: SlaveBase: select() error 9, Bad file descriptor! exiting!
>>debug: SlaveBase: read() error 2, No such file or directory! Exiting!
>>
>>    
>>
>I'm beginning to suspect that it has to do with the remote httpd server 
>closing the connection before konq/e is ready for it....
>
>I've spent the last hour watching this thing crash over and over; the 
>crash happens about the time the httpd child goes away, a second after 
>the flash movie starts.
>
>I'm going to try and patch my httpd server to delay closing the 
>connection, but obviously that won't help you....
>  
>

On further examination, I have no idea what is going on....

Apparently konq/e launches multiple children to deal with some 
things.... I've not traced all that down....

But one of the children attempts to write a big bunch of binary data, 
and fails. I stuck some gDebug statements into connection.cpp to see 
what is going on....

debug: -> KFlashPart::openURL http://127.0.0.1/cgi-bin/pivot.php
debug: KFlashPart::closeURL
debug: KFlashPart::pluginResized()
debug: KFlashPart::pluginResized()
debug: KFlashWidget::initFlash()
launcher: creating slave for protocol http
connection: write failed tried 4096, wrote 4086
connection: data Ú
¸
¸
debug: ERROR: Could not write data - size
debug: ERROR: Could not write data - flush
connection: write failed tried 4096, wrote 4086
connection: data
¸
debug: ERROR: Could not write data - size
debug: ERROR: Could not write data - flush
connection: write failed tried 4096, wrote 4086
connection: data Åj Jû‚îî
debug: ERROR: Could not write data - size
debug: ERROR: Could not write data - flush
connection: write failed tried 4096, wrote 4086
connection: data €f€gjÜJ’
¸
¸
debug: ERROR: Could not write data - size
debug: ERROR: Could not write data - flush
connection: write failed tried 4096, wrote 4086
connection: data Æ
debug: ERROR: Could not write data - size
debug: ERROR: Could not write data - flush
debug: ERROR: Could not write data - flush
debug: ERROR: Could not write data - flush
Segmentation fault

At which point the other processes exit.

What can I provide to track this down? This is potentially a real killer.

--Yan