AW: [omniORB] Array size limit

[BaileyK at schneider.com]

I suspect the use of a fixed size array is causing the marshalling code to
try to allocate it on the stack, and it's too big.  The orb configuration
wouldn't help in that case.

Try using sequence<octet> rather than a fixed size array.


---
Kendall Bailey
Engineering & Research


omniorb-list-bounces at omniorb-support.com wrote on 05/07/2008 07:44:55 AM:

> Here is part from omniOrb.cfg
> Is this what you are looking for?
> Kind Regards, Evgeni
> ....
>
############################################################################

> # giopMaxMsgSize
> #
> #    This value defines the ORB-wide limit on the size of GIOP message
> #    (excluding the header). If this limit is exceeded, the ORB will
> #    refuse to send or receive the message and raise a MARSHAL exception.
> #
> #    Valid values = (n >= 8192)
> #
> giopMaxMsgSize = 2097152    # 2 MBytes.
> ....
>
> -----Ursprüngliche Nachricht-----
> Von: omniorb-list-bounces at omniorb-support.com
> [mailto:omniorb-list-bounces at omniorb-support.com] Im Auftrag von Klas
Nordberg
> Gesendet: Mittwoch, 7. Mai 2008 10:29
> An: omniorb-list at omniorb-support.com
> Betreff: [omniORB] Array size limit
>
>
> Hi
>
> I'm testing how long time it takes to send large arrays of bytes
> (CORBA::Octet) between a server and client using omniORB CORBA by
> writing a client which sends 10^6 bytes to a server which just copies
> the received byte array to a local buffer.
>
> On a linux platform this works fine, but when I tested to run the client
> on a Windows platform, I get a stack overflow runtime error (not an
> exception) from somewhere deep in the ORB.  When I decrease the size of
> the array, however, it works.  Apparently, there is a limit on the
> amount of data in an array which can be processed by the ORB.
>
> The idl looks like
>
> typedef octet inputdata[1000000];
>
> interface server {
>   long receivedata(in inputdata data);
> };
>
>
> QUESTION1: Can I increase the limit so that a 10^6  byte array can be
> sent in one chunk also from a Windows client?  Is there a configuration
> variable?
>
> QUESTION2: Is this limit documented somewhere?  Is it a bug?  It appears
> to me that the stack overflow error is not gracefully managed by the ORB
>
> Regards
>
> Klas Nordberg
>