GitLab

This can only be called after we received a packet, where we already
adjust the clock

Should improve receive performance a bit.

Syncs with libjingle SVN rev 76

Syncs with libjingle SVN rev 56

Also document that fact.

Add to the documentation to note that it can be called several times to
set details for several relay servers during discovery.

Do not unconditionally enable -Werror, since that breaks the build for
people who have stricter compilation warnings enabled.

Instead, add an --enable-compile-warnings configure flag which enables
-Werror when passed as --enable-compile-warnings=error. This mimics the
flag in gnome-common, for consistency. But we enable errors by default
on non-released versions.

This shuts a compiler warning up and allows for format string checking
of debug messages.

This fix may not be entirely cross-platform, and I have been unable to
test whether it is; making it cross-platform is made more difficult by
the fact that the STUN code doesn’t use GLib.

The PRIuPTR macro was defined in POSIX:2004, and later in C99.
http://pubs.opengroup.org/onlinepubs/009696799/basedefs/inttypes.h.html

It gets automatically generated by autotools, so just gets in the way if
checked into git.

Older compilers don't understand all of the flags

It was trying to send queued messages to the final destination instead
of going through the TURN server

div() has an aggregate return, which GCC doesn’t like, although this
seems like a pretty pointless warning because div_t is the same size as
a pointer on 64-bit platforms (probably) and hardly going to cause
performance problems by passing it by value.

Anyway, it seems easier to simplify the code by using explicit / and %
operators inline, than it does to add pragmas and shut the warning up.

If stun_message_append_bytes() is called through
stun_message_append_flag(), data will be NULL and len will be 0. This
will result in a memcpy(ptr, NULL, 0) call. This probably won’t do any
harm (since any reasonable memcpy() implementation will immediately
return if (len == 0)), but the standard allows for memcpy() to explode
if (data == NULL), regardless of the value of len.

In order to be conformant, and to shut up the scan-build static analysis
warning about it, only do the memcpy() if (len > 0).

If (username_len > uname_len), memcpy() would fall off the end of the
uname string and copy (username_len - uname_len) bytes of garbage into
username.

Theoretical bug: not actually attempted to be demonstrated in the wild.

password could be NULL (priv_get_password() can return NULL), but
g_base64_decode() aborts if it receives NULL for the encoded text.

Mostly problems with the specifier for gsize.

This shuts a compiler warning up and allows for format string checking
of debug messages. (See future commits for the fixes this has resulted
in.)

Casting from one struct sockaddr type to another breaks C’s strict
aliasing rules (variables of different types cannot alias). Fix this
cleanly by using unions of struct sockaddrs to convert between the
types (i.e. type-punning).

I wish sockaddr didn’t have to be this painful.

See:
http://gcc.gnu.org/onlinedocs/gcc-4.4.1/gcc/Optimize-Options.html#Type_002dpunning

This shuts GCC’s -Wswitch-default warning, and makes the code flow a
little more explicit.

This introduces no functional changes.

The return value of g_strsplit() can never be NULL.

The latter is deprecated.

Valid IDs will never be 0, which is convenient.

This should be used only when the protocol has it's own keepalive mechanism