Re: gconfd-2 ...

[Havoc Pennington <hp redhat com>]

Michael Meeks <michael ximian com> writes:
> 
> 	I was just getting some super odd problems with gconfd-2 and just
> wondering whether you added any linc stuff to ensure that there would be
> no connections that consumed all the memory.

I don't understand the "connections that consumed all memory" thing here.

> 	I'm just adding some code to linc ( in prototype form ) to stop there
> being an indefinate number of oneways queued locally - so as to block
> the accessibility event emission stuff from over-heating. This may well
> cause gconfd to block too writing to unresponsive clients - it seems you
> didn't use the APIs I created for you in linc/include/linc-connection.h
> to handle this situation, what are you planning to do about it ?

I did use them as far as I know, here is what I have:

     /* Set maximum buffer size, which makes the connection nonblocking
       * if the kernel buffers are full and keeps gconfd from
       * locking up. Set the max to a pretty high number to avoid
       * dropping clients that are just stuck for a while.
       */
      connection = ORBit_small_get_connection (copy);
      ORBit_connection_set_max_buffer (connection, 1024 * 128);
      
Is that wrong?

Havoc