The it is the absolute value of the file descriptor that should not exceed bits in the select() mask, not the number of open descriptors. Different operating systems reuse file descriptor numbers in different ways, TCP sockets that don’t close gracefully can keep descriptors from returning to the pool, raising the fd number over set size. Been many years since I’ve used QNX, but this would be my best guess as to the cause of what you are seeing? Brian

On Mar 31, 2021, at 5:57 AM, George Nistor via Boost wrote:

Hi all,

I am having a problem with the following error message '' The descriptor does not fit into the select call's fd_set"

The setup I have is

QNX , VSomeIP lib (socket abstractisation lib) , lib Boost 1.61 (a very old version) I get the error in this lib VSomeIp from the bellow boost call.

The call to boost::asio::async_write(3 params) gives the above error message.

I have tried to see also on QNX with > pidin -p vsomeip fds How many open file descriptor do I have and I got value 123.

As far as I know there is a limitation 1024 of open file descriptors.

Could you provide some help? Does the latest boost lib solve this particular case?

I appreciate any suggestion, Also please say if any info is missing Thanks

George Nistor

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost