[En-Nut-Discussion] Nut/OS 3.2
NGB
ngbmoreau at yahoo.com.au
Tue Apr 22 14:23:32 CEST 2003
Harald
I submitted 2 pacthes a while ago, one was relating to the incorrect UDP
checksum and the other to code that was incorrect if optimisations were off.
Have they been included in this release ?
Nic
Quoting Harald Kipp <harald.kipp at egnite.de>:
> In order to reduce further delays, we will release Version 3.2 within
> the next days, although there are still open issues.
>
> Here's the list of modifications:
>
> * - Changed: PPP has been redesigned. It is now using a more complete
> * state machine and has been split into a network and a
> * driver part. Ethernet-only application will not link in
> * PPP code anymore.
> * - Added: Chat script interpreter for enhanced modem handling.
> * - Added: LCD driver with VT52 terminal emulation device.
> * - Fixed: Allocating and releasing of NETBUF structures has been
> * inconsistent in case of communication errors and low
> * memory situations. Generally, all transmit routines will
> * now release its NETBUF in case of failures.
> * - Fixed: Undefined signal handler variables with registered timer
> * interrupts. Thanks to Tom Nystrom.
> * - Fixed: Network masks with all bits set had been refused.
> * - Fixed: GCC sets the stack on main entry. As a temporary fix
> * we redefine main during preprocessing.
> * - Changed: Broadcasting to a queue will clear it, even if it had
> * been in signaled state.
> * - Changed: For applications, make install will additionally build a
> * binary and copy it to the bin directory. This is most
> * useful when working with eboot and tftp.
> * - Fixed: When compiling with GCC and different optimizations, NutInit
> * may have failed. Thanks to Ralph Mason.
> * - Fixed: Andre Albsmeier found and fixed a problem with ISC-DHCP V3,
> * which also occured with the DHCP server of some routers.
> * - Fixed: NutUdpReceiveFrom returned the port number in network byte
> * order, which doesn't fit with the documentation and other
> * functions. It will be now returned in host byte order.
> * Thanks to Pavel Chromy.
> * - Added: Two additional devices for debugging output, which can be
> * even used while running in interrupt context.
> * - Changed: NutNetAutoConfig() has been moved to pro/dhcpc and is now
> * deprecated for new applications. A new function named
> * NutDhcpIfConfig() allows for specifying a timeout. Thanks
> * to Pavel Chromy.
> * - Changed: In previous releases network devices had been initialized
> * during interface configuration, while all other devices
> * had been initialized when opening that device. Now all
> * devices are initialized during registration. This will
> * probably break your existing device drivers.
> * - Fixed: Packet overflows in the Realtek network driver caused high
> * interrupt latency times for other interrupt driven devices.
> * - Changed: Complete redesign of VS1001K driver by Pavel Chromy. Great.
> * - Fixed: Memory leak in NutUdpDestroySocket(). Thanks to Ralph Mason.
> * - Fixed: When reading 255, fgetc() returned EOF. Thanks to Ralph Mason.
>
> Before final release, I'd like to send a 3.1.9 preview to people, who are
> interested in testing it and will have the time to report any results ASAP.
> Both, Linux and Windows users are welcome. In the past, this didn't work in
> most cases. I sent out prereleases, wait for a response and nothing
> happened.
> I packed it for a final release and many bug reports came in. :-( So,
> please
> make sure that you are able and willing to spend some time. In the future
> I will release a special developer version in parallel to the final
> release.
> But this will take additional time to prepare.
>
> Thanks again to all contributors,
>
> Harald
>
> _______________________________________________
> En-Nut-Discussion mailing list
> En-Nut-Discussion at egnite.de
> http://www.egnite.de/mailman/listinfo/en-nut-discussion
>
-------------------------------------------------
More information about the En-Nut-Discussion
mailing list