IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
The Samba4 standard process model uses dup() on incoming sockets as an
optimisation (it makes select() a tiny bit faster when used).
Adding dup() to socket wrapper allows us to use the standard process
model in selftest
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
Currently have almost the same logic in swrap_send(), swrap_sendto(),
swrap_writev() and swrap_sendmsg(), this helper functions
let combine all the logic in 2 places.
metze
This reverts commit 710aa773d5.
We rely on FIONREAD in a lot of other parts in the code,
so there's no need to have an ifdef for it in the socket_wrapper code.
On tru64 FIONREAD is defined in <sys/ioctl.h>
and we include <sys/ioctl.h> via "system/network.h".
Tridge: maybe a HAVE_SYS_IOCTL_H was missing at the time you tried
it on tru64?
If we find a platform that doesn't support it,
we need to bail out at configure time or
provide a replacement in libreplace.
metze
This keeps the restriction for stream sockets (where the caller will
retry), without creating problems on datagram sockets (CLDAP is not
defined, as far as I know, across multiple UDP packets).
The commit adding this restriction was
47b106c0ae
Andrew Bartlett
In convert_in_un_remote() the socket family can be accessed either as
si->family or inaddr->sa_family. We were using the si->family to
determine how to cast the inaddr structure, but if si->family !=
inaddr->sa_family then we will incorrectly be casting a in6 structure
as in4 or vice-versa.
Heimdal will, on supporting systems, set these flags in the type
argument of socket(), causing breakage when combined with
socket_wrapper.
For background on these flags, see http://lwn.net/Articles/281965/
Andrew Bartlett
This is required because the deferred connect code skips the connect()
until sending the packet, but unless we catch this call, the connect()
never happens.
Andrew Bartlett
Now we defer the real_connect() for dgram sockets to the first send()/sendto() call,
as the destination might not be there at connect time.
Commit 66dc53ee01 was an incomplete fix for this
problem.
metze
This reverts commit 66dc53ee01.
Somehow this leads to timeouts in some tests, samba4.ldb.ldapi
and samba4.rpc.schannel. I need to look at it more closely later...
metze
The patch simply uses a more descriptive variable name for tcp_seq.
../lib/socket_wrapper/socket_wrapper.c:753: warning: declaration of 'tcp_seq' shadows a global declaration
/usr/include/netinet/tcp.h:40: warning: shadowed declaration is here
../lib/socket_wrapper/socket_wrapper.c: In function `swrap_marshall_packet':
../lib/socket_wrapper/socket_wrapper.c:919: warning: declaration of 'tcp_seq' shadows a global declaration
/usr/include/netinet/tcp.h:40: warning: shadowed declaration is here
