First release candidate of Jamulus 3.8.1

[gilgongo]

We have just released version 3.8.1rc1 of Jamulus and would be happy for you to download and test out this version. If all is well, this will become the 3.8.1 release.

You can view the Release Notes and download versions for Windows, Mac, Linux (Debian/Ubuntu), Android (still experimental) and source code here:

https://github.com/jamulussoftware/jamulus/releases/tag/r3_8_1rc1.

If you find any issues with it, please raise an Issue with as many details as possible.

UPDATE

A German translation has now been added to a second release candidate r3_8_1rc2.

[jujudusud]

I build this RC1 on ArchLinux, everything is ok from my point of view.

[rdica]

RC1 builds/runs fine on Fedora 33, both as server and client.

[softins]

One significant feature in this release is support for IPv6 in direct connections to a compatible server. See here for details.

[rdica]

Yes, thats one of the prime reasons I'm testing both as client and server :)

[rdica]

I played about 5 hours last night on my server with a direct IPv6 connection, seemed solid, same latency as I get from IPv4, perhaps a few ms lower. 12 hops to my server over v6, 14 over v4. I'm pleased overall.

[softins]

Great, thanks for the feedback!

[38github]

I tested it on Debian 11 amd64 connected to a server running it also and during my limited testing I found no issues.

[ghost]

Found Warning while compiling:
src/socket.cpp: In member function ‘void CSocket::SendPacket(const CVector&, const CHostAddress&)’:
src/socket.cpp:328:13: warning: ‘status’ may be used uninitialized in this function [-Wmaybe-uninitialized]
if ( status >= 0 )
^
Fix warning
with (line 256):
int status = 0;
instead of:
int status;

[softins]

I found this once when compiling on CentOS 7 with GCC 4.8.5. It doesn't happen on my Pi with GCC 8.3.0, or Ubuntu 18 with GCC 7.5.0. My guess is the newer GCCs can understand the code paths a bit better. But I'll add the initialisation for tidiness.

[ghost]

I just read the man page for sendto() and found that it would be better to initialise:
int status = -1; // default to failure if sendto() is not called.
(my bad for assuming 0 was ok)

man page for sendto():
On success, these calls return the number of characters sent.
On error, -1 is returned, and errno is set appropriately.

[ghost]

A further correction to avoid ambiguity of the -1 error status of sendto():
int status = -2; // create a local -2 failure if sendto() is not called.

[softins]

In this case, I think it’s best as it is. In the almost-impossible case of the code taking the path that doesn’t call status = sendto(), there is no point in retrying, as the same would happen again, so we may as well break out of the loop anyway.